riscv
/
coriolis
mirror of https://gitlab.lip6.fr/vlsi-eda/coriolis.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity
coriolis/anabatic/src/AutoSegment.cpp

3002 lines
103 KiB
C++
Raw Normal View History

Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
// -*- C++ -*-
//
// This file is part of the Coriolis Software.
Happy New Year 2018 ! Update license years...
2018-01-06 10:55:44 -06:00
// Copyright (c) UPMC 2008-2018, All Rights Reserved
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
//
// +-----------------------------------------------------------------+
// | C O R I O L I S |
// | A n a b a t i c - Routing Toolbox |
// | |
// | Author : Jean-Paul CHAPUT |
// | E-mail : Jean-Paul.Chaput@lip6.fr |
// | =============================================================== |
// | C++ Module : "./AutoSegment.cpp" |
// +-----------------------------------------------------------------+
Manage minimum area for VIAs in the P&R. * New: In Hurricane::Layer, add support for a minimal area (given in microns). Exported in the Python interface. * New: In Anabatic::AutoSegment::bloatStackedStrap(), method to enlarge too small vertically stacked VIAs. For now just create a square ensuring the minimal area. Potentially cause DRC error in some cases so needs to be refined. * New: In Anabatic::_gutAnabatic(), check for too little stacked VIAs and display a report.
2020-11-15 17:55:49 -06:00
#include <cmath>
Improved handling of short nets (fully included in one GCell). The short net mode degrade the routing in some cases. This will be fixed in a next batch of commits. * New: In Hurricane::NetRoutingProperty, added "ShortNet" flag for Nets that are completly inside *one* GCell. * Bug: In CRL::BlifParser::Model::staticInit(), when looking for the output of zero and one cell, also skip the blockage net (as well as automatic and supplies). * New: In Anabatic::AutoSegment, added "ShortNet" flag to know if the segment is part of a short net (fully included in *one* GCell). Also add accessor/mutators for the _analogMode flag (was it ever used before?). * New: In Anabatic::NetBuilder::singleGCell(), if a RoutingPad is vertically small, add a vertical segment to give it some slack. * New: In Anabatic::Dijkstra::_materialize(), detect "short net" as they have only one GCell in their source list... * Bug: In AnabaticEngine::_loadGrbyNet(), reset the AutoSegment "short net" and "analog mode" creation flags between two different nets. * New: In Katana::Configuration, added dedicated ripup for short net segmnts. * New: In Katana: partially implemented support for "short dogleg", that is dogleg that are always kept in same metal because they connect neighboring perpandicular tracks. Not finished neither activated yet. * New: In Katana::TreckElement and derived, export the the *short net* support from AutoSegment. * Bug: In Katana::RoutingEvent::_processRepair(), when a segment is successfully inserted, re-process any perpandicular that is in repair state, as it may have a new chance to be placed. * New: In Katana::SegmentFsm::slackenTopology(), always reject short nets. * Bug: In Katana::Track::check(), correctly handle wide segments instead of issuing false check messages.
2018-07-16 04:16:51 -05:00
#include "hurricane/DebugSession.h"
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
#include "hurricane/Warning.h"
#include "hurricane/Bug.h"
#include "hurricane/DataBase.h"
#include "hurricane/Technology.h"
Basic support for FreePDK 45 completed. * New: In Commons, inspector support for std::pair<T,U>. * New: In Hurricane::Layer, ContactLayer & ViaLayer, support for non square VIAs. The hole (cut) remains square, but the various metal extensions can now be different in X and Y. The ::getEnclosure() method now takes a flag EnclosureH / EnclosureV. * New: In Hurricane::DbU, inspector support for: std::pair<DbU::Unit,DbU::Unit> std::array<DbU::Unit,3> Must be defined here as DbU do not exists yet in Commons.h * Bug: In Hurricane::Interval::getSize(), when the interval is "full span", do not return the difference between min and max, but directly DbU::Max. (the previous result was -1 !) * New: In CRL Core Python/Technology.py, support for non square VIAs in the configuration files. Applied to FreePDK 45. * New: In CRL::RoutingGauge, added a "symbolic" flag to tell if a gauge is for symbolic layout or not. Exported to Python. * New: In Anabatic::AutoHorizontal::updatePosition(), differentiated computation for soure or target taking account of the VIA extension in the right segment metal (due to non-square VIAs). * Change: In Anabatic::AutoHorizontal::_makeDogleg(), the dogleg is UP for HV gauges and DOWN for VH. * New: In Anabatic::AutoSegment::_initialize(), create a cache of the various extension length for each layer (viaToTop, viaToBottom, viaToSame). New implementation of getExtensionCap() using the previous cached extension table. See updatePositions(). New static functions to access the extension cache in the header: getViaTotopCap() ... * Change: In Anabatic::AutoSegment, in various update methods, updateOrient() must always be called *before* updatePositions() as extensions are dependant on source/target. * New: In Anabatic::AutoSegment::getEndAxes() compute the position of the first source and last target position (center/axes) on an *aligned* set of segments. * New: In Anabatic::AutoSegment, add a new state flag SegAxisFixed to signal segments that can be put on only one track. Specific case to VH gauge for a M1 vertical terminal with a M2 vertical segment. The M2 is effectively bound to the M1 axis position. * Bug: In Anabatic::NetBuilderVH::_do_xG_xM1_xM3(), in case of E/W global and only one RoutingPad the connexion to the RoutingPad was duplicated. It was valid, but totally stupid. * Bug: In Anabatic::Session::_canonize(), for an aligned segment set, intersect the user constraints from all segments instead of only considering the canonical one. Issue a warning about too tight constraints only for symbolic gauges. It may be correct for the real ones. * New: In Katata::DataNegociate::update(), more accurate computation of the perpandicular free interval. Use segment extension cap calculation. Create a special case for fixed axis segments allowing them to find alternative free interval, try under source and under target as they are likely to be draggable segments. * Change: In Katana::Manipulator::relax(), use the extension cap value to compute the axis of the perpandicular segemnts. * Change: In Katana::Manipulator::moveUp(), now move up the whole set of aligned segments instead of just the canonical one. * Change: In Katana::NegociateWindow::loadRoutingPads(), more accurate TrackMarkers insertions for fixed terminals. * New: In Katana::RoutingEvent::Key::Compare::operator(), segments with fixed axis are processed prior to any others. * New: In Katana::RoutingEventLoop, store segment pointers instead of ids to generate more accurate error messages. * Change: In Katana::RoutingPlane::create(), perform local track assignment only for HV gauges. * Change: In Katana::SegmentFsm::_slackenLocal(), add a "dragMinimize" step in the automaton. Mutliple states transitions can occurs in a row if an action fails. * New: In Katana::Session::_toIntervalAxis(), normalize interval bounds so they are on track positions (by shrinking the interval). * Bug: In Katana::TrackMarker CTOR, the weigh computation was wrong.
2018-02-17 13:27:38 -06:00
#include "hurricane/ViaLayer.h"
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
#include "hurricane/Horizontal.h"
#include "hurricane/Vertical.h"
#include "crlcore/RoutingGauge.h"
#include "anabatic/Session.h"
#include "anabatic/AutoContact.h"
#include "anabatic/AutoSegment.h"
#include "anabatic/AutoHorizontal.h"
#include "anabatic/AutoVertical.h"
#include "anabatic/GCell.h"
#include "anabatic/AnabaticEngine.h"
namespace {
using namespace std;
using namespace CRL;
using namespace Hurricane;
using namespace Anabatic;
// ---------------------------------------------------------------
// Local Variables.
const char* badAutoSegmentAnchor =
"AutoSegment::create() :\n\n"
" Source and/or target anchor is NOT an <AutoContact> (internal error).\n"
" Source: %s, Target: %s";
const char* dupAutoSegmentAnchor =
"AutoSegment::create() :\n\n"
" Source and Target anchor are the same : %s (internal error).";
const char* badSegment =
"Anabatic::AutoSegment::create () :\n\n"
" Segment between %s and %s\n"
" is neither horizontal nor vertical .\n";
const char* badSegmentSource =
"Anabatic::AutoSegment::create () :\n\n"
" Source anchor of segment %s is not a Contact\n"
" (%s)\n";
const char* badSegmentTarget =
"Anabatic::AutoSegment::create () :\n\n"
" Source anchor of segment %s is not a Contact\n"
" (%s)\n";
const char* mismatchSegmentSource =
"Anabatic::AutoSegment::create () :\n\n"
" Source anchor of segment %s is already an AutoContact\n"
" (%s)\n";
const char* mismatchSegmentTarget =
"Anabatic::AutoSegment::create () :\n\n"
" Target anchor of segment %s is already an AutoContact\n"
" (%s)\n";
// ---------------------------------------------------------------
// Local Functions.
bool getTerminalInterval ( AutoSegment* autoSegment
, AutoContact* fromContact
, bool isHorizontal
, DbU::Unit& min
, DbU::Unit& max
)
{
AutoContact* terminalContact = NULL;
if ( !fromContact ) {
bool found = getTerminalInterval ( autoSegment
, autoSegment->getAutoSource()
, autoSegment->isHorizontal()
, min
, max );
if ( !found )
found = getTerminalInterval ( autoSegment
, autoSegment->getAutoTarget()
, autoSegment->isHorizontal()
, min
, max );
//if ( !found )
// cerr << "[ERROR] Cannot find terminal of " << autoSegment << "." << endl;
return found;
} else {
if ( autoSegment->isGlobal() ) return false;
cdebug_log(145,0) << "Examining " << autoSegment << " " << fromContact << endl;
if ( autoSegment->getSource() == autoSegment->getTarget() ) {
cerr << Error("Source & Target are the same :\n"
" %s\n %s"
,getString(autoSegment).c_str()
,getString(autoSegment->getSource()).c_str()) << endl;
}
terminalContact = autoSegment->getAutoSource();
if ( terminalContact == fromContact ) {
terminalContact = autoSegment->getAutoTarget();
}
if ( !terminalContact->isTerminal() ) {
AutoSegment* segment = NULL;
size_t segmentCount = 0;
forEach ( Component*, icomponent, terminalContact->getSlaveComponents() ) {
if ( *icomponent == autoSegment->base() ) continue;
Segment* connex = dynamic_cast<Segment*>(*icomponent);
if ( !connex ) continue;
segment = Session::lookup ( connex );
if ( not segment or not segment->isWeakTerminal() ) continue;
segmentCount++;
}
if ( segmentCount == 1 ) {
return getTerminalInterval ( segment, terminalContact, isHorizontal, min, max );
return false;
}
} else {
cdebug_log(145,0) << "Terminal is " << terminalContact << endl;
Box constraintBox = terminalContact->getConstraintBox();
if ( isHorizontal ) {
min = constraintBox.getXMin ();
max = constraintBox.getXMax ();
} else {
min = constraintBox.getYMin ();
max = constraintBox.getYMax ();
}
return true;
}
}
return false;
}
// ---------------------------------------------------------------
// Class : "AttractorsMap".
class AttractorsMap {
// Constructor.
public:
inline AttractorsMap ();
inline size_t getAttractorsCount () const;
DbU::Unit getLowerMedian () const;
DbU::Unit getUpperMedian () const;
void addAttractor ( DbU::Unit position );
protected:
map<DbU::Unit,size_t> _attractors;
size_t _attractorsCount;
};
inline AttractorsMap::AttractorsMap ()
: _attractors(), _attractorsCount(0)
{ }
inline size_t AttractorsMap::getAttractorsCount () const
{
return _attractorsCount;
}
void AttractorsMap::addAttractor ( DbU::Unit position )
{
_attractors[position]++;
_attractorsCount++;
First intergration of the Analogic router parts. * New: In Anabatic::AutoSegment, introduce a the kind (associated to a flag) "LongLocal". Analog GCells can be very wide, so at least some carefuly choosen long local segments must be took into account as attractors in the computation of the optimal axis. * New: In Anabatic::AutoSegment::computeOptimal(), take LongLocal into account as attractors. * Change: In ::GCellTopology constructors compare the layers of the RoutingPads using layer masks instead of Layer pointers. Allows to find both "METALx" (symbolic) and "metalX" (real). * Change: In ::GCellTopology::_doHChannel(), _doChannel(), _doStrut() and _doDevice(), tag long locals as "LongLocal". This need to be reviewed as it as bind done a bit too quickly. * Change: In Anabatic::AutoSegment, due too a much bigger span of the analogic GCells the _optimalMin & _optimalMax bitfields must use 16 bits instead of 8 (they where overflowed). * New: In Katana, reorganisation of the initialization procedure to fit both digital and analogic cases. Create an analogInit() method. * Change: In Katana::RoutingEvent, the _tracksNb and _tracksFree bitfields where too short for the Analog GCell size, now uses 16 bits instead of 6. * Bug: In Katana::GraphicKatanEngine::drawGCell(), skip drawing of a GCell if *both* width and height are under 150 pixels. * New: In Katana::Session, add a new isOpen() method.
2016-10-04 10:12:58 -05:00
cdebug_log(145,0) << "add Attractor @" << DbU::getValueString(position)
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
<< " [" << _attractors[position] << "]" << endl;
}
DbU::Unit AttractorsMap::getLowerMedian () const
{
size_t median = (_attractorsCount/2) + (_attractorsCount%2);
size_t lower = 0;
map<DbU::Unit,size_t>::const_iterator it = _attractors.begin ();
for ( ; it != _attractors.end() ; it++ ) {
lower += it->second;
if ( lower >= median ) break;
}
return it->first;
}
DbU::Unit AttractorsMap::getUpperMedian () const
{
size_t median = _attractorsCount / 2;
size_t upper = 0;
map<DbU::Unit,size_t>::const_iterator it = _attractors.begin ();
for ( ; it != _attractors.end() ; it++ ) {
upper += it->second;
if ( upper > median ) break;
}
return it->first;
}
Synchronize priority of TrackSegments connecteds through doglegs. * Bug: In Anabatic::Dijsktra, correct the indentation in cdebug calls (lots of them causing a big shift right). * New: In Anabatic::TrackSegment, new helper structure SideStack to manage a set of aligned GCells and their various sides sizes. * Change: In Anabatic::TrackSegment::computeOptimal(), more accurate computation of attractors from global segments and variable size GCells using SideStack. * Change: In Katana::DataSymmetric::checkPairing(), increase the tolerance for misaligned symmetrics from 2 to 5 tracks (should be enough for our narrow channel routing). * New: In Katana::TrackSegment, add a first flag to enable locking of priority. If it is set, calls to either "computePriority()" or "forcePriority()" will have no effect. Added the uint32_t flags paraphernalia. * New: In ::computeNetPriority(), overall function to control the call of TrackSegment::computeAlignedPriority(). The call is done from NegociateWindow::run(). * New: Katana::TrackSegment::computeAlignedPriority(), order the TrackSegments aligneds through doglegs to the one with the highest priority is routed first and others progressively from him. Done by forcing an ever decreasing priority on the aligneds ones. The amount of decrease is small so the aligned segments got routed (ordered) in close, if not contiguous, sequence. Priority is locked for the order to remain. * Bug: In Katana::TrackSegment::computePriority(), correct computation of the priority when there is more than 10 free tracks (a DbU::toLambda() call was missing, leading to very big priorities). * Change: In katana::TrackCost CTOR, do not compute a distance to fixed in the case of analog segments, this is backfiring. Slight change of the compare function when delta differs. Seems to improve a little.
2017-05-30 15:33:06 -05:00
// ---------------------------------------------------------------
// Class : "SideStack".
class SideStack {
public:
Bad computation of optimal position for analog segments. * Bug: In Anabatic::AutoSegment::computeOptimal(), the "side stack" computing the intersection of all the sides of the GCell that the segment is going through was implicitly supposing that the resulting intersection wasn't empty. But for analog segment, we compute the position of all the segments linked though simple doglegs, and as dogleg occurs there can be a sufficient shift so the intersections of all the sides became empty. Now the "side stack" can manage configuration when instead of having an intersection we have a "hole" in the GCell side span (only one hole is supported at the moment). When a hole is present, the optimal position is the center of the hole. Should be refined so that the position is computed according to the longest GCell that we go through. * Change: In Katana::SegmentFsm::_slackenGlobal(), when slackening a global from a ananlog net, allow to make a dog leg once instead of immediately moving up. The break point is the center of the segment. * Change: In Katana::Manipulator::Makedogleg(DbU::Unit) if the break position is over a device, choose the left of right GCell.
2018-02-21 10:03:43 -06:00
SideStack ( Flags direction, DbU::Unit pitch );
inline bool isHoled () const;
const Interval& getSideAt ( DbU::Unit ) const;
inline const Interval& getGSide () const;
inline DbU::Unit getGSideMin () const;
inline DbU::Unit getGSideMax () const;
inline DbU::Unit getGSideCenter () const;
void addGCell ( const GCell* );
inline bool intersect ( const Interval& ) const;
inline void restrictGSide ( const Interval& );
void show () const;
Synchronize priority of TrackSegments connecteds through doglegs. * Bug: In Anabatic::Dijsktra, correct the indentation in cdebug calls (lots of them causing a big shift right). * New: In Anabatic::TrackSegment, new helper structure SideStack to manage a set of aligned GCells and their various sides sizes. * Change: In Anabatic::TrackSegment::computeOptimal(), more accurate computation of attractors from global segments and variable size GCells using SideStack. * Change: In Katana::DataSymmetric::checkPairing(), increase the tolerance for misaligned symmetrics from 2 to 5 tracks (should be enough for our narrow channel routing). * New: In Katana::TrackSegment, add a first flag to enable locking of priority. If it is set, calls to either "computePriority()" or "forcePriority()" will have no effect. Added the uint32_t flags paraphernalia. * New: In ::computeNetPriority(), overall function to control the call of TrackSegment::computeAlignedPriority(). The call is done from NegociateWindow::run(). * New: Katana::TrackSegment::computeAlignedPriority(), order the TrackSegments aligneds through doglegs to the one with the highest priority is routed first and others progressively from him. Done by forcing an ever decreasing priority on the aligneds ones. The amount of decrease is small so the aligned segments got routed (ordered) in close, if not contiguous, sequence. Priority is locked for the order to remain. * Bug: In Katana::TrackSegment::computePriority(), correct computation of the priority when there is more than 10 free tracks (a DbU::toLambda() call was missing, leading to very big priorities). * Change: In katana::TrackCost CTOR, do not compute a distance to fixed in the case of analog segments, this is backfiring. Slight change of the compare function when delta differs. Seems to improve a little.
2017-05-30 15:33:06 -05:00
private:
Flags _direction;
Bad computation of optimal position for analog segments. * Bug: In Anabatic::AutoSegment::computeOptimal(), the "side stack" computing the intersection of all the sides of the GCell that the segment is going through was implicitly supposing that the resulting intersection wasn't empty. But for analog segment, we compute the position of all the segments linked though simple doglegs, and as dogleg occurs there can be a sufficient shift so the intersections of all the sides became empty. Now the "side stack" can manage configuration when instead of having an intersection we have a "hole" in the GCell side span (only one hole is supported at the moment). When a hole is present, the optimal position is the center of the hole. Should be refined so that the position is computed according to the longest GCell that we go through. * Change: In Katana::SegmentFsm::_slackenGlobal(), when slackening a global from a ananlog net, allow to make a dog leg once instead of immediately moving up. The break point is the center of the segment. * Change: In Katana::Manipulator::Makedogleg(DbU::Unit) if the break position is over a device, choose the left of right GCell.
2018-02-21 10:03:43 -06:00
bool _holed;
Synchronize priority of TrackSegments connecteds through doglegs. * Bug: In Anabatic::Dijsktra, correct the indentation in cdebug calls (lots of them causing a big shift right). * New: In Anabatic::TrackSegment, new helper structure SideStack to manage a set of aligned GCells and their various sides sizes. * Change: In Anabatic::TrackSegment::computeOptimal(), more accurate computation of attractors from global segments and variable size GCells using SideStack. * Change: In Katana::DataSymmetric::checkPairing(), increase the tolerance for misaligned symmetrics from 2 to 5 tracks (should be enough for our narrow channel routing). * New: In Katana::TrackSegment, add a first flag to enable locking of priority. If it is set, calls to either "computePriority()" or "forcePriority()" will have no effect. Added the uint32_t flags paraphernalia. * New: In ::computeNetPriority(), overall function to control the call of TrackSegment::computeAlignedPriority(). The call is done from NegociateWindow::run(). * New: Katana::TrackSegment::computeAlignedPriority(), order the TrackSegments aligneds through doglegs to the one with the highest priority is routed first and others progressively from him. Done by forcing an ever decreasing priority on the aligneds ones. The amount of decrease is small so the aligned segments got routed (ordered) in close, if not contiguous, sequence. Priority is locked for the order to remain. * Bug: In Katana::TrackSegment::computePriority(), correct computation of the priority when there is more than 10 free tracks (a DbU::toLambda() call was missing, leading to very big priorities). * Change: In katana::TrackCost CTOR, do not compute a distance to fixed in the case of analog segments, this is backfiring. Slight change of the compare function when delta differs. Seems to improve a little.
2017-05-30 15:33:06 -05:00
DbU::Unit _pitch;
Interval _full;
Interval _gside;
map<DbU::Unit,Interval> _sides;
};
SideStack::SideStack ( Flags direction, DbU::Unit pitch )
: _direction( (direction & Flags::Horizontal) ? Flags::Vertical : Flags::Horizontal )
Bad computation of optimal position for analog segments. * Bug: In Anabatic::AutoSegment::computeOptimal(), the "side stack" computing the intersection of all the sides of the GCell that the segment is going through was implicitly supposing that the resulting intersection wasn't empty. But for analog segment, we compute the position of all the segments linked though simple doglegs, and as dogleg occurs there can be a sufficient shift so the intersections of all the sides became empty. Now the "side stack" can manage configuration when instead of having an intersection we have a "hole" in the GCell side span (only one hole is supported at the moment). When a hole is present, the optimal position is the center of the hole. Should be refined so that the position is computed according to the longest GCell that we go through. * Change: In Katana::SegmentFsm::_slackenGlobal(), when slackening a global from a ananlog net, allow to make a dog leg once instead of immediately moving up. The break point is the center of the segment. * Change: In Katana::Manipulator::Makedogleg(DbU::Unit) if the break position is over a device, choose the left of right GCell.
2018-02-21 10:03:43 -06:00
, _holed (false)
Synchronize priority of TrackSegments connecteds through doglegs. * Bug: In Anabatic::Dijsktra, correct the indentation in cdebug calls (lots of them causing a big shift right). * New: In Anabatic::TrackSegment, new helper structure SideStack to manage a set of aligned GCells and their various sides sizes. * Change: In Anabatic::TrackSegment::computeOptimal(), more accurate computation of attractors from global segments and variable size GCells using SideStack. * Change: In Katana::DataSymmetric::checkPairing(), increase the tolerance for misaligned symmetrics from 2 to 5 tracks (should be enough for our narrow channel routing). * New: In Katana::TrackSegment, add a first flag to enable locking of priority. If it is set, calls to either "computePriority()" or "forcePriority()" will have no effect. Added the uint32_t flags paraphernalia. * New: In ::computeNetPriority(), overall function to control the call of TrackSegment::computeAlignedPriority(). The call is done from NegociateWindow::run(). * New: Katana::TrackSegment::computeAlignedPriority(), order the TrackSegments aligneds through doglegs to the one with the highest priority is routed first and others progressively from him. Done by forcing an ever decreasing priority on the aligneds ones. The amount of decrease is small so the aligned segments got routed (ordered) in close, if not contiguous, sequence. Priority is locked for the order to remain. * Bug: In Katana::TrackSegment::computePriority(), correct computation of the priority when there is more than 10 free tracks (a DbU::toLambda() call was missing, leading to very big priorities). * Change: In katana::TrackCost CTOR, do not compute a distance to fixed in the case of analog segments, this is backfiring. Slight change of the compare function when delta differs. Seems to improve a little.
2017-05-30 15:33:06 -05:00
, _pitch (pitch)
, _full (false)
, _gside (false)
, _sides ()
{ }
Bad computation of optimal position for analog segments. * Bug: In Anabatic::AutoSegment::computeOptimal(), the "side stack" computing the intersection of all the sides of the GCell that the segment is going through was implicitly supposing that the resulting intersection wasn't empty. But for analog segment, we compute the position of all the segments linked though simple doglegs, and as dogleg occurs there can be a sufficient shift so the intersections of all the sides became empty. Now the "side stack" can manage configuration when instead of having an intersection we have a "hole" in the GCell side span (only one hole is supported at the moment). When a hole is present, the optimal position is the center of the hole. Should be refined so that the position is computed according to the longest GCell that we go through. * Change: In Katana::SegmentFsm::_slackenGlobal(), when slackening a global from a ananlog net, allow to make a dog leg once instead of immediately moving up. The break point is the center of the segment. * Change: In Katana::Manipulator::Makedogleg(DbU::Unit) if the break position is over a device, choose the left of right GCell.
2018-02-21 10:03:43 -06:00
inline bool SideStack::isHoled () const { return _holed; }
inline const Interval& SideStack::getGSide () const { return _gside; }
inline DbU::Unit SideStack::getGSideMin () const { return _gside.getVMin(); }
inline DbU::Unit SideStack::getGSideMax () const { return _gside.getVMax(); }
inline DbU::Unit SideStack::getGSideCenter () const { return _gside.getVMax(); }
inline bool SideStack::intersect ( const Interval& side ) const { return _gside.intersect(side); }
inline void SideStack::restrictGSide ( const Interval& restrict ) { if (not _holed) _gside.intersection(restrict); }
Synchronize priority of TrackSegments connecteds through doglegs. * Bug: In Anabatic::Dijsktra, correct the indentation in cdebug calls (lots of them causing a big shift right). * New: In Anabatic::TrackSegment, new helper structure SideStack to manage a set of aligned GCells and their various sides sizes. * Change: In Anabatic::TrackSegment::computeOptimal(), more accurate computation of attractors from global segments and variable size GCells using SideStack. * Change: In Katana::DataSymmetric::checkPairing(), increase the tolerance for misaligned symmetrics from 2 to 5 tracks (should be enough for our narrow channel routing). * New: In Katana::TrackSegment, add a first flag to enable locking of priority. If it is set, calls to either "computePriority()" or "forcePriority()" will have no effect. Added the uint32_t flags paraphernalia. * New: In ::computeNetPriority(), overall function to control the call of TrackSegment::computeAlignedPriority(). The call is done from NegociateWindow::run(). * New: Katana::TrackSegment::computeAlignedPriority(), order the TrackSegments aligneds through doglegs to the one with the highest priority is routed first and others progressively from him. Done by forcing an ever decreasing priority on the aligneds ones. The amount of decrease is small so the aligned segments got routed (ordered) in close, if not contiguous, sequence. Priority is locked for the order to remain. * Bug: In Katana::TrackSegment::computePriority(), correct computation of the priority when there is more than 10 free tracks (a DbU::toLambda() call was missing, leading to very big priorities). * Change: In katana::TrackCost CTOR, do not compute a distance to fixed in the case of analog segments, this is backfiring. Slight change of the compare function when delta differs. Seems to improve a little.
2017-05-30 15:33:06 -05:00
const Interval& SideStack::getSideAt ( DbU::Unit position ) const
{
if (_sides.empty()) return _full;
if (_sides.size() == 1) return _sides.begin()->second;
if (_sides.begin()->first > position) return _sides.begin()->second;
for ( auto iside = ++_sides.begin() ; iside != _sides.end() ; ++iside ) {
if (iside->first >= position) return (--iside)->second;
}
return _sides.rbegin()->second;
}
void SideStack::addGCell ( const GCell* gcell )
{
Interval side = gcell->getSide( _direction, _pitch );
DbU::Unit position = (_direction & Flags::Vertical) ? gcell->getBoundingBox().getXMin()
: gcell->getBoundingBox().getYMin();
Bad computation of optimal position for analog segments. * Bug: In Anabatic::AutoSegment::computeOptimal(), the "side stack" computing the intersection of all the sides of the GCell that the segment is going through was implicitly supposing that the resulting intersection wasn't empty. But for analog segment, we compute the position of all the segments linked though simple doglegs, and as dogleg occurs there can be a sufficient shift so the intersections of all the sides became empty. Now the "side stack" can manage configuration when instead of having an intersection we have a "hole" in the GCell side span (only one hole is supported at the moment). When a hole is present, the optimal position is the center of the hole. Should be refined so that the position is computed according to the longest GCell that we go through. * Change: In Katana::SegmentFsm::_slackenGlobal(), when slackening a global from a ananlog net, allow to make a dog leg once instead of immediately moving up. The break point is the center of the segment. * Change: In Katana::Manipulator::Makedogleg(DbU::Unit) if the break position is over a device, choose the left of right GCell.
2018-02-21 10:03:43 -06:00
if (not _holed and intersect(side)) _gside.intersection( side );
else {
if (not _holed) {
_holed = true;
if (side.getVMin() > _gside.getVMax()) _gside = Interval( _gside.getVMax(), side.getVMin() );
else _gside = Interval( side.getVMax(), _gside.getVMin() );
} else {
if (not intersect(side)) {
if (side.getVMax() < _gside.getVMin()) _gside.merge( side.getVMax() );
else _gside.merge( side.getVMin() );
}
}
}
Synchronize priority of TrackSegments connecteds through doglegs. * Bug: In Anabatic::Dijsktra, correct the indentation in cdebug calls (lots of them causing a big shift right). * New: In Anabatic::TrackSegment, new helper structure SideStack to manage a set of aligned GCells and their various sides sizes. * Change: In Anabatic::TrackSegment::computeOptimal(), more accurate computation of attractors from global segments and variable size GCells using SideStack. * Change: In Katana::DataSymmetric::checkPairing(), increase the tolerance for misaligned symmetrics from 2 to 5 tracks (should be enough for our narrow channel routing). * New: In Katana::TrackSegment, add a first flag to enable locking of priority. If it is set, calls to either "computePriority()" or "forcePriority()" will have no effect. Added the uint32_t flags paraphernalia. * New: In ::computeNetPriority(), overall function to control the call of TrackSegment::computeAlignedPriority(). The call is done from NegociateWindow::run(). * New: Katana::TrackSegment::computeAlignedPriority(), order the TrackSegments aligneds through doglegs to the one with the highest priority is routed first and others progressively from him. Done by forcing an ever decreasing priority on the aligneds ones. The amount of decrease is small so the aligned segments got routed (ordered) in close, if not contiguous, sequence. Priority is locked for the order to remain. * Bug: In Katana::TrackSegment::computePriority(), correct computation of the priority when there is more than 10 free tracks (a DbU::toLambda() call was missing, leading to very big priorities). * Change: In katana::TrackCost CTOR, do not compute a distance to fixed in the case of analog segments, this is backfiring. Slight change of the compare function when delta differs. Seems to improve a little.
2017-05-30 15:33:06 -05:00
_sides.insert( make_pair(position,side) );
}
void SideStack::show () const
{
cdebug_log(145,0) << "SideStack::show()" << endl;
for ( auto pside : _sides ) {
cdebug_log(145,0) << "@ " << DbU::getValueString(pside.first)
<< " " << pside.second << endl;
}
}
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
} // End of local namespace.
namespace Anabatic {
// -------------------------------------------------------------------
// Class : "Anabatic::AutoSegment::CompareByDepthLength".
bool AutoSegment::CompareByDepthLength::operator() ( AutoSegment* lhs, AutoSegment* rhs ) const
{
int deltaDepth = (int)(Session::getRoutingGauge()->getLayerDepth(lhs->getLayer()))
- (int)(Session::getRoutingGauge()->getLayerDepth(rhs->getLayer()));
if ( deltaDepth < 0 ) return true; // Lowest layer first.
if ( deltaDepth > 0 ) return false;
DbU::Unit deltaUnit = lhs->getSourceU() - rhs->getSourceU();
if ( deltaUnit < 0 ) return true; // Smallest source first.
if ( deltaUnit > 0 ) return false;
deltaUnit = lhs->getLength() - rhs->getLength();
if ( deltaUnit > 0 ) return true; // Longest first.
Groudwork for routing density driven placement. Compliance with clang 5.0.1. This commit contains two set of features that should have been commited separately. 1. Compliance with clang 5.0.1, tested with the RedHat collection llvm-toolset-7. This allow Coriolis to be compiled under Darwin (MacOS) with Xcode & macports. The bootstrap install system has been modificated accordingly. 2. The basic support for routing density driven placement. Related features are: * Bloat property. Each Occurrence of an Instance can be individually bloated. This property not attached to any tool to allow the placer and router to share it as wanted. Nevertheless, it is defined in Etesian. * BloatProfile in Katana, add individual Bloat properties to Instances occurrences based on the East & North overflowed edges of each GCell. * Support in ToolEngine for a "pass number" of a tool. This pass number is mainly used to make "per pass" measurements. The MeasureSet system is improved accordingly to support multiple values of a same measure. * Embryo of "P&R Conductor" to perform the place & route loop until the design is successfully placed. May be the first brick of a Silicon Compiler. * Change: In boostrap/FindBoostrap.cmake, in setup_boost(), added tag to the python component for macport (ex: python27). * Change: In boostrap/build.conf, put etesian before anabatic for instance occurrence BloatProperty dependency. Added option support for the "llvm-toolset-7" collection to build against clang 5.0.1. * Bug: In Hurricane::getRecord( const pair<T,U>& ), the getSlot<> templates for first & second arguments must be called with <const T> and <const U> as the pair itself is const (and not simply <T> & <U>). * Change: In Hurricane::getSlot() temlate, only use "string" arguments and not const string&, simpler for template argument deduction. * Bug: In Hurricane::AnalogCellExtension, the StandardPrivateProperty<> template has a static member "_name". Clang did show that the template for this static number has to be put inside the namespace where the template *is defined* (i.e. Hurricane) instead of the namespace where it is instanciated (i.e. Analog). * Bug: In Isobar, Matrix_FromListOfList(), PyInt_AsPlacementStatus() must be put outside the C linkage back in the Isobar C++ namespace (clang). * Bug: In Hurricane::DBo::~DBo, and derived add a throw() specification (clang). * Bug: In Hurricane::RegularLayer::getEnclosure() & setEnclosure(), change signature so it matches the one of the base class (clang). * Bug: In Hurricane::CellPrinter, use double brackets for initializer list (clang). * Change: In Hurricane::Breakpoint, reverse the meaning of the error level. Only error level *lesser or equal* than the stop level will be enabled. * Bug: In CRL/python/helpers/__init__.loadUserSettings(), must put the current working directory in the sys.path as in certain configuration it may not be included. * Bug: In CRL::ApDriver, DumpSegments(), no longer generate segments when encountering a RoutingPad on a top-level Pin Occurrence. The segment was generated in the wrong direction, creating DRC violations on the "mips_core_flat" example. * Change: In CRL::Measures, partial re-design of the measurements management. Now, each kind of measure can accept multiple values put in a vector. The index is intented to match a tool run number. * Change: In CRL::Histogram, add support for multiple sets of datas, indexeds with tool run number. * Change: In CRL::ToolEngine, add support for multiple pass number, add addMeasure<> templates for the various data-types. * Change: In CRL::gdsDriver & CRL::gdsParser(), comment out unused GDS record name constants. * New: Etesian::BloatProperty, property to attach to Instance occurrences that contains the extra number of pitch to add to the cell width. * Bug: In AutoSegment::CompareByDepthLength, the segment length comparison was wrong, it was always returning true, which broke the "strick weak ordering" of the comparison. This was producing a core-dump in GCell::updateDensity() when sorting a vector<>. The end() iterator was being dereferenced, leading to the problem. * Bug: In Katana::DataSymmetric::checkPairing(), the test for segments whose axis is perpandicular to the symmetry axis was wrong ("!=" instead of "-"). * New: In Katana/GlobalRoute, new ::selectSegments(), selectOverloadedgcells() and selectBloatedInstances() to automatically select segments from overloaded edges, overloaded GCells and bloated cells. * Change: In KatanaEngine, return a more detailed success state, distinguish between global and detailed. Add support for multiple routing iterations. * New: In cumulus/python/plugins/ConductorPlugin.py, embryo of routing driven placement.
2019-12-08 18:57:44 -06:00
if ( deltaUnit < 0 ) return false;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
deltaUnit = lhs->getAxis() - rhs->getAxis();
if ( deltaUnit < 0 ) return true; // Smallest axis first.
if ( deltaUnit > 0 ) return false;
Groudwork for routing density driven placement. Compliance with clang 5.0.1. This commit contains two set of features that should have been commited separately. 1. Compliance with clang 5.0.1, tested with the RedHat collection llvm-toolset-7. This allow Coriolis to be compiled under Darwin (MacOS) with Xcode & macports. The bootstrap install system has been modificated accordingly. 2. The basic support for routing density driven placement. Related features are: * Bloat property. Each Occurrence of an Instance can be individually bloated. This property not attached to any tool to allow the placer and router to share it as wanted. Nevertheless, it is defined in Etesian. * BloatProfile in Katana, add individual Bloat properties to Instances occurrences based on the East & North overflowed edges of each GCell. * Support in ToolEngine for a "pass number" of a tool. This pass number is mainly used to make "per pass" measurements. The MeasureSet system is improved accordingly to support multiple values of a same measure. * Embryo of "P&R Conductor" to perform the place & route loop until the design is successfully placed. May be the first brick of a Silicon Compiler. * Change: In boostrap/FindBoostrap.cmake, in setup_boost(), added tag to the python component for macport (ex: python27). * Change: In boostrap/build.conf, put etesian before anabatic for instance occurrence BloatProperty dependency. Added option support for the "llvm-toolset-7" collection to build against clang 5.0.1. * Bug: In Hurricane::getRecord( const pair<T,U>& ), the getSlot<> templates for first & second arguments must be called with <const T> and <const U> as the pair itself is const (and not simply <T> & <U>). * Change: In Hurricane::getSlot() temlate, only use "string" arguments and not const string&, simpler for template argument deduction. * Bug: In Hurricane::AnalogCellExtension, the StandardPrivateProperty<> template has a static member "_name". Clang did show that the template for this static number has to be put inside the namespace where the template *is defined* (i.e. Hurricane) instead of the namespace where it is instanciated (i.e. Analog). * Bug: In Isobar, Matrix_FromListOfList(), PyInt_AsPlacementStatus() must be put outside the C linkage back in the Isobar C++ namespace (clang). * Bug: In Hurricane::DBo::~DBo, and derived add a throw() specification (clang). * Bug: In Hurricane::RegularLayer::getEnclosure() & setEnclosure(), change signature so it matches the one of the base class (clang). * Bug: In Hurricane::CellPrinter, use double brackets for initializer list (clang). * Change: In Hurricane::Breakpoint, reverse the meaning of the error level. Only error level *lesser or equal* than the stop level will be enabled. * Bug: In CRL/python/helpers/__init__.loadUserSettings(), must put the current working directory in the sys.path as in certain configuration it may not be included. * Bug: In CRL::ApDriver, DumpSegments(), no longer generate segments when encountering a RoutingPad on a top-level Pin Occurrence. The segment was generated in the wrong direction, creating DRC violations on the "mips_core_flat" example. * Change: In CRL::Measures, partial re-design of the measurements management. Now, each kind of measure can accept multiple values put in a vector. The index is intented to match a tool run number. * Change: In CRL::Histogram, add support for multiple sets of datas, indexeds with tool run number. * Change: In CRL::ToolEngine, add support for multiple pass number, add addMeasure<> templates for the various data-types. * Change: In CRL::gdsDriver & CRL::gdsParser(), comment out unused GDS record name constants. * New: Etesian::BloatProperty, property to attach to Instance occurrences that contains the extra number of pitch to add to the cell width. * Bug: In AutoSegment::CompareByDepthLength, the segment length comparison was wrong, it was always returning true, which broke the "strick weak ordering" of the comparison. This was producing a core-dump in GCell::updateDensity() when sorting a vector<>. The end() iterator was being dereferenced, leading to the problem. * Bug: In Katana::DataSymmetric::checkPairing(), the test for segments whose axis is perpandicular to the symmetry axis was wrong ("!=" instead of "-"). * New: In Katana/GlobalRoute, new ::selectSegments(), selectOverloadedgcells() and selectBloatedInstances() to automatically select segments from overloaded edges, overloaded GCells and bloated cells. * Change: In KatanaEngine, return a more detailed success state, distinguish between global and detailed. Add support for multiple routing iterations. * New: In cumulus/python/plugins/ConductorPlugin.py, embryo of routing driven placement.
2019-12-08 18:57:44 -06:00
#if THIS_IS_DISABLED
if ( lhs->isCanonical () xor rhs->isCanonical () ) return lhs->isCanonical();
if ( lhs->isCollapsed () xor rhs->isCollapsed () ) return rhs->isCollapsed();
if ( lhs->isSlackenStrap() xor rhs->isSlackenStrap() ) return lhs->isSlackenStrap();
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
Groudwork for routing density driven placement. Compliance with clang 5.0.1. This commit contains two set of features that should have been commited separately. 1. Compliance with clang 5.0.1, tested with the RedHat collection llvm-toolset-7. This allow Coriolis to be compiled under Darwin (MacOS) with Xcode & macports. The bootstrap install system has been modificated accordingly. 2. The basic support for routing density driven placement. Related features are: * Bloat property. Each Occurrence of an Instance can be individually bloated. This property not attached to any tool to allow the placer and router to share it as wanted. Nevertheless, it is defined in Etesian. * BloatProfile in Katana, add individual Bloat properties to Instances occurrences based on the East & North overflowed edges of each GCell. * Support in ToolEngine for a "pass number" of a tool. This pass number is mainly used to make "per pass" measurements. The MeasureSet system is improved accordingly to support multiple values of a same measure. * Embryo of "P&R Conductor" to perform the place & route loop until the design is successfully placed. May be the first brick of a Silicon Compiler. * Change: In boostrap/FindBoostrap.cmake, in setup_boost(), added tag to the python component for macport (ex: python27). * Change: In boostrap/build.conf, put etesian before anabatic for instance occurrence BloatProperty dependency. Added option support for the "llvm-toolset-7" collection to build against clang 5.0.1. * Bug: In Hurricane::getRecord( const pair<T,U>& ), the getSlot<> templates for first & second arguments must be called with <const T> and <const U> as the pair itself is const (and not simply <T> & <U>). * Change: In Hurricane::getSlot() temlate, only use "string" arguments and not const string&, simpler for template argument deduction. * Bug: In Hurricane::AnalogCellExtension, the StandardPrivateProperty<> template has a static member "_name". Clang did show that the template for this static number has to be put inside the namespace where the template *is defined* (i.e. Hurricane) instead of the namespace where it is instanciated (i.e. Analog). * Bug: In Isobar, Matrix_FromListOfList(), PyInt_AsPlacementStatus() must be put outside the C linkage back in the Isobar C++ namespace (clang). * Bug: In Hurricane::DBo::~DBo, and derived add a throw() specification (clang). * Bug: In Hurricane::RegularLayer::getEnclosure() & setEnclosure(), change signature so it matches the one of the base class (clang). * Bug: In Hurricane::CellPrinter, use double brackets for initializer list (clang). * Change: In Hurricane::Breakpoint, reverse the meaning of the error level. Only error level *lesser or equal* than the stop level will be enabled. * Bug: In CRL/python/helpers/__init__.loadUserSettings(), must put the current working directory in the sys.path as in certain configuration it may not be included. * Bug: In CRL::ApDriver, DumpSegments(), no longer generate segments when encountering a RoutingPad on a top-level Pin Occurrence. The segment was generated in the wrong direction, creating DRC violations on the "mips_core_flat" example. * Change: In CRL::Measures, partial re-design of the measurements management. Now, each kind of measure can accept multiple values put in a vector. The index is intented to match a tool run number. * Change: In CRL::Histogram, add support for multiple sets of datas, indexeds with tool run number. * Change: In CRL::ToolEngine, add support for multiple pass number, add addMeasure<> templates for the various data-types. * Change: In CRL::gdsDriver & CRL::gdsParser(), comment out unused GDS record name constants. * New: Etesian::BloatProperty, property to attach to Instance occurrences that contains the extra number of pitch to add to the cell width. * Bug: In AutoSegment::CompareByDepthLength, the segment length comparison was wrong, it was always returning true, which broke the "strick weak ordering" of the comparison. This was producing a core-dump in GCell::updateDensity() when sorting a vector<>. The end() iterator was being dereferenced, leading to the problem. * Bug: In Katana::DataSymmetric::checkPairing(), the test for segments whose axis is perpandicular to the symmetry axis was wrong ("!=" instead of "-"). * New: In Katana/GlobalRoute, new ::selectSegments(), selectOverloadedgcells() and selectBloatedInstances() to automatically select segments from overloaded edges, overloaded GCells and bloated cells. * Change: In KatanaEngine, return a more detailed success state, distinguish between global and detailed. Add support for multiple routing iterations. * New: In cumulus/python/plugins/ConductorPlugin.py, embryo of routing driven placement.
2019-12-08 18:57:44 -06:00
if ( lhs->isGlobal () xor rhs->isGlobal () ) return lhs->isGlobal();
if ( lhs->isTerminal () xor rhs->isTerminal () ) return rhs->isTerminal();
if ( lhs->isHorizontal() xor rhs->isHorizontal() ) return lhs->isHorizontal();
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
Groudwork for routing density driven placement. Compliance with clang 5.0.1. This commit contains two set of features that should have been commited separately. 1. Compliance with clang 5.0.1, tested with the RedHat collection llvm-toolset-7. This allow Coriolis to be compiled under Darwin (MacOS) with Xcode & macports. The bootstrap install system has been modificated accordingly. 2. The basic support for routing density driven placement. Related features are: * Bloat property. Each Occurrence of an Instance can be individually bloated. This property not attached to any tool to allow the placer and router to share it as wanted. Nevertheless, it is defined in Etesian. * BloatProfile in Katana, add individual Bloat properties to Instances occurrences based on the East & North overflowed edges of each GCell. * Support in ToolEngine for a "pass number" of a tool. This pass number is mainly used to make "per pass" measurements. The MeasureSet system is improved accordingly to support multiple values of a same measure. * Embryo of "P&R Conductor" to perform the place & route loop until the design is successfully placed. May be the first brick of a Silicon Compiler. * Change: In boostrap/FindBoostrap.cmake, in setup_boost(), added tag to the python component for macport (ex: python27). * Change: In boostrap/build.conf, put etesian before anabatic for instance occurrence BloatProperty dependency. Added option support for the "llvm-toolset-7" collection to build against clang 5.0.1. * Bug: In Hurricane::getRecord( const pair<T,U>& ), the getSlot<> templates for first & second arguments must be called with <const T> and <const U> as the pair itself is const (and not simply <T> & <U>). * Change: In Hurricane::getSlot() temlate, only use "string" arguments and not const string&, simpler for template argument deduction. * Bug: In Hurricane::AnalogCellExtension, the StandardPrivateProperty<> template has a static member "_name". Clang did show that the template for this static number has to be put inside the namespace where the template *is defined* (i.e. Hurricane) instead of the namespace where it is instanciated (i.e. Analog). * Bug: In Isobar, Matrix_FromListOfList(), PyInt_AsPlacementStatus() must be put outside the C linkage back in the Isobar C++ namespace (clang). * Bug: In Hurricane::DBo::~DBo, and derived add a throw() specification (clang). * Bug: In Hurricane::RegularLayer::getEnclosure() & setEnclosure(), change signature so it matches the one of the base class (clang). * Bug: In Hurricane::CellPrinter, use double brackets for initializer list (clang). * Change: In Hurricane::Breakpoint, reverse the meaning of the error level. Only error level *lesser or equal* than the stop level will be enabled. * Bug: In CRL/python/helpers/__init__.loadUserSettings(), must put the current working directory in the sys.path as in certain configuration it may not be included. * Bug: In CRL::ApDriver, DumpSegments(), no longer generate segments when encountering a RoutingPad on a top-level Pin Occurrence. The segment was generated in the wrong direction, creating DRC violations on the "mips_core_flat" example. * Change: In CRL::Measures, partial re-design of the measurements management. Now, each kind of measure can accept multiple values put in a vector. The index is intented to match a tool run number. * Change: In CRL::Histogram, add support for multiple sets of datas, indexeds with tool run number. * Change: In CRL::ToolEngine, add support for multiple pass number, add addMeasure<> templates for the various data-types. * Change: In CRL::gdsDriver & CRL::gdsParser(), comment out unused GDS record name constants. * New: Etesian::BloatProperty, property to attach to Instance occurrences that contains the extra number of pitch to add to the cell width. * Bug: In AutoSegment::CompareByDepthLength, the segment length comparison was wrong, it was always returning true, which broke the "strick weak ordering" of the comparison. This was producing a core-dump in GCell::updateDensity() when sorting a vector<>. The end() iterator was being dereferenced, leading to the problem. * Bug: In Katana::DataSymmetric::checkPairing(), the test for segments whose axis is perpandicular to the symmetry axis was wrong ("!=" instead of "-"). * New: In Katana/GlobalRoute, new ::selectSegments(), selectOverloadedgcells() and selectBloatedInstances() to automatically select segments from overloaded edges, overloaded GCells and bloated cells. * Change: In KatanaEngine, return a more detailed success state, distinguish between global and detailed. Add support for multiple routing iterations. * New: In cumulus/python/plugins/ConductorPlugin.py, embryo of routing driven placement.
2019-12-08 18:57:44 -06:00
if ( lhs->isFixed() xor rhs->isFixed() ) return lhs->isFixed();
#endif
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
return lhs->getId() < rhs->getId();
}
// -------------------------------------------------------------------
// Class : "Anabatic::AutoSegment::CompareByDepthAxis".
bool AutoSegment::CompareByDepthAxis::operator() ( AutoSegment* lhs, AutoSegment* rhs ) const
{
int deltaDepth = (int)(Session::getRoutingGauge()->getLayerDepth(lhs->getLayer()))
- (int)(Session::getRoutingGauge()->getLayerDepth(rhs->getLayer()));
if ( deltaDepth < 0 ) return true; // Lowest layer first.
if ( deltaDepth > 0 ) return false;
DbU::Unit deltaUnit = lhs->getAxis() - rhs->getAxis();
if ( deltaUnit < 0 ) return true; // Smallest axis first.
if ( deltaUnit > 0 ) return false;
deltaUnit = lhs->getSourceU() - rhs->getSourceU();
if ( deltaUnit < 0 ) return true; // Smallest source first.
if ( deltaUnit > 0 ) return false;
return lhs->getId() < rhs->getId(); // Smallest Id first.
}
Basic support for FreePDK 45 completed. * New: In Commons, inspector support for std::pair<T,U>. * New: In Hurricane::Layer, ContactLayer & ViaLayer, support for non square VIAs. The hole (cut) remains square, but the various metal extensions can now be different in X and Y. The ::getEnclosure() method now takes a flag EnclosureH / EnclosureV. * New: In Hurricane::DbU, inspector support for: std::pair<DbU::Unit,DbU::Unit> std::array<DbU::Unit,3> Must be defined here as DbU do not exists yet in Commons.h * Bug: In Hurricane::Interval::getSize(), when the interval is "full span", do not return the difference between min and max, but directly DbU::Max. (the previous result was -1 !) * New: In CRL Core Python/Technology.py, support for non square VIAs in the configuration files. Applied to FreePDK 45. * New: In CRL::RoutingGauge, added a "symbolic" flag to tell if a gauge is for symbolic layout or not. Exported to Python. * New: In Anabatic::AutoHorizontal::updatePosition(), differentiated computation for soure or target taking account of the VIA extension in the right segment metal (due to non-square VIAs). * Change: In Anabatic::AutoHorizontal::_makeDogleg(), the dogleg is UP for HV gauges and DOWN for VH. * New: In Anabatic::AutoSegment::_initialize(), create a cache of the various extension length for each layer (viaToTop, viaToBottom, viaToSame). New implementation of getExtensionCap() using the previous cached extension table. See updatePositions(). New static functions to access the extension cache in the header: getViaTotopCap() ... * Change: In Anabatic::AutoSegment, in various update methods, updateOrient() must always be called *before* updatePositions() as extensions are dependant on source/target. * New: In Anabatic::AutoSegment::getEndAxes() compute the position of the first source and last target position (center/axes) on an *aligned* set of segments. * New: In Anabatic::AutoSegment, add a new state flag SegAxisFixed to signal segments that can be put on only one track. Specific case to VH gauge for a M1 vertical terminal with a M2 vertical segment. The M2 is effectively bound to the M1 axis position. * Bug: In Anabatic::NetBuilderVH::_do_xG_xM1_xM3(), in case of E/W global and only one RoutingPad the connexion to the RoutingPad was duplicated. It was valid, but totally stupid. * Bug: In Anabatic::Session::_canonize(), for an aligned segment set, intersect the user constraints from all segments instead of only considering the canonical one. Issue a warning about too tight constraints only for symbolic gauges. It may be correct for the real ones. * New: In Katata::DataNegociate::update(), more accurate computation of the perpandicular free interval. Use segment extension cap calculation. Create a special case for fixed axis segments allowing them to find alternative free interval, try under source and under target as they are likely to be draggable segments. * Change: In Katana::Manipulator::relax(), use the extension cap value to compute the axis of the perpandicular segemnts. * Change: In Katana::Manipulator::moveUp(), now move up the whole set of aligned segments instead of just the canonical one. * Change: In Katana::NegociateWindow::loadRoutingPads(), more accurate TrackMarkers insertions for fixed terminals. * New: In Katana::RoutingEvent::Key::Compare::operator(), segments with fixed axis are processed prior to any others. * New: In Katana::RoutingEventLoop, store segment pointers instead of ids to generate more accurate error messages. * Change: In Katana::RoutingPlane::create(), perform local track assignment only for HV gauges. * Change: In Katana::SegmentFsm::_slackenLocal(), add a "dragMinimize" step in the automaton. Mutliple states transitions can occurs in a row if an action fails. * New: In Katana::Session::_toIntervalAxis(), normalize interval bounds so they are on track positions (by shrinking the interval). * Bug: In Katana::TrackMarker CTOR, the weigh computation was wrong.
2018-02-17 13:27:38 -06:00
// -------------------------------------------------------------------
// Class : "Anabatic::AutoSegment::CompareBySourceU".
bool AutoSegment::CompareBySourceU::operator() ( AutoSegment* lhs, AutoSegment* rhs ) const
{
DbU::Unit deltaUnit = lhs->getSourceU() - rhs->getSourceU();
if (deltaUnit < 0) return true; // Smallest source first.
if (deltaUnit > 0) return false;
return lhs->getId() < rhs->getId(); // Smallest Id first.
}
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
// -------------------------------------------------------------------
// Class : "Anabatic::AutoSegment".
Basic support for FreePDK 45 completed. * New: In Commons, inspector support for std::pair<T,U>. * New: In Hurricane::Layer, ContactLayer & ViaLayer, support for non square VIAs. The hole (cut) remains square, but the various metal extensions can now be different in X and Y. The ::getEnclosure() method now takes a flag EnclosureH / EnclosureV. * New: In Hurricane::DbU, inspector support for: std::pair<DbU::Unit,DbU::Unit> std::array<DbU::Unit,3> Must be defined here as DbU do not exists yet in Commons.h * Bug: In Hurricane::Interval::getSize(), when the interval is "full span", do not return the difference between min and max, but directly DbU::Max. (the previous result was -1 !) * New: In CRL Core Python/Technology.py, support for non square VIAs in the configuration files. Applied to FreePDK 45. * New: In CRL::RoutingGauge, added a "symbolic" flag to tell if a gauge is for symbolic layout or not. Exported to Python. * New: In Anabatic::AutoHorizontal::updatePosition(), differentiated computation for soure or target taking account of the VIA extension in the right segment metal (due to non-square VIAs). * Change: In Anabatic::AutoHorizontal::_makeDogleg(), the dogleg is UP for HV gauges and DOWN for VH. * New: In Anabatic::AutoSegment::_initialize(), create a cache of the various extension length for each layer (viaToTop, viaToBottom, viaToSame). New implementation of getExtensionCap() using the previous cached extension table. See updatePositions(). New static functions to access the extension cache in the header: getViaTotopCap() ... * Change: In Anabatic::AutoSegment, in various update methods, updateOrient() must always be called *before* updatePositions() as extensions are dependant on source/target. * New: In Anabatic::AutoSegment::getEndAxes() compute the position of the first source and last target position (center/axes) on an *aligned* set of segments. * New: In Anabatic::AutoSegment, add a new state flag SegAxisFixed to signal segments that can be put on only one track. Specific case to VH gauge for a M1 vertical terminal with a M2 vertical segment. The M2 is effectively bound to the M1 axis position. * Bug: In Anabatic::NetBuilderVH::_do_xG_xM1_xM3(), in case of E/W global and only one RoutingPad the connexion to the RoutingPad was duplicated. It was valid, but totally stupid. * Bug: In Anabatic::Session::_canonize(), for an aligned segment set, intersect the user constraints from all segments instead of only considering the canonical one. Issue a warning about too tight constraints only for symbolic gauges. It may be correct for the real ones. * New: In Katata::DataNegociate::update(), more accurate computation of the perpandicular free interval. Use segment extension cap calculation. Create a special case for fixed axis segments allowing them to find alternative free interval, try under source and under target as they are likely to be draggable segments. * Change: In Katana::Manipulator::relax(), use the extension cap value to compute the axis of the perpandicular segemnts. * Change: In Katana::Manipulator::moveUp(), now move up the whole set of aligned segments instead of just the canonical one. * Change: In Katana::NegociateWindow::loadRoutingPads(), more accurate TrackMarkers insertions for fixed terminals. * New: In Katana::RoutingEvent::Key::Compare::operator(), segments with fixed axis are processed prior to any others. * New: In Katana::RoutingEventLoop, store segment pointers instead of ids to generate more accurate error messages. * Change: In Katana::RoutingPlane::create(), perform local track assignment only for HV gauges. * Change: In Katana::SegmentFsm::_slackenLocal(), add a "dragMinimize" step in the automaton. Mutliple states transitions can occurs in a row if an action fails. * New: In Katana::Session::_toIntervalAxis(), normalize interval bounds so they are on track positions (by shrinking the interval). * Bug: In Katana::TrackMarker CTOR, the weigh computation was wrong.
2018-02-17 13:27:38 -06:00
size_t AutoSegment::_allocateds = 0;
size_t AutoSegment::_globalsCount = 0;
bool AutoSegment::_analogMode = false;
Improved handling of short nets (fully included in one GCell). The short net mode degrade the routing in some cases. This will be fixed in a next batch of commits. * New: In Hurricane::NetRoutingProperty, added "ShortNet" flag for Nets that are completly inside *one* GCell. * Bug: In CRL::BlifParser::Model::staticInit(), when looking for the output of zero and one cell, also skip the blockage net (as well as automatic and supplies). * New: In Anabatic::AutoSegment, added "ShortNet" flag to know if the segment is part of a short net (fully included in *one* GCell). Also add accessor/mutators for the _analogMode flag (was it ever used before?). * New: In Anabatic::NetBuilder::singleGCell(), if a RoutingPad is vertically small, add a vertical segment to give it some slack. * New: In Anabatic::Dijkstra::_materialize(), detect "short net" as they have only one GCell in their source list... * Bug: In AnabaticEngine::_loadGrbyNet(), reset the AutoSegment "short net" and "analog mode" creation flags between two different nets. * New: In Katana::Configuration, added dedicated ripup for short net segmnts. * New: In Katana: partially implemented support for "short dogleg", that is dogleg that are always kept in same metal because they connect neighboring perpandicular tracks. Not finished neither activated yet. * New: In Katana::TreckElement and derived, export the the *short net* support from AutoSegment. * Bug: In Katana::RoutingEvent::_processRepair(), when a segment is successfully inserted, re-process any perpandicular that is in repair state, as it may have a new chance to be placed. * New: In Katana::SegmentFsm::slackenTopology(), always reject short nets. * Bug: In Katana::Track::check(), correctly handle wide segments instead of issuing false check messages.
2018-07-16 04:16:51 -05:00
bool AutoSegment::_shortNetMode = false;
Basic support for FreePDK 45 completed. * New: In Commons, inspector support for std::pair<T,U>. * New: In Hurricane::Layer, ContactLayer & ViaLayer, support for non square VIAs. The hole (cut) remains square, but the various metal extensions can now be different in X and Y. The ::getEnclosure() method now takes a flag EnclosureH / EnclosureV. * New: In Hurricane::DbU, inspector support for: std::pair<DbU::Unit,DbU::Unit> std::array<DbU::Unit,3> Must be defined here as DbU do not exists yet in Commons.h * Bug: In Hurricane::Interval::getSize(), when the interval is "full span", do not return the difference between min and max, but directly DbU::Max. (the previous result was -1 !) * New: In CRL Core Python/Technology.py, support for non square VIAs in the configuration files. Applied to FreePDK 45. * New: In CRL::RoutingGauge, added a "symbolic" flag to tell if a gauge is for symbolic layout or not. Exported to Python. * New: In Anabatic::AutoHorizontal::updatePosition(), differentiated computation for soure or target taking account of the VIA extension in the right segment metal (due to non-square VIAs). * Change: In Anabatic::AutoHorizontal::_makeDogleg(), the dogleg is UP for HV gauges and DOWN for VH. * New: In Anabatic::AutoSegment::_initialize(), create a cache of the various extension length for each layer (viaToTop, viaToBottom, viaToSame). New implementation of getExtensionCap() using the previous cached extension table. See updatePositions(). New static functions to access the extension cache in the header: getViaTotopCap() ... * Change: In Anabatic::AutoSegment, in various update methods, updateOrient() must always be called *before* updatePositions() as extensions are dependant on source/target. * New: In Anabatic::AutoSegment::getEndAxes() compute the position of the first source and last target position (center/axes) on an *aligned* set of segments. * New: In Anabatic::AutoSegment, add a new state flag SegAxisFixed to signal segments that can be put on only one track. Specific case to VH gauge for a M1 vertical terminal with a M2 vertical segment. The M2 is effectively bound to the M1 axis position. * Bug: In Anabatic::NetBuilderVH::_do_xG_xM1_xM3(), in case of E/W global and only one RoutingPad the connexion to the RoutingPad was duplicated. It was valid, but totally stupid. * Bug: In Anabatic::Session::_canonize(), for an aligned segment set, intersect the user constraints from all segments instead of only considering the canonical one. Issue a warning about too tight constraints only for symbolic gauges. It may be correct for the real ones. * New: In Katata::DataNegociate::update(), more accurate computation of the perpandicular free interval. Use segment extension cap calculation. Create a special case for fixed axis segments allowing them to find alternative free interval, try under source and under target as they are likely to be draggable segments. * Change: In Katana::Manipulator::relax(), use the extension cap value to compute the axis of the perpandicular segemnts. * Change: In Katana::Manipulator::moveUp(), now move up the whole set of aligned segments instead of just the canonical one. * Change: In Katana::NegociateWindow::loadRoutingPads(), more accurate TrackMarkers insertions for fixed terminals. * New: In Katana::RoutingEvent::Key::Compare::operator(), segments with fixed axis are processed prior to any others. * New: In Katana::RoutingEventLoop, store segment pointers instead of ids to generate more accurate error messages. * Change: In Katana::RoutingPlane::create(), perform local track assignment only for HV gauges. * Change: In Katana::SegmentFsm::_slackenLocal(), add a "dragMinimize" step in the automaton. Mutliple states transitions can occurs in a row if an action fails. * New: In Katana::Session::_toIntervalAxis(), normalize interval bounds so they are on track positions (by shrinking the interval). * Bug: In Katana::TrackMarker CTOR, the weigh computation was wrong.
2018-02-17 13:27:38 -06:00
bool AutoSegment::_initialized = false;
vector< array<DbU::Unit*,3> > AutoSegment::_extensionCaps;
Added analog type on segment NetRoutingProperty. * New: In Anabatic_AutoSegments collection, added a Flag to the constructors to allow different behavior between digital and analog modes. For "Aligneds" and "Perpandiculars" collections, now manage a new Flag WithDoglegs to follow aligned globals through local doglegs (for analog nets). Adjust the log level of collections to 144 (formerly 145). * New: In Anabatic::AutoSegment, new flag SegAnalog for segments that are part of an analog net. Note that with this flag, we reach the 32 bits limit... * Change: In Anabatic::Constants, Flags are now declared as BaseFlags objects and *not* uint64_t. This avoids overload resolution problems with arithmetical overload of the operators. The BaseFlags/Flags types are now completly "isolated" from the uint64_t, it has the advantage of showing where unwanted previous implicit conversions where occuring. * Change: In Katana::Constants, Flags values are now of BaseFlags type instead of uint64_t. * Change: In Anabatic::Dijkstra, lots of log cleanup. * Change: In Anabatic::GCell::getSide(), make the "shrink" parameter visible to allow to substract the topmost and rightmost track for axis span computation in AutoSegment::computeOptimal(). Used for analog mode. * Change: In NetRoutingState, added a flag for analog mode. Use uint32_t for the flags type. * New: In Isobar, export the NetRoutingState and NetRoutingExtension objects.
2017-05-20 05:33:12 -05:00
Improved handling of short nets (fully included in one GCell). The short net mode degrade the routing in some cases. This will be fixed in a next batch of commits. * New: In Hurricane::NetRoutingProperty, added "ShortNet" flag for Nets that are completly inside *one* GCell. * Bug: In CRL::BlifParser::Model::staticInit(), when looking for the output of zero and one cell, also skip the blockage net (as well as automatic and supplies). * New: In Anabatic::AutoSegment, added "ShortNet" flag to know if the segment is part of a short net (fully included in *one* GCell). Also add accessor/mutators for the _analogMode flag (was it ever used before?). * New: In Anabatic::NetBuilder::singleGCell(), if a RoutingPad is vertically small, add a vertical segment to give it some slack. * New: In Anabatic::Dijkstra::_materialize(), detect "short net" as they have only one GCell in their source list... * Bug: In AnabaticEngine::_loadGrbyNet(), reset the AutoSegment "short net" and "analog mode" creation flags between two different nets. * New: In Katana::Configuration, added dedicated ripup for short net segmnts. * New: In Katana: partially implemented support for "short dogleg", that is dogleg that are always kept in same metal because they connect neighboring perpandicular tracks. Not finished neither activated yet. * New: In Katana::TreckElement and derived, export the the *short net* support from AutoSegment. * Bug: In Katana::RoutingEvent::_processRepair(), when a segment is successfully inserted, re-process any perpandicular that is in repair state, as it may have a new chance to be placed. * New: In Katana::SegmentFsm::slackenTopology(), always reject short nets. * Bug: In Katana::Track::check(), correctly handle wide segments instead of issuing false check messages.
2018-07-16 04:16:51 -05:00
void AutoSegment::setAnalogMode ( bool state ) { _analogMode = state; }
bool AutoSegment::getAnalogMode () { return _analogMode; }
void AutoSegment::setShortNetMode ( bool state ) { _shortNetMode = state; }
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
Basic support for FreePDK 45 completed. * New: In Commons, inspector support for std::pair<T,U>. * New: In Hurricane::Layer, ContactLayer & ViaLayer, support for non square VIAs. The hole (cut) remains square, but the various metal extensions can now be different in X and Y. The ::getEnclosure() method now takes a flag EnclosureH / EnclosureV. * New: In Hurricane::DbU, inspector support for: std::pair<DbU::Unit,DbU::Unit> std::array<DbU::Unit,3> Must be defined here as DbU do not exists yet in Commons.h * Bug: In Hurricane::Interval::getSize(), when the interval is "full span", do not return the difference between min and max, but directly DbU::Max. (the previous result was -1 !) * New: In CRL Core Python/Technology.py, support for non square VIAs in the configuration files. Applied to FreePDK 45. * New: In CRL::RoutingGauge, added a "symbolic" flag to tell if a gauge is for symbolic layout or not. Exported to Python. * New: In Anabatic::AutoHorizontal::updatePosition(), differentiated computation for soure or target taking account of the VIA extension in the right segment metal (due to non-square VIAs). * Change: In Anabatic::AutoHorizontal::_makeDogleg(), the dogleg is UP for HV gauges and DOWN for VH. * New: In Anabatic::AutoSegment::_initialize(), create a cache of the various extension length for each layer (viaToTop, viaToBottom, viaToSame). New implementation of getExtensionCap() using the previous cached extension table. See updatePositions(). New static functions to access the extension cache in the header: getViaTotopCap() ... * Change: In Anabatic::AutoSegment, in various update methods, updateOrient() must always be called *before* updatePositions() as extensions are dependant on source/target. * New: In Anabatic::AutoSegment::getEndAxes() compute the position of the first source and last target position (center/axes) on an *aligned* set of segments. * New: In Anabatic::AutoSegment, add a new state flag SegAxisFixed to signal segments that can be put on only one track. Specific case to VH gauge for a M1 vertical terminal with a M2 vertical segment. The M2 is effectively bound to the M1 axis position. * Bug: In Anabatic::NetBuilderVH::_do_xG_xM1_xM3(), in case of E/W global and only one RoutingPad the connexion to the RoutingPad was duplicated. It was valid, but totally stupid. * Bug: In Anabatic::Session::_canonize(), for an aligned segment set, intersect the user constraints from all segments instead of only considering the canonical one. Issue a warning about too tight constraints only for symbolic gauges. It may be correct for the real ones. * New: In Katata::DataNegociate::update(), more accurate computation of the perpandicular free interval. Use segment extension cap calculation. Create a special case for fixed axis segments allowing them to find alternative free interval, try under source and under target as they are likely to be draggable segments. * Change: In Katana::Manipulator::relax(), use the extension cap value to compute the axis of the perpandicular segemnts. * Change: In Katana::Manipulator::moveUp(), now move up the whole set of aligned segments instead of just the canonical one. * Change: In Katana::NegociateWindow::loadRoutingPads(), more accurate TrackMarkers insertions for fixed terminals. * New: In Katana::RoutingEvent::Key::Compare::operator(), segments with fixed axis are processed prior to any others. * New: In Katana::RoutingEventLoop, store segment pointers instead of ids to generate more accurate error messages. * Change: In Katana::RoutingPlane::create(), perform local track assignment only for HV gauges. * Change: In Katana::SegmentFsm::_slackenLocal(), add a "dragMinimize" step in the automaton. Mutliple states transitions can occurs in a row if an action fails. * New: In Katana::Session::_toIntervalAxis(), normalize interval bounds so they are on track positions (by shrinking the interval). * Bug: In Katana::TrackMarker CTOR, the weigh computation was wrong.
2018-02-17 13:27:38 -06:00
void AutoSegment::_initialize ()
{
Capacity managment by layer in Edges & GCells (plus fixes). * New: In Anabatic::EdgeCapacity, dedicated object to manage the capacities of an edge by layer. This needed now because with real technologies layers capacities differs (unlike with symbolic technologies). This object is separated to be shared between Edges with identical characteristics (direction+interval). Deletion is automatic and done through refcounting. All the already allocateds EdgeCapacity are kept into a set in the AnabaticEngine (key is (direction,interval)). * New: In Anabatic::Edge, capacities are stored in a shared EdgeCapacity object. The total capacity can be annotated (i.e. decreased). EdgeCapacity attribute is created during the materialize() call. The capacities are computed at this time. The incCapacity() function is renamed in reserveCapacity(). * New: In Anabatic::AnabaticEngine, added attribute _edgeCapacitiesLut to store the shared EdgeCapacity. Lookup/Creation of an EdgeCapacity is done through _createCapacity(). * Change: In Anabatic::Constants, rename Flags::IllimitedCapacity into Flags::InfiniteCapacity. Add Flags::NullCapacity (both for Edges). * Change: In Anabatic::GCell, implement a by depth (for layer) getCapacity(). This modification did expose a bug in the density calculation : per depth density where divided by the complete density instead of the density's depth. This was leading to greatly underestimated densities. Thoses underestimations where preventing Dijkstra and layer assignement to manage congestion correctly (in fact, it was acting as if there never was congestion). Also avoid a divide by zero (thus -NAN showing in densities). * Change: In Anabatic::GCell, rename setEdgeOccupancy() into the more accurate forceEdgesCapacities(). Note for Eric: only the first Edge on each side has it's capacity forced. What if there's more than one Edge ?
2018-02-20 17:16:50 -06:00
//cerr << "AutoSegment::_initialize()" << endl;
Basic support for FreePDK 45 completed. * New: In Commons, inspector support for std::pair<T,U>. * New: In Hurricane::Layer, ContactLayer & ViaLayer, support for non square VIAs. The hole (cut) remains square, but the various metal extensions can now be different in X and Y. The ::getEnclosure() method now takes a flag EnclosureH / EnclosureV. * New: In Hurricane::DbU, inspector support for: std::pair<DbU::Unit,DbU::Unit> std::array<DbU::Unit,3> Must be defined here as DbU do not exists yet in Commons.h * Bug: In Hurricane::Interval::getSize(), when the interval is "full span", do not return the difference between min and max, but directly DbU::Max. (the previous result was -1 !) * New: In CRL Core Python/Technology.py, support for non square VIAs in the configuration files. Applied to FreePDK 45. * New: In CRL::RoutingGauge, added a "symbolic" flag to tell if a gauge is for symbolic layout or not. Exported to Python. * New: In Anabatic::AutoHorizontal::updatePosition(), differentiated computation for soure or target taking account of the VIA extension in the right segment metal (due to non-square VIAs). * Change: In Anabatic::AutoHorizontal::_makeDogleg(), the dogleg is UP for HV gauges and DOWN for VH. * New: In Anabatic::AutoSegment::_initialize(), create a cache of the various extension length for each layer (viaToTop, viaToBottom, viaToSame). New implementation of getExtensionCap() using the previous cached extension table. See updatePositions(). New static functions to access the extension cache in the header: getViaTotopCap() ... * Change: In Anabatic::AutoSegment, in various update methods, updateOrient() must always be called *before* updatePositions() as extensions are dependant on source/target. * New: In Anabatic::AutoSegment::getEndAxes() compute the position of the first source and last target position (center/axes) on an *aligned* set of segments. * New: In Anabatic::AutoSegment, add a new state flag SegAxisFixed to signal segments that can be put on only one track. Specific case to VH gauge for a M1 vertical terminal with a M2 vertical segment. The M2 is effectively bound to the M1 axis position. * Bug: In Anabatic::NetBuilderVH::_do_xG_xM1_xM3(), in case of E/W global and only one RoutingPad the connexion to the RoutingPad was duplicated. It was valid, but totally stupid. * Bug: In Anabatic::Session::_canonize(), for an aligned segment set, intersect the user constraints from all segments instead of only considering the canonical one. Issue a warning about too tight constraints only for symbolic gauges. It may be correct for the real ones. * New: In Katata::DataNegociate::update(), more accurate computation of the perpandicular free interval. Use segment extension cap calculation. Create a special case for fixed axis segments allowing them to find alternative free interval, try under source and under target as they are likely to be draggable segments. * Change: In Katana::Manipulator::relax(), use the extension cap value to compute the axis of the perpandicular segemnts. * Change: In Katana::Manipulator::moveUp(), now move up the whole set of aligned segments instead of just the canonical one. * Change: In Katana::NegociateWindow::loadRoutingPads(), more accurate TrackMarkers insertions for fixed terminals. * New: In Katana::RoutingEvent::Key::Compare::operator(), segments with fixed axis are processed prior to any others. * New: In Katana::RoutingEventLoop, store segment pointers instead of ids to generate more accurate error messages. * Change: In Katana::RoutingPlane::create(), perform local track assignment only for HV gauges. * Change: In Katana::SegmentFsm::_slackenLocal(), add a "dragMinimize" step in the automaton. Mutliple states transitions can occurs in a row if an action fails. * New: In Katana::Session::_toIntervalAxis(), normalize interval bounds so they are on track positions (by shrinking the interval). * Bug: In Katana::TrackMarker CTOR, the weigh computation was wrong.
2018-02-17 13:27:38 -06:00
_initialized = true;
for ( size_t depth=0 ; depth<Session::getDepth() ; ++depth ) {
DbU::Unit* viaToTopCap = new DbU::Unit ( 0 );
DbU::Unit* viaToBottomCap = new DbU::Unit ( 0 );
DbU::Unit* viaToSameCap = new DbU::Unit ( 0 );
bool isVertical = (depth == 0) or (Session::getLayerGauge(depth)->isVertical());
uint32_t flags = (isVertical) ? Layer::EnclosureV : Layer::EnclosureH ;
Capacity managment by layer in Edges & GCells (plus fixes). * New: In Anabatic::EdgeCapacity, dedicated object to manage the capacities of an edge by layer. This needed now because with real technologies layers capacities differs (unlike with symbolic technologies). This object is separated to be shared between Edges with identical characteristics (direction+interval). Deletion is automatic and done through refcounting. All the already allocateds EdgeCapacity are kept into a set in the AnabaticEngine (key is (direction,interval)). * New: In Anabatic::Edge, capacities are stored in a shared EdgeCapacity object. The total capacity can be annotated (i.e. decreased). EdgeCapacity attribute is created during the materialize() call. The capacities are computed at this time. The incCapacity() function is renamed in reserveCapacity(). * New: In Anabatic::AnabaticEngine, added attribute _edgeCapacitiesLut to store the shared EdgeCapacity. Lookup/Creation of an EdgeCapacity is done through _createCapacity(). * Change: In Anabatic::Constants, rename Flags::IllimitedCapacity into Flags::InfiniteCapacity. Add Flags::NullCapacity (both for Edges). * Change: In Anabatic::GCell, implement a by depth (for layer) getCapacity(). This modification did expose a bug in the density calculation : per depth density where divided by the complete density instead of the density's depth. This was leading to greatly underestimated densities. Thoses underestimations where preventing Dijkstra and layer assignement to manage congestion correctly (in fact, it was acting as if there never was congestion). Also avoid a divide by zero (thus -NAN showing in densities). * Change: In Anabatic::GCell, rename setEdgeOccupancy() into the more accurate forceEdgesCapacities(). Note for Eric: only the first Edge on each side has it's capacity forced. What if there's more than one Edge ?
2018-02-20 17:16:50 -06:00
//cerr << depth << ":" << Session::getLayerGauge(depth)->getLayer()->getName()
// << " isVertical:" << Session::getLayerGauge(depth)->isVertical() << endl;
Basic support for FreePDK 45 completed. * New: In Commons, inspector support for std::pair<T,U>. * New: In Hurricane::Layer, ContactLayer & ViaLayer, support for non square VIAs. The hole (cut) remains square, but the various metal extensions can now be different in X and Y. The ::getEnclosure() method now takes a flag EnclosureH / EnclosureV. * New: In Hurricane::DbU, inspector support for: std::pair<DbU::Unit,DbU::Unit> std::array<DbU::Unit,3> Must be defined here as DbU do not exists yet in Commons.h * Bug: In Hurricane::Interval::getSize(), when the interval is "full span", do not return the difference between min and max, but directly DbU::Max. (the previous result was -1 !) * New: In CRL Core Python/Technology.py, support for non square VIAs in the configuration files. Applied to FreePDK 45. * New: In CRL::RoutingGauge, added a "symbolic" flag to tell if a gauge is for symbolic layout or not. Exported to Python. * New: In Anabatic::AutoHorizontal::updatePosition(), differentiated computation for soure or target taking account of the VIA extension in the right segment metal (due to non-square VIAs). * Change: In Anabatic::AutoHorizontal::_makeDogleg(), the dogleg is UP for HV gauges and DOWN for VH. * New: In Anabatic::AutoSegment::_initialize(), create a cache of the various extension length for each layer (viaToTop, viaToBottom, viaToSame). New implementation of getExtensionCap() using the previous cached extension table. See updatePositions(). New static functions to access the extension cache in the header: getViaTotopCap() ... * Change: In Anabatic::AutoSegment, in various update methods, updateOrient() must always be called *before* updatePositions() as extensions are dependant on source/target. * New: In Anabatic::AutoSegment::getEndAxes() compute the position of the first source and last target position (center/axes) on an *aligned* set of segments. * New: In Anabatic::AutoSegment, add a new state flag SegAxisFixed to signal segments that can be put on only one track. Specific case to VH gauge for a M1 vertical terminal with a M2 vertical segment. The M2 is effectively bound to the M1 axis position. * Bug: In Anabatic::NetBuilderVH::_do_xG_xM1_xM3(), in case of E/W global and only one RoutingPad the connexion to the RoutingPad was duplicated. It was valid, but totally stupid. * Bug: In Anabatic::Session::_canonize(), for an aligned segment set, intersect the user constraints from all segments instead of only considering the canonical one. Issue a warning about too tight constraints only for symbolic gauges. It may be correct for the real ones. * New: In Katata::DataNegociate::update(), more accurate computation of the perpandicular free interval. Use segment extension cap calculation. Create a special case for fixed axis segments allowing them to find alternative free interval, try under source and under target as they are likely to be draggable segments. * Change: In Katana::Manipulator::relax(), use the extension cap value to compute the axis of the perpandicular segemnts. * Change: In Katana::Manipulator::moveUp(), now move up the whole set of aligned segments instead of just the canonical one. * Change: In Katana::NegociateWindow::loadRoutingPads(), more accurate TrackMarkers insertions for fixed terminals. * New: In Katana::RoutingEvent::Key::Compare::operator(), segments with fixed axis are processed prior to any others. * New: In Katana::RoutingEventLoop, store segment pointers instead of ids to generate more accurate error messages. * Change: In Katana::RoutingPlane::create(), perform local track assignment only for HV gauges. * Change: In Katana::SegmentFsm::_slackenLocal(), add a "dragMinimize" step in the automaton. Mutliple states transitions can occurs in a row if an action fails. * New: In Katana::Session::_toIntervalAxis(), normalize interval bounds so they are on track positions (by shrinking the interval). * Bug: In Katana::TrackMarker CTOR, the weigh computation was wrong.
2018-02-17 13:27:38 -06:00
DRC correct on Arlet6505 / TSMC C180. Integrate new features and bug fixes so the Arlet 6502 benchs successfully passes real DRC with reference industrial tools. Short summary: * Manage minimum area for VIAs in Katana::Tracks. * Allow different wire width for wires perpandicular to the prefered routing direction. * StackedVIAs used in the clock tree no longer assume an uniform routing grid (same offset & pitch all the way up). * Some hard-coded patches in PowerRails for FlexLib. * New: In CRL/symbolic/cmos/kite.py & cmos45/kite.py, update the RoutingLayerGauges by adding the new PWireWidth parameter. Always zero in case of symbolic layout (too fine tuning). * New: In CRL::RoutingGauge, add accessor to PWireWidth parameter. Modify the clone method. * New: In CRL::RoutingLayerGauge, add new parameter "PWireWidth" to give the width of a wire when it not drawn in the prefered routing direction. If it is set to zero, the normal width is used. * New: In CRL::PyRoutingGauge, export the updated constructor interface. It is *not* backward compatible, one must add the PWireWidth parameter in the various kite.py configuration files (in etc/). * Change: In AnabaticEngine::_gutAnabatic(), disable the minimum area detection mechanism, replaced by a more complete one in Katana::Track. Left commented out for now, but will be removed in the future. * Change: In Anabatic::AutoContact::updateLayer(), now systematically calls setLayerAndWidth() to potentially resize the VIAs. This is needed in real mode as VIAs are *not* macro-generated but have their real final size. * Change: In Anabatic::AutoContact::setLayerAndWidth(), select the width and height of the contact using the gauge wire width *and* perpandicular *wire width*. * Change: In Anabatic::AutoSegment::_initialize(), the "VIA to same cap" to PWireWidth/2, this will be the size of the VIA in the non-preferred direction at the end cap (non-square in real mode). * Change: In Anabatic::AutoSegment::getExtensionCap(), makes different cases for symbolic and real. Use raw length in real, add half the wire width in symbolic. Add a flag to get the extension cap *only*, not increased of half the minimal spacing. * Change: In Anabatic::AutoSegment::bloatStackedStrap(), enhanced, but finally unused... * New: In Anabatic::AutoSegment::create(), use the PWireWidth when the segment is not in the preferred routing direction (and of minimal width). * New: In Anabatic::Configuration, add new getPWirewidth(), DPHorizontalWidth() and DPVerticalWidth() accessors. * Change: In AnabaticEngine::setupPreRouteds(), skip components in in "cut" material. We are only interested in objects containing some metal (happens in real mode when VIAs cuts are really there). * New: In Katana::PowerRailsPlanes::Rail::doLayout(), add an hard-coded patch that artificially enlarge the *wide wire* so the spacing for wide wire is enforced. For now, two pitches on each side for "FlexLib" gauge. * New: In Katana::Track, add support to find and correct small wire chunks so they respect the minimum area rules. Two helper functions: * ::hasSameLayerTurn(), to find if a a TrackElement as non-zero length perpandicular is same layer connected to it. * ::toFoundryGrid(), to ensure that all coordinates will be on the foundry grid (may move in a more shared location). * ::expandToMinArea(), try to expand, *in the routing direction* the too small wire so it respect the minimal area. Check for the free space in the track. Track::minExpandArea() go through all the TrackElements in the track to look for too small ones and correct them. * Change: In Katana::RoutingPlane, add an accessor to get the tracks. * New: In KatanaEngine::finalizeLayout(), add a post-treatment to find for minimal area violations. * Change: In cumulus/plugins.block.configuration.GaugeConf, add a routingBb attribute that will serve as a common reference to all the functions calculation track positions. We must not have two different reference for the core and the corona. The reference is always the corona when we working on a complete chip. * New: In cumulus/plugins.block.configuration.GaugeConf.getTrack(), Simplified and more reliable way of getting tracks positions. Use the routingBb. * New: In cumulus/plugins.block.configuration.GaugeConf.rpAccess(), Make use of getTrack() to get every metal strap on the right X/Y position. * New: In cumulus/plugins.block.configuration.GaugeConf.expandMinArea(), As those wires are left alone by the router, it is our responsability to abide by the minimal area rule here. Hence the code duplication from the router (bad). Mainly wires made for the clock tree, I mean. * Bug: In cumulus/plugins.chip.configuration.ChipConf.setupICore(), the core instance must be placed on the GCell grid, defined by the slice height (X *and* Y). * Bug: In cumulus/plugins.chip.corona.Builder(), forgot to use bigvia for the corners of the inner ring. * Bug: In cumulus/plugins.chip.pads.corona._createCoreWire(), hard-coded patch for LibreSOCIO, the power/ground connectors toward the core are too wide and can create DRC errors when put side by side. Shrink them by the minimal distance.
2020-11-23 16:07:15 -06:00
*viaToSameCap = Session::getPWireWidth(depth)/2;
Basic support for FreePDK 45 completed. * New: In Commons, inspector support for std::pair<T,U>. * New: In Hurricane::Layer, ContactLayer & ViaLayer, support for non square VIAs. The hole (cut) remains square, but the various metal extensions can now be different in X and Y. The ::getEnclosure() method now takes a flag EnclosureH / EnclosureV. * New: In Hurricane::DbU, inspector support for: std::pair<DbU::Unit,DbU::Unit> std::array<DbU::Unit,3> Must be defined here as DbU do not exists yet in Commons.h * Bug: In Hurricane::Interval::getSize(), when the interval is "full span", do not return the difference between min and max, but directly DbU::Max. (the previous result was -1 !) * New: In CRL Core Python/Technology.py, support for non square VIAs in the configuration files. Applied to FreePDK 45. * New: In CRL::RoutingGauge, added a "symbolic" flag to tell if a gauge is for symbolic layout or not. Exported to Python. * New: In Anabatic::AutoHorizontal::updatePosition(), differentiated computation for soure or target taking account of the VIA extension in the right segment metal (due to non-square VIAs). * Change: In Anabatic::AutoHorizontal::_makeDogleg(), the dogleg is UP for HV gauges and DOWN for VH. * New: In Anabatic::AutoSegment::_initialize(), create a cache of the various extension length for each layer (viaToTop, viaToBottom, viaToSame). New implementation of getExtensionCap() using the previous cached extension table. See updatePositions(). New static functions to access the extension cache in the header: getViaTotopCap() ... * Change: In Anabatic::AutoSegment, in various update methods, updateOrient() must always be called *before* updatePositions() as extensions are dependant on source/target. * New: In Anabatic::AutoSegment::getEndAxes() compute the position of the first source and last target position (center/axes) on an *aligned* set of segments. * New: In Anabatic::AutoSegment, add a new state flag SegAxisFixed to signal segments that can be put on only one track. Specific case to VH gauge for a M1 vertical terminal with a M2 vertical segment. The M2 is effectively bound to the M1 axis position. * Bug: In Anabatic::NetBuilderVH::_do_xG_xM1_xM3(), in case of E/W global and only one RoutingPad the connexion to the RoutingPad was duplicated. It was valid, but totally stupid. * Bug: In Anabatic::Session::_canonize(), for an aligned segment set, intersect the user constraints from all segments instead of only considering the canonical one. Issue a warning about too tight constraints only for symbolic gauges. It may be correct for the real ones. * New: In Katata::DataNegociate::update(), more accurate computation of the perpandicular free interval. Use segment extension cap calculation. Create a special case for fixed axis segments allowing them to find alternative free interval, try under source and under target as they are likely to be draggable segments. * Change: In Katana::Manipulator::relax(), use the extension cap value to compute the axis of the perpandicular segemnts. * Change: In Katana::Manipulator::moveUp(), now move up the whole set of aligned segments instead of just the canonical one. * Change: In Katana::NegociateWindow::loadRoutingPads(), more accurate TrackMarkers insertions for fixed terminals. * New: In Katana::RoutingEvent::Key::Compare::operator(), segments with fixed axis are processed prior to any others. * New: In Katana::RoutingEventLoop, store segment pointers instead of ids to generate more accurate error messages. * Change: In Katana::RoutingPlane::create(), perform local track assignment only for HV gauges. * Change: In Katana::SegmentFsm::_slackenLocal(), add a "dragMinimize" step in the automaton. Mutliple states transitions can occurs in a row if an action fails. * New: In Katana::Session::_toIntervalAxis(), normalize interval bounds so they are on track positions (by shrinking the interval). * Bug: In Katana::TrackMarker CTOR, the weigh computation was wrong.
2018-02-17 13:27:38 -06:00
// Bottom metal of the VIA going *up*.
const Layer* viaLayer = dynamic_cast<const ViaLayer*>( Session::getContactLayer(depth) );
if (viaLayer)
*viaToTopCap = Session::getViaWidth(depth)/2 + viaLayer->getBottomEnclosure( flags );
// Top metal of the VIA going *down*.
if (depth > 0) {
viaLayer = dynamic_cast<const ViaLayer*>( Session::getContactLayer(depth-1) );
if (viaLayer)
*viaToBottomCap = Session::getViaWidth(depth-1)/2 + viaLayer->getTopEnclosure( flags );
}
Capacity managment by layer in Edges & GCells (plus fixes). * New: In Anabatic::EdgeCapacity, dedicated object to manage the capacities of an edge by layer. This needed now because with real technologies layers capacities differs (unlike with symbolic technologies). This object is separated to be shared between Edges with identical characteristics (direction+interval). Deletion is automatic and done through refcounting. All the already allocateds EdgeCapacity are kept into a set in the AnabaticEngine (key is (direction,interval)). * New: In Anabatic::Edge, capacities are stored in a shared EdgeCapacity object. The total capacity can be annotated (i.e. decreased). EdgeCapacity attribute is created during the materialize() call. The capacities are computed at this time. The incCapacity() function is renamed in reserveCapacity(). * New: In Anabatic::AnabaticEngine, added attribute _edgeCapacitiesLut to store the shared EdgeCapacity. Lookup/Creation of an EdgeCapacity is done through _createCapacity(). * Change: In Anabatic::Constants, rename Flags::IllimitedCapacity into Flags::InfiniteCapacity. Add Flags::NullCapacity (both for Edges). * Change: In Anabatic::GCell, implement a by depth (for layer) getCapacity(). This modification did expose a bug in the density calculation : per depth density where divided by the complete density instead of the density's depth. This was leading to greatly underestimated densities. Thoses underestimations where preventing Dijkstra and layer assignement to manage congestion correctly (in fact, it was acting as if there never was congestion). Also avoid a divide by zero (thus -NAN showing in densities). * Change: In Anabatic::GCell, rename setEdgeOccupancy() into the more accurate forceEdgesCapacities(). Note for Eric: only the first Edge on each side has it's capacity forced. What if there's more than one Edge ?
2018-02-20 17:16:50 -06:00
//cerr << " viaToTop width: " << DbU::getValueString( Session::getViaWidth(depth) ) << endl;
//cerr << " viaToTopCap: " << DbU::getValueString(*viaToTopCap ) << endl;
//if (depth > 0)
// cerr << " viaToBottom width:" << DbU::getValueString( Session::getViaWidth(depth-1)/2 ) << endl;
//cerr << " viaToBottomCap: " << DbU::getValueString(*viaToBottomCap) << endl;
//cerr << " viaToSameCap: " << DbU::getValueString(*viaToSameCap ) << endl;
Basic support for FreePDK 45 completed. * New: In Commons, inspector support for std::pair<T,U>. * New: In Hurricane::Layer, ContactLayer & ViaLayer, support for non square VIAs. The hole (cut) remains square, but the various metal extensions can now be different in X and Y. The ::getEnclosure() method now takes a flag EnclosureH / EnclosureV. * New: In Hurricane::DbU, inspector support for: std::pair<DbU::Unit,DbU::Unit> std::array<DbU::Unit,3> Must be defined here as DbU do not exists yet in Commons.h * Bug: In Hurricane::Interval::getSize(), when the interval is "full span", do not return the difference between min and max, but directly DbU::Max. (the previous result was -1 !) * New: In CRL Core Python/Technology.py, support for non square VIAs in the configuration files. Applied to FreePDK 45. * New: In CRL::RoutingGauge, added a "symbolic" flag to tell if a gauge is for symbolic layout or not. Exported to Python. * New: In Anabatic::AutoHorizontal::updatePosition(), differentiated computation for soure or target taking account of the VIA extension in the right segment metal (due to non-square VIAs). * Change: In Anabatic::AutoHorizontal::_makeDogleg(), the dogleg is UP for HV gauges and DOWN for VH. * New: In Anabatic::AutoSegment::_initialize(), create a cache of the various extension length for each layer (viaToTop, viaToBottom, viaToSame). New implementation of getExtensionCap() using the previous cached extension table. See updatePositions(). New static functions to access the extension cache in the header: getViaTotopCap() ... * Change: In Anabatic::AutoSegment, in various update methods, updateOrient() must always be called *before* updatePositions() as extensions are dependant on source/target. * New: In Anabatic::AutoSegment::getEndAxes() compute the position of the first source and last target position (center/axes) on an *aligned* set of segments. * New: In Anabatic::AutoSegment, add a new state flag SegAxisFixed to signal segments that can be put on only one track. Specific case to VH gauge for a M1 vertical terminal with a M2 vertical segment. The M2 is effectively bound to the M1 axis position. * Bug: In Anabatic::NetBuilderVH::_do_xG_xM1_xM3(), in case of E/W global and only one RoutingPad the connexion to the RoutingPad was duplicated. It was valid, but totally stupid. * Bug: In Anabatic::Session::_canonize(), for an aligned segment set, intersect the user constraints from all segments instead of only considering the canonical one. Issue a warning about too tight constraints only for symbolic gauges. It may be correct for the real ones. * New: In Katata::DataNegociate::update(), more accurate computation of the perpandicular free interval. Use segment extension cap calculation. Create a special case for fixed axis segments allowing them to find alternative free interval, try under source and under target as they are likely to be draggable segments. * Change: In Katana::Manipulator::relax(), use the extension cap value to compute the axis of the perpandicular segemnts. * Change: In Katana::Manipulator::moveUp(), now move up the whole set of aligned segments instead of just the canonical one. * Change: In Katana::NegociateWindow::loadRoutingPads(), more accurate TrackMarkers insertions for fixed terminals. * New: In Katana::RoutingEvent::Key::Compare::operator(), segments with fixed axis are processed prior to any others. * New: In Katana::RoutingEventLoop, store segment pointers instead of ids to generate more accurate error messages. * Change: In Katana::RoutingPlane::create(), perform local track assignment only for HV gauges. * Change: In Katana::SegmentFsm::_slackenLocal(), add a "dragMinimize" step in the automaton. Mutliple states transitions can occurs in a row if an action fails. * New: In Katana::Session::_toIntervalAxis(), normalize interval bounds so they are on track positions (by shrinking the interval). * Bug: In Katana::TrackMarker CTOR, the weigh computation was wrong.
2018-02-17 13:27:38 -06:00
Groudwork for routing density driven placement. Compliance with clang 5.0.1. This commit contains two set of features that should have been commited separately. 1. Compliance with clang 5.0.1, tested with the RedHat collection llvm-toolset-7. This allow Coriolis to be compiled under Darwin (MacOS) with Xcode & macports. The bootstrap install system has been modificated accordingly. 2. The basic support for routing density driven placement. Related features are: * Bloat property. Each Occurrence of an Instance can be individually bloated. This property not attached to any tool to allow the placer and router to share it as wanted. Nevertheless, it is defined in Etesian. * BloatProfile in Katana, add individual Bloat properties to Instances occurrences based on the East & North overflowed edges of each GCell. * Support in ToolEngine for a "pass number" of a tool. This pass number is mainly used to make "per pass" measurements. The MeasureSet system is improved accordingly to support multiple values of a same measure. * Embryo of "P&R Conductor" to perform the place & route loop until the design is successfully placed. May be the first brick of a Silicon Compiler. * Change: In boostrap/FindBoostrap.cmake, in setup_boost(), added tag to the python component for macport (ex: python27). * Change: In boostrap/build.conf, put etesian before anabatic for instance occurrence BloatProperty dependency. Added option support for the "llvm-toolset-7" collection to build against clang 5.0.1. * Bug: In Hurricane::getRecord( const pair<T,U>& ), the getSlot<> templates for first & second arguments must be called with <const T> and <const U> as the pair itself is const (and not simply <T> & <U>). * Change: In Hurricane::getSlot() temlate, only use "string" arguments and not const string&, simpler for template argument deduction. * Bug: In Hurricane::AnalogCellExtension, the StandardPrivateProperty<> template has a static member "_name". Clang did show that the template for this static number has to be put inside the namespace where the template *is defined* (i.e. Hurricane) instead of the namespace where it is instanciated (i.e. Analog). * Bug: In Isobar, Matrix_FromListOfList(), PyInt_AsPlacementStatus() must be put outside the C linkage back in the Isobar C++ namespace (clang). * Bug: In Hurricane::DBo::~DBo, and derived add a throw() specification (clang). * Bug: In Hurricane::RegularLayer::getEnclosure() & setEnclosure(), change signature so it matches the one of the base class (clang). * Bug: In Hurricane::CellPrinter, use double brackets for initializer list (clang). * Change: In Hurricane::Breakpoint, reverse the meaning of the error level. Only error level *lesser or equal* than the stop level will be enabled. * Bug: In CRL/python/helpers/__init__.loadUserSettings(), must put the current working directory in the sys.path as in certain configuration it may not be included. * Bug: In CRL::ApDriver, DumpSegments(), no longer generate segments when encountering a RoutingPad on a top-level Pin Occurrence. The segment was generated in the wrong direction, creating DRC violations on the "mips_core_flat" example. * Change: In CRL::Measures, partial re-design of the measurements management. Now, each kind of measure can accept multiple values put in a vector. The index is intented to match a tool run number. * Change: In CRL::Histogram, add support for multiple sets of datas, indexeds with tool run number. * Change: In CRL::ToolEngine, add support for multiple pass number, add addMeasure<> templates for the various data-types. * Change: In CRL::gdsDriver & CRL::gdsParser(), comment out unused GDS record name constants. * New: Etesian::BloatProperty, property to attach to Instance occurrences that contains the extra number of pitch to add to the cell width. * Bug: In AutoSegment::CompareByDepthLength, the segment length comparison was wrong, it was always returning true, which broke the "strick weak ordering" of the comparison. This was producing a core-dump in GCell::updateDensity() when sorting a vector<>. The end() iterator was being dereferenced, leading to the problem. * Bug: In Katana::DataSymmetric::checkPairing(), the test for segments whose axis is perpandicular to the symmetry axis was wrong ("!=" instead of "-"). * New: In Katana/GlobalRoute, new ::selectSegments(), selectOverloadedgcells() and selectBloatedInstances() to automatically select segments from overloaded edges, overloaded GCells and bloated cells. * Change: In KatanaEngine, return a more detailed success state, distinguish between global and detailed. Add support for multiple routing iterations. * New: In cumulus/python/plugins/ConductorPlugin.py, embryo of routing driven placement.
2019-12-08 18:57:44 -06:00
_extensionCaps.push_back( std::array<DbU::Unit*,3>( {{ viaToTopCap, viaToBottomCap, viaToSameCap }} ) );
Basic support for FreePDK 45 completed. * New: In Commons, inspector support for std::pair<T,U>. * New: In Hurricane::Layer, ContactLayer & ViaLayer, support for non square VIAs. The hole (cut) remains square, but the various metal extensions can now be different in X and Y. The ::getEnclosure() method now takes a flag EnclosureH / EnclosureV. * New: In Hurricane::DbU, inspector support for: std::pair<DbU::Unit,DbU::Unit> std::array<DbU::Unit,3> Must be defined here as DbU do not exists yet in Commons.h * Bug: In Hurricane::Interval::getSize(), when the interval is "full span", do not return the difference between min and max, but directly DbU::Max. (the previous result was -1 !) * New: In CRL Core Python/Technology.py, support for non square VIAs in the configuration files. Applied to FreePDK 45. * New: In CRL::RoutingGauge, added a "symbolic" flag to tell if a gauge is for symbolic layout or not. Exported to Python. * New: In Anabatic::AutoHorizontal::updatePosition(), differentiated computation for soure or target taking account of the VIA extension in the right segment metal (due to non-square VIAs). * Change: In Anabatic::AutoHorizontal::_makeDogleg(), the dogleg is UP for HV gauges and DOWN for VH. * New: In Anabatic::AutoSegment::_initialize(), create a cache of the various extension length for each layer (viaToTop, viaToBottom, viaToSame). New implementation of getExtensionCap() using the previous cached extension table. See updatePositions(). New static functions to access the extension cache in the header: getViaTotopCap() ... * Change: In Anabatic::AutoSegment, in various update methods, updateOrient() must always be called *before* updatePositions() as extensions are dependant on source/target. * New: In Anabatic::AutoSegment::getEndAxes() compute the position of the first source and last target position (center/axes) on an *aligned* set of segments. * New: In Anabatic::AutoSegment, add a new state flag SegAxisFixed to signal segments that can be put on only one track. Specific case to VH gauge for a M1 vertical terminal with a M2 vertical segment. The M2 is effectively bound to the M1 axis position. * Bug: In Anabatic::NetBuilderVH::_do_xG_xM1_xM3(), in case of E/W global and only one RoutingPad the connexion to the RoutingPad was duplicated. It was valid, but totally stupid. * Bug: In Anabatic::Session::_canonize(), for an aligned segment set, intersect the user constraints from all segments instead of only considering the canonical one. Issue a warning about too tight constraints only for symbolic gauges. It may be correct for the real ones. * New: In Katata::DataNegociate::update(), more accurate computation of the perpandicular free interval. Use segment extension cap calculation. Create a special case for fixed axis segments allowing them to find alternative free interval, try under source and under target as they are likely to be draggable segments. * Change: In Katana::Manipulator::relax(), use the extension cap value to compute the axis of the perpandicular segemnts. * Change: In Katana::Manipulator::moveUp(), now move up the whole set of aligned segments instead of just the canonical one. * Change: In Katana::NegociateWindow::loadRoutingPads(), more accurate TrackMarkers insertions for fixed terminals. * New: In Katana::RoutingEvent::Key::Compare::operator(), segments with fixed axis are processed prior to any others. * New: In Katana::RoutingEventLoop, store segment pointers instead of ids to generate more accurate error messages. * Change: In Katana::RoutingPlane::create(), perform local track assignment only for HV gauges. * Change: In Katana::SegmentFsm::_slackenLocal(), add a "dragMinimize" step in the automaton. Mutliple states transitions can occurs in a row if an action fails. * New: In Katana::Session::_toIntervalAxis(), normalize interval bounds so they are on track positions (by shrinking the interval). * Bug: In Katana::TrackMarker CTOR, the weigh computation was wrong.
2018-02-17 13:27:38 -06:00
}
}
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
AutoSegment::AutoSegment ( Segment* segment )
: _id (segment->getId())
, _gcell (NULL)
, _flags (SegCreated)
, _depth (Session::getLayerDepth(segment->getLayer()))
, _optimalMin (0)
, _optimalMax (0)
, _reduceds (0)
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
, _rpDistance (15)
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
, _sourcePosition (0)
, _targetPosition (0)
, _userConstraints (false)
, _nativeConstraints(false)
, _parent (NULL)
, _observers ()
{
Basic support for FreePDK 45 completed. * New: In Commons, inspector support for std::pair<T,U>. * New: In Hurricane::Layer, ContactLayer & ViaLayer, support for non square VIAs. The hole (cut) remains square, but the various metal extensions can now be different in X and Y. The ::getEnclosure() method now takes a flag EnclosureH / EnclosureV. * New: In Hurricane::DbU, inspector support for: std::pair<DbU::Unit,DbU::Unit> std::array<DbU::Unit,3> Must be defined here as DbU do not exists yet in Commons.h * Bug: In Hurricane::Interval::getSize(), when the interval is "full span", do not return the difference between min and max, but directly DbU::Max. (the previous result was -1 !) * New: In CRL Core Python/Technology.py, support for non square VIAs in the configuration files. Applied to FreePDK 45. * New: In CRL::RoutingGauge, added a "symbolic" flag to tell if a gauge is for symbolic layout or not. Exported to Python. * New: In Anabatic::AutoHorizontal::updatePosition(), differentiated computation for soure or target taking account of the VIA extension in the right segment metal (due to non-square VIAs). * Change: In Anabatic::AutoHorizontal::_makeDogleg(), the dogleg is UP for HV gauges and DOWN for VH. * New: In Anabatic::AutoSegment::_initialize(), create a cache of the various extension length for each layer (viaToTop, viaToBottom, viaToSame). New implementation of getExtensionCap() using the previous cached extension table. See updatePositions(). New static functions to access the extension cache in the header: getViaTotopCap() ... * Change: In Anabatic::AutoSegment, in various update methods, updateOrient() must always be called *before* updatePositions() as extensions are dependant on source/target. * New: In Anabatic::AutoSegment::getEndAxes() compute the position of the first source and last target position (center/axes) on an *aligned* set of segments. * New: In Anabatic::AutoSegment, add a new state flag SegAxisFixed to signal segments that can be put on only one track. Specific case to VH gauge for a M1 vertical terminal with a M2 vertical segment. The M2 is effectively bound to the M1 axis position. * Bug: In Anabatic::NetBuilderVH::_do_xG_xM1_xM3(), in case of E/W global and only one RoutingPad the connexion to the RoutingPad was duplicated. It was valid, but totally stupid. * Bug: In Anabatic::Session::_canonize(), for an aligned segment set, intersect the user constraints from all segments instead of only considering the canonical one. Issue a warning about too tight constraints only for symbolic gauges. It may be correct for the real ones. * New: In Katata::DataNegociate::update(), more accurate computation of the perpandicular free interval. Use segment extension cap calculation. Create a special case for fixed axis segments allowing them to find alternative free interval, try under source and under target as they are likely to be draggable segments. * Change: In Katana::Manipulator::relax(), use the extension cap value to compute the axis of the perpandicular segemnts. * Change: In Katana::Manipulator::moveUp(), now move up the whole set of aligned segments instead of just the canonical one. * Change: In Katana::NegociateWindow::loadRoutingPads(), more accurate TrackMarkers insertions for fixed terminals. * New: In Katana::RoutingEvent::Key::Compare::operator(), segments with fixed axis are processed prior to any others. * New: In Katana::RoutingEventLoop, store segment pointers instead of ids to generate more accurate error messages. * Change: In Katana::RoutingPlane::create(), perform local track assignment only for HV gauges. * Change: In Katana::SegmentFsm::_slackenLocal(), add a "dragMinimize" step in the automaton. Mutliple states transitions can occurs in a row if an action fails. * New: In Katana::Session::_toIntervalAxis(), normalize interval bounds so they are on track positions (by shrinking the interval). * Bug: In Katana::TrackMarker CTOR, the weigh computation was wrong.
2018-02-17 13:27:38 -06:00
if (not _initialized) _initialize();
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
_allocateds++;
if (dynamic_cast<Horizontal*>(segment)) setFlags( SegHorizontal );
_globalsCount += isGlobal() ? 1 : 0;
Basic support for FreePDK 45 completed. * New: In Commons, inspector support for std::pair<T,U>. * New: In Hurricane::Layer, ContactLayer & ViaLayer, support for non square VIAs. The hole (cut) remains square, but the various metal extensions can now be different in X and Y. The ::getEnclosure() method now takes a flag EnclosureH / EnclosureV. * New: In Hurricane::DbU, inspector support for: std::pair<DbU::Unit,DbU::Unit> std::array<DbU::Unit,3> Must be defined here as DbU do not exists yet in Commons.h * Bug: In Hurricane::Interval::getSize(), when the interval is "full span", do not return the difference between min and max, but directly DbU::Max. (the previous result was -1 !) * New: In CRL Core Python/Technology.py, support for non square VIAs in the configuration files. Applied to FreePDK 45. * New: In CRL::RoutingGauge, added a "symbolic" flag to tell if a gauge is for symbolic layout or not. Exported to Python. * New: In Anabatic::AutoHorizontal::updatePosition(), differentiated computation for soure or target taking account of the VIA extension in the right segment metal (due to non-square VIAs). * Change: In Anabatic::AutoHorizontal::_makeDogleg(), the dogleg is UP for HV gauges and DOWN for VH. * New: In Anabatic::AutoSegment::_initialize(), create a cache of the various extension length for each layer (viaToTop, viaToBottom, viaToSame). New implementation of getExtensionCap() using the previous cached extension table. See updatePositions(). New static functions to access the extension cache in the header: getViaTotopCap() ... * Change: In Anabatic::AutoSegment, in various update methods, updateOrient() must always be called *before* updatePositions() as extensions are dependant on source/target. * New: In Anabatic::AutoSegment::getEndAxes() compute the position of the first source and last target position (center/axes) on an *aligned* set of segments. * New: In Anabatic::AutoSegment, add a new state flag SegAxisFixed to signal segments that can be put on only one track. Specific case to VH gauge for a M1 vertical terminal with a M2 vertical segment. The M2 is effectively bound to the M1 axis position. * Bug: In Anabatic::NetBuilderVH::_do_xG_xM1_xM3(), in case of E/W global and only one RoutingPad the connexion to the RoutingPad was duplicated. It was valid, but totally stupid. * Bug: In Anabatic::Session::_canonize(), for an aligned segment set, intersect the user constraints from all segments instead of only considering the canonical one. Issue a warning about too tight constraints only for symbolic gauges. It may be correct for the real ones. * New: In Katata::DataNegociate::update(), more accurate computation of the perpandicular free interval. Use segment extension cap calculation. Create a special case for fixed axis segments allowing them to find alternative free interval, try under source and under target as they are likely to be draggable segments. * Change: In Katana::Manipulator::relax(), use the extension cap value to compute the axis of the perpandicular segemnts. * Change: In Katana::Manipulator::moveUp(), now move up the whole set of aligned segments instead of just the canonical one. * Change: In Katana::NegociateWindow::loadRoutingPads(), more accurate TrackMarkers insertions for fixed terminals. * New: In Katana::RoutingEvent::Key::Compare::operator(), segments with fixed axis are processed prior to any others. * New: In Katana::RoutingEventLoop, store segment pointers instead of ids to generate more accurate error messages. * Change: In Katana::RoutingPlane::create(), perform local track assignment only for HV gauges. * Change: In Katana::SegmentFsm::_slackenLocal(), add a "dragMinimize" step in the automaton. Mutliple states transitions can occurs in a row if an action fails. * New: In Katana::Session::_toIntervalAxis(), normalize interval bounds so they are on track positions (by shrinking the interval). * Bug: In Katana::TrackMarker CTOR, the weigh computation was wrong.
2018-02-17 13:27:38 -06:00
AutoContact* source = Session::lookup(dynamic_cast<Contact*>(segment->getSource()));
AutoContact* target = Session::lookup(dynamic_cast<Contact*>(segment->getTarget()));
Forgot to perform Track re-order after removing zero-length segments. * Change: In Anabatic::Autocontact, replace getMinDepth() and getMaxDepth() by getDepthSpan(). * New: In Anabatic::AutoSegment::canMoveUp(), add an optional check of low up density (Flags::CheckLowUpDensity). Allows to move up a segment if the up density is (very) low, and in this case it's more efficient than breaking it to fit in the lower layer. canMoveUp() is now able to perform the same work as canPivotUp() if *not* supplied the flag Flags::IgnoreContacts. * New: In Katana, in GlobalRouting::DigitalDistance() now take into account the cost of a VIA (currently set to 2.5). Need to known the Net currently routed in the DigitalDistance object itself. * Change: In Katana::Track::Element::canPivotUp(), now also takes a flag parameter. * Change: In Katana::Manipulator, new flag IgnoreContacts to mirror the one of Anabatic. * Change: In Katana::SegmentFsm, allocate once a Manipulator object instead of many times on the fly. In SegmentFsm::_slackenGlobal(), in the slacken state, if the up density is (very) low, bypass to move up instead of slackening. This solve better the routing of the control part of the register file. The register file having a pathological case of terminal placement: many punctual terminals aligneds in METAL2 *and* a grid of METAL2 and METAL3 blockages near below... * Bug: In Katana::Session::_revalidate(), after removing the zero-length segments, forgot to re-order the track, leading to many stranges effects as the indexes where no longer coherent in the Track.
2016-09-20 04:30:45 -05:00
if (source->isTerminal()) setFlags( SegSourceTerminal );
if (target->isTerminal()) setFlags( SegTargetTerminal );
Added analog type on segment NetRoutingProperty. * New: In Anabatic_AutoSegments collection, added a Flag to the constructors to allow different behavior between digital and analog modes. For "Aligneds" and "Perpandiculars" collections, now manage a new Flag WithDoglegs to follow aligned globals through local doglegs (for analog nets). Adjust the log level of collections to 144 (formerly 145). * New: In Anabatic::AutoSegment, new flag SegAnalog for segments that are part of an analog net. Note that with this flag, we reach the 32 bits limit... * Change: In Anabatic::Constants, Flags are now declared as BaseFlags objects and *not* uint64_t. This avoids overload resolution problems with arithmetical overload of the operators. The BaseFlags/Flags types are now completly "isolated" from the uint64_t, it has the advantage of showing where unwanted previous implicit conversions where occuring. * Change: In Katana::Constants, Flags values are now of BaseFlags type instead of uint64_t. * Change: In Anabatic::Dijkstra, lots of log cleanup. * Change: In Anabatic::GCell::getSide(), make the "shrink" parameter visible to allow to substract the topmost and rightmost track for axis span computation in AutoSegment::computeOptimal(). Used for analog mode. * Change: In NetRoutingState, added a flag for analog mode. Use uint32_t for the flags type. * New: In Isobar, export the NetRoutingState and NetRoutingExtension objects.
2017-05-20 05:33:12 -05:00
if (_analogMode) setFlags( SegAnalog );
Improved handling of short nets (fully included in one GCell). The short net mode degrade the routing in some cases. This will be fixed in a next batch of commits. * New: In Hurricane::NetRoutingProperty, added "ShortNet" flag for Nets that are completly inside *one* GCell. * Bug: In CRL::BlifParser::Model::staticInit(), when looking for the output of zero and one cell, also skip the blockage net (as well as automatic and supplies). * New: In Anabatic::AutoSegment, added "ShortNet" flag to know if the segment is part of a short net (fully included in *one* GCell). Also add accessor/mutators for the _analogMode flag (was it ever used before?). * New: In Anabatic::NetBuilder::singleGCell(), if a RoutingPad is vertically small, add a vertical segment to give it some slack. * New: In Anabatic::Dijkstra::_materialize(), detect "short net" as they have only one GCell in their source list... * Bug: In AnabaticEngine::_loadGrbyNet(), reset the AutoSegment "short net" and "analog mode" creation flags between two different nets. * New: In Katana::Configuration, added dedicated ripup for short net segmnts. * New: In Katana: partially implemented support for "short dogleg", that is dogleg that are always kept in same metal because they connect neighboring perpandicular tracks. Not finished neither activated yet. * New: In Katana::TreckElement and derived, export the the *short net* support from AutoSegment. * Bug: In Katana::RoutingEvent::_processRepair(), when a segment is successfully inserted, re-process any perpandicular that is in repair state, as it may have a new chance to be placed. * New: In Katana::SegmentFsm::slackenTopology(), always reject short nets. * Bug: In Katana::Track::check(), correctly handle wide segments instead of issuing false check messages.
2018-07-16 04:16:51 -05:00
if (_shortNetMode) setFlags( SegShortNet );
Forgot to perform Track re-order after removing zero-length segments. * Change: In Anabatic::Autocontact, replace getMinDepth() and getMaxDepth() by getDepthSpan(). * New: In Anabatic::AutoSegment::canMoveUp(), add an optional check of low up density (Flags::CheckLowUpDensity). Allows to move up a segment if the up density is (very) low, and in this case it's more efficient than breaking it to fit in the lower layer. canMoveUp() is now able to perform the same work as canPivotUp() if *not* supplied the flag Flags::IgnoreContacts. * New: In Katana, in GlobalRouting::DigitalDistance() now take into account the cost of a VIA (currently set to 2.5). Need to known the Net currently routed in the DigitalDistance object itself. * Change: In Katana::Track::Element::canPivotUp(), now also takes a flag parameter. * Change: In Katana::Manipulator, new flag IgnoreContacts to mirror the one of Anabatic. * Change: In Katana::SegmentFsm, allocate once a Manipulator object instead of many times on the fly. In SegmentFsm::_slackenGlobal(), in the slacken state, if the up density is (very) low, bypass to move up instead of slackening. This solve better the routing of the control part of the register file. The register file having a pathological case of terminal placement: many punctual terminals aligneds in METAL2 *and* a grid of METAL2 and METAL3 blockages near below... * Bug: In Katana::Session::_revalidate(), after removing the zero-length segments, forgot to re-order the track, leading to many stranges effects as the indexes where no longer coherent in the Track.
2016-09-20 04:30:45 -05:00
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
source->invalidate( Flags::Topology );
}
void AutoSegment::_preCreate ( AutoContact* source, AutoContact* target )
{
if ( (source == NULL) or (target == NULL) )
throw Error( badAutoSegmentAnchor
, ((source)?getString(source).c_str():"NULL")
, ((target)?getString(target).c_str():"NULL")
);
if (source == target)
throw Error( dupAutoSegmentAnchor, getString(source).c_str() );
}
void AutoSegment::_postCreate ()
{
Session::invalidate( getNet() );
Session::link( this );
updateOrient();
updatePositions();
invalidate( Flags::Topology );
_observers.notify( Create );
}
void AutoSegment::_preDestroy ()
{
cdebug_log(149,0) << "AutoSegment::_preDestroy() - " << (void*)this << endl;
cdebug_tabw(145,1);
_observers.notify( Destroy );
Fix for Coriolis issue #6 (Katana core dump on ALU16). * Bug: In Katana::AutoSegment::_preDestroy(), remove the segment from the source & target AutoContact cache. In LibreSOC/experiment7, weird placement caused fixed AutoSegment overlaping blockage to be deleteds. It seems to have never occured before (or at least, no ended up in core dump).
2020-04-15 03:41:46 -05:00
AutoContact* contact = getAutoSource();
if (contact) contact->cacheDetach( this );
contact = getAutoTarget();
if (contact) contact->cacheDetach( this );
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
Session::unlink( this );
cdebug_tabw(145,-1);
}
AutoSegment::~AutoSegment ()
{
_allocateds--;
if ( isGlobal() and (_globalsCount > 0) ) _globalsCount--;
}
DbU::Unit AutoSegment::getX () const
{ return base()->getX(); }
DbU::Unit AutoSegment::getY () const
{ return base()->getY(); }
AutoContact* AutoSegment::getOppositeAnchor ( AutoContact* anchor ) const
{ return Session::lookup(static_cast<Contact*>(getOppositeAnchor(anchor->base()))); }
Interval& AutoSegment::getOptimal ( Interval& i ) const
{
i.getVMin() = getOptimalMin();
i.getVMax() = getOptimalMax();
return i;
}
bool AutoSegment::checkNotInvalidated () const
{
if (isInvalidated())
cerr << Error("%s is invalidated.",getString(this).c_str()) << endl;
return not isInvalidated();
}
Replace "unsigned int" by "Flags" in all AutoSegments collections. * Change: In Anabatic::AutoSegments collections, change the type of all the flags that where in "unsigned int" (32 bits) to Flags (uint64_t) as there is now more than 32 flags for functions. * New: In Ababatic::Constants, added new flag Flags::WithPerpands, which makes the number of flags tip over 32 bits, thus making mandatory to uses Flags and not unsigned int. * New: In Anabatic::AutoSegments_Perpandiculars, manage a new flag Flags::WithDoglegs to allow to propagate through global segments that are connecteds via doglegs on local segments. Meaning that there is a good chance that they could be aligned. Slighly change the way we propagate on aligned segments: no longer check for VTee or HTee, but only for same direction and layer as master. * New: In Anabatic & Katana, replace all the "int", "long" and their variants by the less implementation ambiguous "int32_t", "int64_t" (and variant). This should help to better detect bit trucation in flags. Use the type to give a hint about the flags kind: - Type "Flags", for flags shared among Anabatic & Katana functions/methods (may also appear in some objects states). - Type "uint32_t" for flags belonging to an object internal state of from Hurricane functions flags (those should be grouped in a Flag subclass in a perfect world).
2017-05-16 07:53:33 -05:00
void AutoSegment::invalidate ( Flags flags )
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
{
if (Session::doDestroyTool()) return;
Topology update was not correctly done after layer change. * Bug: In Anabatic::AutoSegment, Anabatic::AutoContact and derived classes, the recursive invalidation mechanism did have cross-call troubles. - The various AutoContact::_invalidate() now take the flag into account instead of resetting it. This is for Flags::Topology to be passed along from segments to contacts. - In AutoSegment::invalidate(), do not invalidate S/T contacts already invalidateds. In the same way, do not invalidate aligned segments already invalidateds (save some recursion). - In AutoSegment::_changeDepth(), call invalidate with Flags::Topology (to force contact update) and Flags::NoCheckLayer to avoid disgraceful warnings. * New: Anabatic::NetBuilder::_do_xG_1M1() with (x >= 2) new set of optimized topologies when there is multiple globals but only one terminal. The default one was making two connections to the same RoutingPad, which was making clutter.
2018-03-27 11:03:51 -05:00
cdebug_log(149,0) << "AutoSegment::invalidate() " << flags.asString(FlagsFunction)
<< " " << this << endl;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
cdebug_tabw(149,1);
Improved management of AutoContactTerminal for VH gauges (real ones). * New: In Anabatic & Katana, add the new "drag" feature. With VH gauges used by real technologies (M1-H, M2-V, M3-H) a new routing configuration that was not efficiently handled did appear. While the preferred routing direction for metal1 is officially horizontal, due to the way the standard cell must be designed, their metal1 terminals are still verticals (or punctuals). Thus, when connecting to them, we face the case where the metal1 terminal (RoutingPad) is vertical *and* the metal2 wire is also vertical. With that setup, the position of the AutoContactTerminal via12 cannot be deduced, it may range all the way over the metal1 RoutingPad. What may define it's position is the metal3 the metal2 finally connects to. That, is, when we have one horizontal (the metal3) and one vertical (the metal1 RoutingPad). The intermediate wire of metal2 can be kept to a minimum size by "dragging" the via12 close to the via23 when the metal3 wire is moved. * New: In Anabatic & Katana, problem of closely vertically aligneds RoutingPads in metal1 is managed first in PreProcess by restricting the span of the connecteds metal3 and in _makeDogleg also by restricting the span even more tightly (to the RoutingPad itself). * New: In Anabatic::AutoContactTerminal, add the "drag" support. Automatically check if the connecting segment is in the same direction as the RoutingPad, if so, sets the "SegDrag" flag. The dragging state can be known with the "::canDrag()" predicate. * New: In Anabatic::AutoHorizontal, add the "drag" support. The drag state can be known with the "::isDrag()" predicate. In "::_makeDogleg()", when making a dogleg on a dragable segment pass the drag state correctly and restrict the perpandicular span of the perpandicular to the RoutingPad (though segment user constraints). If we make a dogleg on the metal2 is it likely than we cannot go straigth out vertically from the RoutingPad, so the new perpandicular *is* restricted to the RoutingPad span. Idem for AutoVertical. * New: In Katana::Manipulator, add method "::dragMinimize()" which find a hole where to minimize a draggable segment. We finally did not use it, but keep it for potential further use. * New: In Katana::PreProcess, adds a "protectAlignedaccesses()" local function to check for vertically aligned metal1 RoutingPads, in that case setup user constraints on the metal3 segments so they cannot completly cover the other RoutingPad with metal2. We also keep a "metal2protect()" function that create a fixed segment to lock/protect a RoutingPad. Not used for now. * New: In Katana::Session, add a RoutingPad locking event mechanism. This allows us to request the creation of a locking (fixed segment) over a draggable segment. Not used for now. Lock events are processeds before all others as they create new TrackElements. * New: In Katana::Track, "::getNextFree()" and "::getPreviousFree()" method to find the nearest free interval in a Track after/before a position. * Bug: In Anabatic::AutoHorizontal::getConstraints(), merge with user constraints *only* if it's not an empty interval (as we use min/max functions). Idem for AutoVertical. * Bug: In AutoSegments_OnContacts::Locator::isValid(), the boolean test must be inverted. Seems it never worked, but we never used it until now...
2018-01-25 04:58:04 -06:00
if (flags & Flags::Source) setFlags( SegInvalidatedSource );
if (flags & Flags::Target) setFlags( SegInvalidatedTarget );
Topology update was not correctly done after layer change. * Bug: In Anabatic::AutoSegment, Anabatic::AutoContact and derived classes, the recursive invalidation mechanism did have cross-call troubles. - The various AutoContact::_invalidate() now take the flag into account instead of resetting it. This is for Flags::Topology to be passed along from segments to contacts. - In AutoSegment::invalidate(), do not invalidate S/T contacts already invalidateds. In the same way, do not invalidate aligned segments already invalidateds (save some recursion). - In AutoSegment::_changeDepth(), call invalidate with Flags::Topology (to force contact update) and Flags::NoCheckLayer to avoid disgraceful warnings. * New: Anabatic::NetBuilder::_do_xG_1M1() with (x >= 2) new set of optimized topologies when there is multiple globals but only one terminal. The default one was making two connections to the same RoutingPad, which was making clutter.
2018-03-27 11:03:51 -05:00
if ( (getFlags() & SegSourceTerminal)
and getAutoSource()
and getAutoSource()->canDrag()
and not getAutoSource()->isInvalidated() )
getAutoSource()->invalidate( flags );
Improved management of AutoContactTerminal for VH gauges (real ones). * New: In Anabatic & Katana, add the new "drag" feature. With VH gauges used by real technologies (M1-H, M2-V, M3-H) a new routing configuration that was not efficiently handled did appear. While the preferred routing direction for metal1 is officially horizontal, due to the way the standard cell must be designed, their metal1 terminals are still verticals (or punctuals). Thus, when connecting to them, we face the case where the metal1 terminal (RoutingPad) is vertical *and* the metal2 wire is also vertical. With that setup, the position of the AutoContactTerminal via12 cannot be deduced, it may range all the way over the metal1 RoutingPad. What may define it's position is the metal3 the metal2 finally connects to. That, is, when we have one horizontal (the metal3) and one vertical (the metal1 RoutingPad). The intermediate wire of metal2 can be kept to a minimum size by "dragging" the via12 close to the via23 when the metal3 wire is moved. * New: In Anabatic & Katana, problem of closely vertically aligneds RoutingPads in metal1 is managed first in PreProcess by restricting the span of the connecteds metal3 and in _makeDogleg also by restricting the span even more tightly (to the RoutingPad itself). * New: In Anabatic::AutoContactTerminal, add the "drag" support. Automatically check if the connecting segment is in the same direction as the RoutingPad, if so, sets the "SegDrag" flag. The dragging state can be known with the "::canDrag()" predicate. * New: In Anabatic::AutoHorizontal, add the "drag" support. The drag state can be known with the "::isDrag()" predicate. In "::_makeDogleg()", when making a dogleg on a dragable segment pass the drag state correctly and restrict the perpandicular span of the perpandicular to the RoutingPad (though segment user constraints). If we make a dogleg on the metal2 is it likely than we cannot go straigth out vertically from the RoutingPad, so the new perpandicular *is* restricted to the RoutingPad span. Idem for AutoVertical. * New: In Katana::Manipulator, add method "::dragMinimize()" which find a hole where to minimize a draggable segment. We finally did not use it, but keep it for potential further use. * New: In Katana::PreProcess, adds a "protectAlignedaccesses()" local function to check for vertically aligned metal1 RoutingPads, in that case setup user constraints on the metal3 segments so they cannot completly cover the other RoutingPad with metal2. We also keep a "metal2protect()" function that create a fixed segment to lock/protect a RoutingPad. Not used for now. * New: In Katana::Session, add a RoutingPad locking event mechanism. This allows us to request the creation of a locking (fixed segment) over a draggable segment. Not used for now. Lock events are processeds before all others as they create new TrackElements. * New: In Katana::Track, "::getNextFree()" and "::getPreviousFree()" method to find the nearest free interval in a Track after/before a position. * Bug: In Anabatic::AutoHorizontal::getConstraints(), merge with user constraints *only* if it's not an empty interval (as we use min/max functions). Idem for AutoVertical. * Bug: In AutoSegments_OnContacts::Locator::isValid(), the boolean test must be inverted. Seems it never worked, but we never used it until now...
2018-01-25 04:58:04 -06:00
Topology update was not correctly done after layer change. * Bug: In Anabatic::AutoSegment, Anabatic::AutoContact and derived classes, the recursive invalidation mechanism did have cross-call troubles. - The various AutoContact::_invalidate() now take the flag into account instead of resetting it. This is for Flags::Topology to be passed along from segments to contacts. - In AutoSegment::invalidate(), do not invalidate S/T contacts already invalidateds. In the same way, do not invalidate aligned segments already invalidateds (save some recursion). - In AutoSegment::_changeDepth(), call invalidate with Flags::Topology (to force contact update) and Flags::NoCheckLayer to avoid disgraceful warnings. * New: Anabatic::NetBuilder::_do_xG_1M1() with (x >= 2) new set of optimized topologies when there is multiple globals but only one terminal. The default one was making two connections to the same RoutingPad, which was making clutter.
2018-03-27 11:03:51 -05:00
if ( (getFlags() & SegTargetTerminal)
and getAutoTarget()
and getAutoTarget()->canDrag()
and not getAutoTarget()->isInvalidated() )
getAutoTarget()->invalidate( flags );
Improved management of AutoContactTerminal for VH gauges (real ones). * New: In Anabatic & Katana, add the new "drag" feature. With VH gauges used by real technologies (M1-H, M2-V, M3-H) a new routing configuration that was not efficiently handled did appear. While the preferred routing direction for metal1 is officially horizontal, due to the way the standard cell must be designed, their metal1 terminals are still verticals (or punctuals). Thus, when connecting to them, we face the case where the metal1 terminal (RoutingPad) is vertical *and* the metal2 wire is also vertical. With that setup, the position of the AutoContactTerminal via12 cannot be deduced, it may range all the way over the metal1 RoutingPad. What may define it's position is the metal3 the metal2 finally connects to. That, is, when we have one horizontal (the metal3) and one vertical (the metal1 RoutingPad). The intermediate wire of metal2 can be kept to a minimum size by "dragging" the via12 close to the via23 when the metal3 wire is moved. * New: In Anabatic & Katana, problem of closely vertically aligneds RoutingPads in metal1 is managed first in PreProcess by restricting the span of the connecteds metal3 and in _makeDogleg also by restricting the span even more tightly (to the RoutingPad itself). * New: In Anabatic::AutoContactTerminal, add the "drag" support. Automatically check if the connecting segment is in the same direction as the RoutingPad, if so, sets the "SegDrag" flag. The dragging state can be known with the "::canDrag()" predicate. * New: In Anabatic::AutoHorizontal, add the "drag" support. The drag state can be known with the "::isDrag()" predicate. In "::_makeDogleg()", when making a dogleg on a dragable segment pass the drag state correctly and restrict the perpandicular span of the perpandicular to the RoutingPad (though segment user constraints). If we make a dogleg on the metal2 is it likely than we cannot go straigth out vertically from the RoutingPad, so the new perpandicular *is* restricted to the RoutingPad span. Idem for AutoVertical. * New: In Katana::Manipulator, add method "::dragMinimize()" which find a hole where to minimize a draggable segment. We finally did not use it, but keep it for potential further use. * New: In Katana::PreProcess, adds a "protectAlignedaccesses()" local function to check for vertically aligned metal1 RoutingPads, in that case setup user constraints on the metal3 segments so they cannot completly cover the other RoutingPad with metal2. We also keep a "metal2protect()" function that create a fixed segment to lock/protect a RoutingPad. Not used for now. * New: In Katana::Session, add a RoutingPad locking event mechanism. This allows us to request the creation of a locking (fixed segment) over a draggable segment. Not used for now. Lock events are processeds before all others as they create new TrackElements. * New: In Katana::Track, "::getNextFree()" and "::getPreviousFree()" method to find the nearest free interval in a Track after/before a position. * Bug: In Anabatic::AutoHorizontal::getConstraints(), merge with user constraints *only* if it's not an empty interval (as we use min/max functions). Idem for AutoVertical. * Bug: In AutoSegments_OnContacts::Locator::isValid(), the boolean test must be inverted. Seems it never worked, but we never used it until now...
2018-01-25 04:58:04 -06:00
if (isInvalidated()) { cdebug_tabw(149,-1); return; }
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
_invalidate();
if ((flags & Flags::Propagate) and not isNotAligned()) {
Topology update was not correctly done after layer change. * Bug: In Anabatic::AutoSegment, Anabatic::AutoContact and derived classes, the recursive invalidation mechanism did have cross-call troubles. - The various AutoContact::_invalidate() now take the flag into account instead of resetting it. This is for Flags::Topology to be passed along from segments to contacts. - In AutoSegment::invalidate(), do not invalidate S/T contacts already invalidateds. In the same way, do not invalidate aligned segments already invalidateds (save some recursion). - In AutoSegment::_changeDepth(), call invalidate with Flags::Topology (to force contact update) and Flags::NoCheckLayer to avoid disgraceful warnings. * New: Anabatic::NetBuilder::_do_xG_1M1() with (x >= 2) new set of optimized topologies when there is multiple globals but only one terminal. The default one was making two connections to the same RoutingPad, which was making clutter.
2018-03-27 11:03:51 -05:00
for ( AutoSegment* segment : getAligneds(flags & Flags::NoCheckLayer) ) {
if (not segment->isInvalidated()) segment->_invalidate();
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
}
}
cdebug_tabw(149,-1);
}
void AutoSegment::_invalidate ()
{
if (isInvalidated()) return;
cdebug_log(145,0) << "AutoSegment::_invalidate() " << this << endl;
setFlags( SegInvalidated );
Session::invalidate( this );
_observers.notify( Invalidate );
}
void AutoSegment::invalidate ( AutoContact* contact )
{
if (Session::doDestroyTool()) return;
if (contact == getAutoSource()) setFlags( SegInvalidatedSource );
if (contact == getAutoTarget()) setFlags( SegInvalidatedTarget );
}
void AutoSegment::revalidate ()
{
cdebug_log(149,0) << "AutoSegment::revalidate() " << this << endl;
if (not isInvalidated()) return;
cdebug_tabw(149,1);
Basic support for FreePDK 45 completed. * New: In Commons, inspector support for std::pair<T,U>. * New: In Hurricane::Layer, ContactLayer & ViaLayer, support for non square VIAs. The hole (cut) remains square, but the various metal extensions can now be different in X and Y. The ::getEnclosure() method now takes a flag EnclosureH / EnclosureV. * New: In Hurricane::DbU, inspector support for: std::pair<DbU::Unit,DbU::Unit> std::array<DbU::Unit,3> Must be defined here as DbU do not exists yet in Commons.h * Bug: In Hurricane::Interval::getSize(), when the interval is "full span", do not return the difference between min and max, but directly DbU::Max. (the previous result was -1 !) * New: In CRL Core Python/Technology.py, support for non square VIAs in the configuration files. Applied to FreePDK 45. * New: In CRL::RoutingGauge, added a "symbolic" flag to tell if a gauge is for symbolic layout or not. Exported to Python. * New: In Anabatic::AutoHorizontal::updatePosition(), differentiated computation for soure or target taking account of the VIA extension in the right segment metal (due to non-square VIAs). * Change: In Anabatic::AutoHorizontal::_makeDogleg(), the dogleg is UP for HV gauges and DOWN for VH. * New: In Anabatic::AutoSegment::_initialize(), create a cache of the various extension length for each layer (viaToTop, viaToBottom, viaToSame). New implementation of getExtensionCap() using the previous cached extension table. See updatePositions(). New static functions to access the extension cache in the header: getViaTotopCap() ... * Change: In Anabatic::AutoSegment, in various update methods, updateOrient() must always be called *before* updatePositions() as extensions are dependant on source/target. * New: In Anabatic::AutoSegment::getEndAxes() compute the position of the first source and last target position (center/axes) on an *aligned* set of segments. * New: In Anabatic::AutoSegment, add a new state flag SegAxisFixed to signal segments that can be put on only one track. Specific case to VH gauge for a M1 vertical terminal with a M2 vertical segment. The M2 is effectively bound to the M1 axis position. * Bug: In Anabatic::NetBuilderVH::_do_xG_xM1_xM3(), in case of E/W global and only one RoutingPad the connexion to the RoutingPad was duplicated. It was valid, but totally stupid. * Bug: In Anabatic::Session::_canonize(), for an aligned segment set, intersect the user constraints from all segments instead of only considering the canonical one. Issue a warning about too tight constraints only for symbolic gauges. It may be correct for the real ones. * New: In Katata::DataNegociate::update(), more accurate computation of the perpandicular free interval. Use segment extension cap calculation. Create a special case for fixed axis segments allowing them to find alternative free interval, try under source and under target as they are likely to be draggable segments. * Change: In Katana::Manipulator::relax(), use the extension cap value to compute the axis of the perpandicular segemnts. * Change: In Katana::Manipulator::moveUp(), now move up the whole set of aligned segments instead of just the canonical one. * Change: In Katana::NegociateWindow::loadRoutingPads(), more accurate TrackMarkers insertions for fixed terminals. * New: In Katana::RoutingEvent::Key::Compare::operator(), segments with fixed axis are processed prior to any others. * New: In Katana::RoutingEventLoop, store segment pointers instead of ids to generate more accurate error messages. * Change: In Katana::RoutingPlane::create(), perform local track assignment only for HV gauges. * Change: In Katana::SegmentFsm::_slackenLocal(), add a "dragMinimize" step in the automaton. Mutliple states transitions can occurs in a row if an action fails. * New: In Katana::Session::_toIntervalAxis(), normalize interval bounds so they are on track positions (by shrinking the interval). * Bug: In Katana::TrackMarker CTOR, the weigh computation was wrong.
2018-02-17 13:27:38 -06:00
updateOrient();
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
Katana manage wide wires, and they can also be symmetric. * New: In Anabatic::AutoContact and the derived classes, manages wide wires. The contact self dimension itself according to the segments it is connected to. Special case for the AutoContactTerminal which also read the size of the component it is anchored upon. New refresh method "updateSize()" and flag CntInvalidatedWidth. to compute the size. In AutoContactTerminal, compute the constraint box according to the width of the segment. * New: In Anabatic::AutoSegment, flags are now implemented as "static const" attributes of the class. The flags are stored into a uint64_t as they are more than 32. Added new flag "SegWide" and associated predicates. * Change: In GCellTopology::_doHChannel() and GCellTopology::_doVChannel(), uses the simpler overload of AutoSegment::create() in order to detect the wire width automatically. * New: In Katana::Manipulator, split insertToTrack() and forceToTrack() into a one-track method and a segment level method that iterate over the track span of the segment. * New: In Katana::SegmentFsm, for each cost in the table, now allow access to a specific track. So the base functions have now two parameters: "icost" and "itrack" (has a cost can have multiple tracks in the case of wide segments). * Change: In Katana::TrackElement, remove the index of the element inside it's track, as for a wide segment it will not be meaningful for the non-base track. This means that we have to use the Track::find() method each time instead. Remove the wide flag, as it is a duplicate of the one in AutoSegment. Added a getTrackCount() method to tell the number of track the segment is inserted into. Needed in the Track destroy step to delete a segment only when the last track that refers it is destroyed. Added getSymmetricAxis() to correct the computation of the symmetric base track in case of wide segment as the base track is not centered but the the leftmost one. * Change: In Track::insert() insert wide segments in their whole track span. * Change: In TrackCost, create an array of costs according to the segment track span. * Change: In TrackSegment::create(), now activate the factory and create wide segments. * Bug: In Katana::AutoSegments_Perpandicular, correct the debug indentation problem (ever shifting to the right).
2017-07-28 08:30:22 -05:00
uint64_t oldSpinFlags = _flags & SegDepthSpin;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
if (_flags & (SegInvalidatedSource|SegCreated)) {
AutoContact* source = getAutoSource();
const Layer* contactLayer = source->getLayer();
const Layer* segmentLayer = getLayer();
cdebug_log(149,0) << "Changed source: " << source << endl;
unsetFlags( SegSourceTop|SegSourceBottom );
Redesigning access to analog terminals in devices. * New: In Anabatic::AutoContact, added forceOnGrid() method to allow a AutoContactTerminal to be set (and fixed) outside it's constraint Box. Be aware that this may generate gaps in the routing. Stored using the UserNativeConstraints mechanism. * New: In Anabatic::Session() new method getNearestgridpoint() to easily compute the nearest grid point from a position. * Change: In Anabatic::AutoSegment::_revalidate(), use layer mask to compute the spin of the segment instead of the layers (more robust for analogic). * New: In ::GCellTopology::doRp_Access() new method to create terminal contacts in _doDevice(). Issue more warnings and force the terminal to on grid (even if it create gaps). This a temporary hack.
2016-10-05 16:43:07 -05:00
if (contactLayer->getMask() != segmentLayer->getMask())
In Anabatic/Katana, add support for VH gauges (real technos). * Change: In Anabatic::AutoContactTerminal::getNativeConstraintBox(), when the anchor is a RoutingPad (which must be always the case), perform the true computation of it's position based on the segment occurrence. It is a important change, previously the area was in fact the "center line" of the connector while now it is really an area (mandatory for "half-offgrid" terminals of real technologies). The change is not complete yet, the area should be shrinked by the half size of a VIA, because the area applies to the center coordinate of the VIA (to be done quickly). * Bug: In Anabatic::AutoContactTurn::updateTopology(), when a dogleg is created (restore connexity after a layer change) the layer of the VIA, based on the segments it connects to must be re-computed *after* the dogleg has been made. * Change: In all files of Anabatic, when comparing two layers, no longer use the Layer pointer itself, but the layer mask. This allow a transparent management of both real and symbolic layers (which do share the same mask). Real metal layers (not VIAs) will be BasicLayer and symbolic metal layers will be RegularLayer. * New: Anabatic::Configuration::selectRpComponent(), select the best RoutingPad component for metal1 terminals. Look for the metal1 component with the biggest accessibility on-grid. RoutingPad using other metals are left untoucheds. * New: New function Anabatic::Vertex::getNeighbor(Edge*) to get the neighbor Vertex through an Edge*. This method allows to write clearer code as we no longer need to access the neighbor through the underlying GCell. Also add proxies for GCell methods in Vertex. * Bug: In Anabatic::Dijkstra::_toSources(), in the ripup stage, when a component with multiples vertexes is reached *and* two of it's vertexes are reached *at the same time* (one from which we backtrack and one still in the queue) extraneous edges may be created by _materialize(). Case occurs on snx/c35b4, "abc_5360_n903_1". To solve this, Dijkstra::_toSource() is modificated, the "from" edges of the newly reacheds vertexes are reset to NULL, *except* for the one we will be backtracking from. That is, the one given in the source argument. * Change: In Anabatic::NetBuilder class, put the various Hooks and RoutingPad sorting functions as class ones. * Bug: In AutoSegment::setLayer(), raise the SegInvalidatedFayer flag. This unset flag was causing AutoContactTurn::updateTopology() to not work as expected and making gaps, this was the cause of the last remaining warnings about layer connexity.
2018-01-06 09:55:53 -06:00
setFlags( (segmentLayer->getMask() == contactLayer->getTop()->getMask()) ? SegSourceBottom : SegSourceTop );
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
if (source->isTurn() and source->getPerpandicular(this)->isReduced())
incReduceds();
}
if (_flags & (SegInvalidatedTarget|SegCreated)) {
AutoContact* target = getAutoTarget();
const Layer* contactLayer = target->getLayer();
const Layer* segmentLayer = getLayer();
cdebug_log(149,0) << "Changed target: " << target << endl;
unsetFlags( SegTargetTop|SegTargetBottom );
Redesigning access to analog terminals in devices. * New: In Anabatic::AutoContact, added forceOnGrid() method to allow a AutoContactTerminal to be set (and fixed) outside it's constraint Box. Be aware that this may generate gaps in the routing. Stored using the UserNativeConstraints mechanism. * New: In Anabatic::Session() new method getNearestgridpoint() to easily compute the nearest grid point from a position. * Change: In Anabatic::AutoSegment::_revalidate(), use layer mask to compute the spin of the segment instead of the layers (more robust for analogic). * New: In ::GCellTopology::doRp_Access() new method to create terminal contacts in _doDevice(). Issue more warnings and force the terminal to on grid (even if it create gaps). This a temporary hack.
2016-10-05 16:43:07 -05:00
if (contactLayer->getMask() != segmentLayer->getMask())
In Anabatic/Katana, add support for VH gauges (real technos). * Change: In Anabatic::AutoContactTerminal::getNativeConstraintBox(), when the anchor is a RoutingPad (which must be always the case), perform the true computation of it's position based on the segment occurrence. It is a important change, previously the area was in fact the "center line" of the connector while now it is really an area (mandatory for "half-offgrid" terminals of real technologies). The change is not complete yet, the area should be shrinked by the half size of a VIA, because the area applies to the center coordinate of the VIA (to be done quickly). * Bug: In Anabatic::AutoContactTurn::updateTopology(), when a dogleg is created (restore connexity after a layer change) the layer of the VIA, based on the segments it connects to must be re-computed *after* the dogleg has been made. * Change: In all files of Anabatic, when comparing two layers, no longer use the Layer pointer itself, but the layer mask. This allow a transparent management of both real and symbolic layers (which do share the same mask). Real metal layers (not VIAs) will be BasicLayer and symbolic metal layers will be RegularLayer. * New: Anabatic::Configuration::selectRpComponent(), select the best RoutingPad component for metal1 terminals. Look for the metal1 component with the biggest accessibility on-grid. RoutingPad using other metals are left untoucheds. * New: New function Anabatic::Vertex::getNeighbor(Edge*) to get the neighbor Vertex through an Edge*. This method allows to write clearer code as we no longer need to access the neighbor through the underlying GCell. Also add proxies for GCell methods in Vertex. * Bug: In Anabatic::Dijkstra::_toSources(), in the ripup stage, when a component with multiples vertexes is reached *and* two of it's vertexes are reached *at the same time* (one from which we backtrack and one still in the queue) extraneous edges may be created by _materialize(). Case occurs on snx/c35b4, "abc_5360_n903_1". To solve this, Dijkstra::_toSource() is modificated, the "from" edges of the newly reacheds vertexes are reset to NULL, *except* for the one we will be backtracking from. That is, the one given in the source argument. * Change: In Anabatic::NetBuilder class, put the various Hooks and RoutingPad sorting functions as class ones. * Bug: In AutoSegment::setLayer(), raise the SegInvalidatedFayer flag. This unset flag was causing AutoContactTurn::updateTopology() to not work as expected and making gaps, this was the cause of the last remaining warnings about layer connexity.
2018-01-06 09:55:53 -06:00
setFlags( (segmentLayer->getMask() == contactLayer->getTop()->getMask()) ? SegTargetBottom : SegTargetTop );
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
if (target->isTurn() and target->getPerpandicular(this)->isReduced())
incReduceds();
}
Basic support for FreePDK 45 completed. * New: In Commons, inspector support for std::pair<T,U>. * New: In Hurricane::Layer, ContactLayer & ViaLayer, support for non square VIAs. The hole (cut) remains square, but the various metal extensions can now be different in X and Y. The ::getEnclosure() method now takes a flag EnclosureH / EnclosureV. * New: In Hurricane::DbU, inspector support for: std::pair<DbU::Unit,DbU::Unit> std::array<DbU::Unit,3> Must be defined here as DbU do not exists yet in Commons.h * Bug: In Hurricane::Interval::getSize(), when the interval is "full span", do not return the difference between min and max, but directly DbU::Max. (the previous result was -1 !) * New: In CRL Core Python/Technology.py, support for non square VIAs in the configuration files. Applied to FreePDK 45. * New: In CRL::RoutingGauge, added a "symbolic" flag to tell if a gauge is for symbolic layout or not. Exported to Python. * New: In Anabatic::AutoHorizontal::updatePosition(), differentiated computation for soure or target taking account of the VIA extension in the right segment metal (due to non-square VIAs). * Change: In Anabatic::AutoHorizontal::_makeDogleg(), the dogleg is UP for HV gauges and DOWN for VH. * New: In Anabatic::AutoSegment::_initialize(), create a cache of the various extension length for each layer (viaToTop, viaToBottom, viaToSame). New implementation of getExtensionCap() using the previous cached extension table. See updatePositions(). New static functions to access the extension cache in the header: getViaTotopCap() ... * Change: In Anabatic::AutoSegment, in various update methods, updateOrient() must always be called *before* updatePositions() as extensions are dependant on source/target. * New: In Anabatic::AutoSegment::getEndAxes() compute the position of the first source and last target position (center/axes) on an *aligned* set of segments. * New: In Anabatic::AutoSegment, add a new state flag SegAxisFixed to signal segments that can be put on only one track. Specific case to VH gauge for a M1 vertical terminal with a M2 vertical segment. The M2 is effectively bound to the M1 axis position. * Bug: In Anabatic::NetBuilderVH::_do_xG_xM1_xM3(), in case of E/W global and only one RoutingPad the connexion to the RoutingPad was duplicated. It was valid, but totally stupid. * Bug: In Anabatic::Session::_canonize(), for an aligned segment set, intersect the user constraints from all segments instead of only considering the canonical one. Issue a warning about too tight constraints only for symbolic gauges. It may be correct for the real ones. * New: In Katata::DataNegociate::update(), more accurate computation of the perpandicular free interval. Use segment extension cap calculation. Create a special case for fixed axis segments allowing them to find alternative free interval, try under source and under target as they are likely to be draggable segments. * Change: In Katana::Manipulator::relax(), use the extension cap value to compute the axis of the perpandicular segemnts. * Change: In Katana::Manipulator::moveUp(), now move up the whole set of aligned segments instead of just the canonical one. * Change: In Katana::NegociateWindow::loadRoutingPads(), more accurate TrackMarkers insertions for fixed terminals. * New: In Katana::RoutingEvent::Key::Compare::operator(), segments with fixed axis are processed prior to any others. * New: In Katana::RoutingEventLoop, store segment pointers instead of ids to generate more accurate error messages. * Change: In Katana::RoutingPlane::create(), perform local track assignment only for HV gauges. * Change: In Katana::SegmentFsm::_slackenLocal(), add a "dragMinimize" step in the automaton. Mutliple states transitions can occurs in a row if an action fails. * New: In Katana::Session::_toIntervalAxis(), normalize interval bounds so they are on track positions (by shrinking the interval). * Bug: In Katana::TrackMarker CTOR, the weigh computation was wrong.
2018-02-17 13:27:38 -06:00
updatePositions();
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
unsigned int observerFlags = Revalidate;
if ( (_flags & SegCreated) or (oldSpinFlags != (_flags & SegDepthSpin)) )
observerFlags |= RevalidatePPitch;
unsetFlags( SegInvalidated
| SegInvalidatedSource
| SegInvalidatedTarget
| SegInvalidatedLayer
| SegCreated
);
_observers.notify( observerFlags );
cdebug_log(149,0) << "Updated: " << this << endl;
cdebug_tabw(149,-1);
}
Replace "unsigned int" by "Flags" in all AutoSegments collections. * Change: In Anabatic::AutoSegments collections, change the type of all the flags that where in "unsigned int" (32 bits) to Flags (uint64_t) as there is now more than 32 flags for functions. * New: In Ababatic::Constants, added new flag Flags::WithPerpands, which makes the number of flags tip over 32 bits, thus making mandatory to uses Flags and not unsigned int. * New: In Anabatic::AutoSegments_Perpandiculars, manage a new flag Flags::WithDoglegs to allow to propagate through global segments that are connecteds via doglegs on local segments. Meaning that there is a good chance that they could be aligned. Slighly change the way we propagate on aligned segments: no longer check for VTee or HTee, but only for same direction and layer as master. * New: In Anabatic & Katana, replace all the "int", "long" and their variants by the less implementation ambiguous "int32_t", "int64_t" (and variant). This should help to better detect bit trucation in flags. Use the type to give a hint about the flags kind: - Type "Flags", for flags shared among Anabatic & Katana functions/methods (may also appear in some objects states). - Type "uint32_t" for flags belonging to an object internal state of from Hurricane functions flags (those should be grouped in a Flag subclass in a perfect world).
2017-05-16 07:53:33 -05:00
bool AutoSegment::isStrongTerminal ( Flags flags ) const
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
{
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
if (isTerminal()) return true;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
if ((flags & Flags::Propagate) and not isNotAligned()) {
Replace "unsigned int" by "Flags" in all AutoSegments collections. * Change: In Anabatic::AutoSegments collections, change the type of all the flags that where in "unsigned int" (32 bits) to Flags (uint64_t) as there is now more than 32 flags for functions. * New: In Ababatic::Constants, added new flag Flags::WithPerpands, which makes the number of flags tip over 32 bits, thus making mandatory to uses Flags and not unsigned int. * New: In Anabatic::AutoSegments_Perpandiculars, manage a new flag Flags::WithDoglegs to allow to propagate through global segments that are connecteds via doglegs on local segments. Meaning that there is a good chance that they could be aligned. Slighly change the way we propagate on aligned segments: no longer check for VTee or HTee, but only for same direction and layer as master. * New: In Anabatic & Katana, replace all the "int", "long" and their variants by the less implementation ambiguous "int32_t", "int64_t" (and variant). This should help to better detect bit trucation in flags. Use the type to give a hint about the flags kind: - Type "Flags", for flags shared among Anabatic & Katana functions/methods (may also appear in some objects states). - Type "uint32_t" for flags belonging to an object internal state of from Hurricane functions flags (those should be grouped in a Flag subclass in a perfect world).
2017-05-16 07:53:33 -05:00
for ( AutoSegment* segment : const_cast<AutoSegment*>(this)->getAligneds() ) {
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
if (segment->isTerminal()) return true;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
}
}
return false;
}
DbU::Unit AutoSegment::getPPitch () const
{
unsigned int depth = getDepth();
In Anabatic/Katana, add support for VH gauges (real technos). * Change: In Anabatic::AutoContactTerminal::getNativeConstraintBox(), when the anchor is a RoutingPad (which must be always the case), perform the true computation of it's position based on the segment occurrence. It is a important change, previously the area was in fact the "center line" of the connector while now it is really an area (mandatory for "half-offgrid" terminals of real technologies). The change is not complete yet, the area should be shrinked by the half size of a VIA, because the area applies to the center coordinate of the VIA (to be done quickly). * Bug: In Anabatic::AutoContactTurn::updateTopology(), when a dogleg is created (restore connexity after a layer change) the layer of the VIA, based on the segments it connects to must be re-computed *after* the dogleg has been made. * Change: In all files of Anabatic, when comparing two layers, no longer use the Layer pointer itself, but the layer mask. This allow a transparent management of both real and symbolic layers (which do share the same mask). Real metal layers (not VIAs) will be BasicLayer and symbolic metal layers will be RegularLayer. * New: Anabatic::Configuration::selectRpComponent(), select the best RoutingPad component for metal1 terminals. Look for the metal1 component with the biggest accessibility on-grid. RoutingPad using other metals are left untoucheds. * New: New function Anabatic::Vertex::getNeighbor(Edge*) to get the neighbor Vertex through an Edge*. This method allows to write clearer code as we no longer need to access the neighbor through the underlying GCell. Also add proxies for GCell methods in Vertex. * Bug: In Anabatic::Dijkstra::_toSources(), in the ripup stage, when a component with multiples vertexes is reached *and* two of it's vertexes are reached *at the same time* (one from which we backtrack and one still in the queue) extraneous edges may be created by _materialize(). Case occurs on snx/c35b4, "abc_5360_n903_1". To solve this, Dijkstra::_toSource() is modificated, the "from" edges of the newly reacheds vertexes are reset to NULL, *except* for the one we will be backtracking from. That is, the one given in the source argument. * Change: In Anabatic::NetBuilder class, put the various Hooks and RoutingPad sorting functions as class ones. * Bug: In AutoSegment::setLayer(), raise the SegInvalidatedFayer flag. This unset flag was causing AutoContactTurn::updateTopology() to not work as expected and making gaps, this was the cause of the last remaining warnings about layer connexity.
2018-01-06 09:55:53 -06:00
DbU::Unit topPPitch = Session::getPitch( depth + ( ((_flags & SegSpinTop ) and (depth+1 < Session::getDepth())) ? 1 : 0) );
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
DbU::Unit bottomPPitch = Session::getPitch( depth - ( ((_flags & SegSpinBottom) and (depth > 0))? 1 : 0) );
return std::max( topPPitch, bottomPPitch );
}
Basic support for FreePDK 45 completed. * New: In Commons, inspector support for std::pair<T,U>. * New: In Hurricane::Layer, ContactLayer & ViaLayer, support for non square VIAs. The hole (cut) remains square, but the various metal extensions can now be different in X and Y. The ::getEnclosure() method now takes a flag EnclosureH / EnclosureV. * New: In Hurricane::DbU, inspector support for: std::pair<DbU::Unit,DbU::Unit> std::array<DbU::Unit,3> Must be defined here as DbU do not exists yet in Commons.h * Bug: In Hurricane::Interval::getSize(), when the interval is "full span", do not return the difference between min and max, but directly DbU::Max. (the previous result was -1 !) * New: In CRL Core Python/Technology.py, support for non square VIAs in the configuration files. Applied to FreePDK 45. * New: In CRL::RoutingGauge, added a "symbolic" flag to tell if a gauge is for symbolic layout or not. Exported to Python. * New: In Anabatic::AutoHorizontal::updatePosition(), differentiated computation for soure or target taking account of the VIA extension in the right segment metal (due to non-square VIAs). * Change: In Anabatic::AutoHorizontal::_makeDogleg(), the dogleg is UP for HV gauges and DOWN for VH. * New: In Anabatic::AutoSegment::_initialize(), create a cache of the various extension length for each layer (viaToTop, viaToBottom, viaToSame). New implementation of getExtensionCap() using the previous cached extension table. See updatePositions(). New static functions to access the extension cache in the header: getViaTotopCap() ... * Change: In Anabatic::AutoSegment, in various update methods, updateOrient() must always be called *before* updatePositions() as extensions are dependant on source/target. * New: In Anabatic::AutoSegment::getEndAxes() compute the position of the first source and last target position (center/axes) on an *aligned* set of segments. * New: In Anabatic::AutoSegment, add a new state flag SegAxisFixed to signal segments that can be put on only one track. Specific case to VH gauge for a M1 vertical terminal with a M2 vertical segment. The M2 is effectively bound to the M1 axis position. * Bug: In Anabatic::NetBuilderVH::_do_xG_xM1_xM3(), in case of E/W global and only one RoutingPad the connexion to the RoutingPad was duplicated. It was valid, but totally stupid. * Bug: In Anabatic::Session::_canonize(), for an aligned segment set, intersect the user constraints from all segments instead of only considering the canonical one. Issue a warning about too tight constraints only for symbolic gauges. It may be correct for the real ones. * New: In Katata::DataNegociate::update(), more accurate computation of the perpandicular free interval. Use segment extension cap calculation. Create a special case for fixed axis segments allowing them to find alternative free interval, try under source and under target as they are likely to be draggable segments. * Change: In Katana::Manipulator::relax(), use the extension cap value to compute the axis of the perpandicular segemnts. * Change: In Katana::Manipulator::moveUp(), now move up the whole set of aligned segments instead of just the canonical one. * Change: In Katana::NegociateWindow::loadRoutingPads(), more accurate TrackMarkers insertions for fixed terminals. * New: In Katana::RoutingEvent::Key::Compare::operator(), segments with fixed axis are processed prior to any others. * New: In Katana::RoutingEventLoop, store segment pointers instead of ids to generate more accurate error messages. * Change: In Katana::RoutingPlane::create(), perform local track assignment only for HV gauges. * Change: In Katana::SegmentFsm::_slackenLocal(), add a "dragMinimize" step in the automaton. Mutliple states transitions can occurs in a row if an action fails. * New: In Katana::Session::_toIntervalAxis(), normalize interval bounds so they are on track positions (by shrinking the interval). * Bug: In Katana::TrackMarker CTOR, the weigh computation was wrong.
2018-02-17 13:27:38 -06:00
DbU::Unit AutoSegment::getExtensionCap ( Flags flags ) const
TrackCost clean support for symmetric and wide segments. * New: In Katana::TrackCost, the TrackElement and it's optional symmetric are now kept as attribute of a TrackCost. The cost is completly computed inside the constructor. TrackCost now support any mix of symmetric event and wide segments. The cost is now computed by adding directly to the current one instead of creating secondaries that are merged afterwards. As a consequence, remove all copy construction and merge capabilities. All the various methods used to compute the cost are renamed "addOverlapcost()" in all the various related objects. As a reminder, the overal cost method call is as follow: 1. TrackCost constructor on a TrackElement. 2. Call TrackElement::addOverlapcost() 3. For all Track under the TrackElement, call Track::addOverlapCost() 4. For all other TrackElement intersecting with the overlap interval call: TrackElement::incOverlapCost() 5. The callback overlap function for segments is called (defined in NegociateWidow). Don't confuse: - TrackElement::addOverlapCost(), which compute the cost of inserting the segment inside a track (or a set of). - TrackElement::incOverlapCost(), which compute the cost of overlaping with this already inserted segment. It is the other way around of the previous one. * Change: In Katana::SegmentFsm, use a vector of pointer to TrackCost instead of an object to avoid copy construction.
2017-06-10 05:27:25 -05:00
{
Basic support for FreePDK 45 completed. * New: In Commons, inspector support for std::pair<T,U>. * New: In Hurricane::Layer, ContactLayer & ViaLayer, support for non square VIAs. The hole (cut) remains square, but the various metal extensions can now be different in X and Y. The ::getEnclosure() method now takes a flag EnclosureH / EnclosureV. * New: In Hurricane::DbU, inspector support for: std::pair<DbU::Unit,DbU::Unit> std::array<DbU::Unit,3> Must be defined here as DbU do not exists yet in Commons.h * Bug: In Hurricane::Interval::getSize(), when the interval is "full span", do not return the difference between min and max, but directly DbU::Max. (the previous result was -1 !) * New: In CRL Core Python/Technology.py, support for non square VIAs in the configuration files. Applied to FreePDK 45. * New: In CRL::RoutingGauge, added a "symbolic" flag to tell if a gauge is for symbolic layout or not. Exported to Python. * New: In Anabatic::AutoHorizontal::updatePosition(), differentiated computation for soure or target taking account of the VIA extension in the right segment metal (due to non-square VIAs). * Change: In Anabatic::AutoHorizontal::_makeDogleg(), the dogleg is UP for HV gauges and DOWN for VH. * New: In Anabatic::AutoSegment::_initialize(), create a cache of the various extension length for each layer (viaToTop, viaToBottom, viaToSame). New implementation of getExtensionCap() using the previous cached extension table. See updatePositions(). New static functions to access the extension cache in the header: getViaTotopCap() ... * Change: In Anabatic::AutoSegment, in various update methods, updateOrient() must always be called *before* updatePositions() as extensions are dependant on source/target. * New: In Anabatic::AutoSegment::getEndAxes() compute the position of the first source and last target position (center/axes) on an *aligned* set of segments. * New: In Anabatic::AutoSegment, add a new state flag SegAxisFixed to signal segments that can be put on only one track. Specific case to VH gauge for a M1 vertical terminal with a M2 vertical segment. The M2 is effectively bound to the M1 axis position. * Bug: In Anabatic::NetBuilderVH::_do_xG_xM1_xM3(), in case of E/W global and only one RoutingPad the connexion to the RoutingPad was duplicated. It was valid, but totally stupid. * Bug: In Anabatic::Session::_canonize(), for an aligned segment set, intersect the user constraints from all segments instead of only considering the canonical one. Issue a warning about too tight constraints only for symbolic gauges. It may be correct for the real ones. * New: In Katata::DataNegociate::update(), more accurate computation of the perpandicular free interval. Use segment extension cap calculation. Create a special case for fixed axis segments allowing them to find alternative free interval, try under source and under target as they are likely to be draggable segments. * Change: In Katana::Manipulator::relax(), use the extension cap value to compute the axis of the perpandicular segemnts. * Change: In Katana::Manipulator::moveUp(), now move up the whole set of aligned segments instead of just the canonical one. * Change: In Katana::NegociateWindow::loadRoutingPads(), more accurate TrackMarkers insertions for fixed terminals. * New: In Katana::RoutingEvent::Key::Compare::operator(), segments with fixed axis are processed prior to any others. * New: In Katana::RoutingEventLoop, store segment pointers instead of ids to generate more accurate error messages. * Change: In Katana::RoutingPlane::create(), perform local track assignment only for HV gauges. * Change: In Katana::SegmentFsm::_slackenLocal(), add a "dragMinimize" step in the automaton. Mutliple states transitions can occurs in a row if an action fails. * New: In Katana::Session::_toIntervalAxis(), normalize interval bounds so they are on track positions (by shrinking the interval). * Bug: In Katana::TrackMarker CTOR, the weigh computation was wrong.
2018-02-17 13:27:38 -06:00
size_t depth = Session::getLayerDepth( getLayer() );
DbU::Unit cap = 0;
if (flags & Flags::Source) {
if (getFlags() & SegSourceTop ) cap = getViaToTopCap (depth);
else if (getFlags() & SegSourceBottom) cap = getViaToBottomCap(depth);
else cap = getViaToSameCap (depth);
DRC correct on Arlet6505 / TSMC C180. Integrate new features and bug fixes so the Arlet 6502 benchs successfully passes real DRC with reference industrial tools. Short summary: * Manage minimum area for VIAs in Katana::Tracks. * Allow different wire width for wires perpandicular to the prefered routing direction. * StackedVIAs used in the clock tree no longer assume an uniform routing grid (same offset & pitch all the way up). * Some hard-coded patches in PowerRails for FlexLib. * New: In CRL/symbolic/cmos/kite.py & cmos45/kite.py, update the RoutingLayerGauges by adding the new PWireWidth parameter. Always zero in case of symbolic layout (too fine tuning). * New: In CRL::RoutingGauge, add accessor to PWireWidth parameter. Modify the clone method. * New: In CRL::RoutingLayerGauge, add new parameter "PWireWidth" to give the width of a wire when it not drawn in the prefered routing direction. If it is set to zero, the normal width is used. * New: In CRL::PyRoutingGauge, export the updated constructor interface. It is *not* backward compatible, one must add the PWireWidth parameter in the various kite.py configuration files (in etc/). * Change: In AnabaticEngine::_gutAnabatic(), disable the minimum area detection mechanism, replaced by a more complete one in Katana::Track. Left commented out for now, but will be removed in the future. * Change: In Anabatic::AutoContact::updateLayer(), now systematically calls setLayerAndWidth() to potentially resize the VIAs. This is needed in real mode as VIAs are *not* macro-generated but have their real final size. * Change: In Anabatic::AutoContact::setLayerAndWidth(), select the width and height of the contact using the gauge wire width *and* perpandicular *wire width*. * Change: In Anabatic::AutoSegment::_initialize(), the "VIA to same cap" to PWireWidth/2, this will be the size of the VIA in the non-preferred direction at the end cap (non-square in real mode). * Change: In Anabatic::AutoSegment::getExtensionCap(), makes different cases for symbolic and real. Use raw length in real, add half the wire width in symbolic. Add a flag to get the extension cap *only*, not increased of half the minimal spacing. * Change: In Anabatic::AutoSegment::bloatStackedStrap(), enhanced, but finally unused... * New: In Anabatic::AutoSegment::create(), use the PWireWidth when the segment is not in the preferred routing direction (and of minimal width). * New: In Anabatic::Configuration, add new getPWirewidth(), DPHorizontalWidth() and DPVerticalWidth() accessors. * Change: In AnabaticEngine::setupPreRouteds(), skip components in in "cut" material. We are only interested in objects containing some metal (happens in real mode when VIAs cuts are really there). * New: In Katana::PowerRailsPlanes::Rail::doLayout(), add an hard-coded patch that artificially enlarge the *wide wire* so the spacing for wide wire is enforced. For now, two pitches on each side for "FlexLib" gauge. * New: In Katana::Track, add support to find and correct small wire chunks so they respect the minimum area rules. Two helper functions: * ::hasSameLayerTurn(), to find if a a TrackElement as non-zero length perpandicular is same layer connected to it. * ::toFoundryGrid(), to ensure that all coordinates will be on the foundry grid (may move in a more shared location). * ::expandToMinArea(), try to expand, *in the routing direction* the too small wire so it respect the minimal area. Check for the free space in the track. Track::minExpandArea() go through all the TrackElements in the track to look for too small ones and correct them. * Change: In Katana::RoutingPlane, add an accessor to get the tracks. * New: In KatanaEngine::finalizeLayout(), add a post-treatment to find for minimal area violations. * Change: In cumulus/plugins.block.configuration.GaugeConf, add a routingBb attribute that will serve as a common reference to all the functions calculation track positions. We must not have two different reference for the core and the corona. The reference is always the corona when we working on a complete chip. * New: In cumulus/plugins.block.configuration.GaugeConf.getTrack(), Simplified and more reliable way of getting tracks positions. Use the routingBb. * New: In cumulus/plugins.block.configuration.GaugeConf.rpAccess(), Make use of getTrack() to get every metal strap on the right X/Y position. * New: In cumulus/plugins.block.configuration.GaugeConf.expandMinArea(), As those wires are left alone by the router, it is our responsability to abide by the minimal area rule here. Hence the code duplication from the router (bad). Mainly wires made for the clock tree, I mean. * Bug: In cumulus/plugins.chip.configuration.ChipConf.setupICore(), the core instance must be placed on the GCell grid, defined by the slice height (X *and* Y). * Bug: In cumulus/plugins.chip.corona.Builder(), forgot to use bigvia for the corners of the inner ring. * Bug: In cumulus/plugins.chip.pads.corona._createCoreWire(), hard-coded patch for LibreSOCIO, the power/ground connectors toward the core are too wide and can create DRC errors when put side by side. Shrink them by the minimal distance.
2020-11-23 16:07:15 -06:00
cdebug_log(150,0) << "getExtensionCap(): flags:" << getFlags()
Basic support for FreePDK 45 completed. * New: In Commons, inspector support for std::pair<T,U>. * New: In Hurricane::Layer, ContactLayer & ViaLayer, support for non square VIAs. The hole (cut) remains square, but the various metal extensions can now be different in X and Y. The ::getEnclosure() method now takes a flag EnclosureH / EnclosureV. * New: In Hurricane::DbU, inspector support for: std::pair<DbU::Unit,DbU::Unit> std::array<DbU::Unit,3> Must be defined here as DbU do not exists yet in Commons.h * Bug: In Hurricane::Interval::getSize(), when the interval is "full span", do not return the difference between min and max, but directly DbU::Max. (the previous result was -1 !) * New: In CRL Core Python/Technology.py, support for non square VIAs in the configuration files. Applied to FreePDK 45. * New: In CRL::RoutingGauge, added a "symbolic" flag to tell if a gauge is for symbolic layout or not. Exported to Python. * New: In Anabatic::AutoHorizontal::updatePosition(), differentiated computation for soure or target taking account of the VIA extension in the right segment metal (due to non-square VIAs). * Change: In Anabatic::AutoHorizontal::_makeDogleg(), the dogleg is UP for HV gauges and DOWN for VH. * New: In Anabatic::AutoSegment::_initialize(), create a cache of the various extension length for each layer (viaToTop, viaToBottom, viaToSame). New implementation of getExtensionCap() using the previous cached extension table. See updatePositions(). New static functions to access the extension cache in the header: getViaTotopCap() ... * Change: In Anabatic::AutoSegment, in various update methods, updateOrient() must always be called *before* updatePositions() as extensions are dependant on source/target. * New: In Anabatic::AutoSegment::getEndAxes() compute the position of the first source and last target position (center/axes) on an *aligned* set of segments. * New: In Anabatic::AutoSegment, add a new state flag SegAxisFixed to signal segments that can be put on only one track. Specific case to VH gauge for a M1 vertical terminal with a M2 vertical segment. The M2 is effectively bound to the M1 axis position. * Bug: In Anabatic::NetBuilderVH::_do_xG_xM1_xM3(), in case of E/W global and only one RoutingPad the connexion to the RoutingPad was duplicated. It was valid, but totally stupid. * Bug: In Anabatic::Session::_canonize(), for an aligned segment set, intersect the user constraints from all segments instead of only considering the canonical one. Issue a warning about too tight constraints only for symbolic gauges. It may be correct for the real ones. * New: In Katata::DataNegociate::update(), more accurate computation of the perpandicular free interval. Use segment extension cap calculation. Create a special case for fixed axis segments allowing them to find alternative free interval, try under source and under target as they are likely to be draggable segments. * Change: In Katana::Manipulator::relax(), use the extension cap value to compute the axis of the perpandicular segemnts. * Change: In Katana::Manipulator::moveUp(), now move up the whole set of aligned segments instead of just the canonical one. * Change: In Katana::NegociateWindow::loadRoutingPads(), more accurate TrackMarkers insertions for fixed terminals. * New: In Katana::RoutingEvent::Key::Compare::operator(), segments with fixed axis are processed prior to any others. * New: In Katana::RoutingEventLoop, store segment pointers instead of ids to generate more accurate error messages. * Change: In Katana::RoutingPlane::create(), perform local track assignment only for HV gauges. * Change: In Katana::SegmentFsm::_slackenLocal(), add a "dragMinimize" step in the automaton. Mutliple states transitions can occurs in a row if an action fails. * New: In Katana::Session::_toIntervalAxis(), normalize interval bounds so they are on track positions (by shrinking the interval). * Bug: In Katana::TrackMarker CTOR, the weigh computation was wrong.
2018-02-17 13:27:38 -06:00
<< " VIA cap:" << DbU::getValueString(cap)
DRC correct on Arlet6505 / TSMC C180. Integrate new features and bug fixes so the Arlet 6502 benchs successfully passes real DRC with reference industrial tools. Short summary: * Manage minimum area for VIAs in Katana::Tracks. * Allow different wire width for wires perpandicular to the prefered routing direction. * StackedVIAs used in the clock tree no longer assume an uniform routing grid (same offset & pitch all the way up). * Some hard-coded patches in PowerRails for FlexLib. * New: In CRL/symbolic/cmos/kite.py & cmos45/kite.py, update the RoutingLayerGauges by adding the new PWireWidth parameter. Always zero in case of symbolic layout (too fine tuning). * New: In CRL::RoutingGauge, add accessor to PWireWidth parameter. Modify the clone method. * New: In CRL::RoutingLayerGauge, add new parameter "PWireWidth" to give the width of a wire when it not drawn in the prefered routing direction. If it is set to zero, the normal width is used. * New: In CRL::PyRoutingGauge, export the updated constructor interface. It is *not* backward compatible, one must add the PWireWidth parameter in the various kite.py configuration files (in etc/). * Change: In AnabaticEngine::_gutAnabatic(), disable the minimum area detection mechanism, replaced by a more complete one in Katana::Track. Left commented out for now, but will be removed in the future. * Change: In Anabatic::AutoContact::updateLayer(), now systematically calls setLayerAndWidth() to potentially resize the VIAs. This is needed in real mode as VIAs are *not* macro-generated but have their real final size. * Change: In Anabatic::AutoContact::setLayerAndWidth(), select the width and height of the contact using the gauge wire width *and* perpandicular *wire width*. * Change: In Anabatic::AutoSegment::_initialize(), the "VIA to same cap" to PWireWidth/2, this will be the size of the VIA in the non-preferred direction at the end cap (non-square in real mode). * Change: In Anabatic::AutoSegment::getExtensionCap(), makes different cases for symbolic and real. Use raw length in real, add half the wire width in symbolic. Add a flag to get the extension cap *only*, not increased of half the minimal spacing. * Change: In Anabatic::AutoSegment::bloatStackedStrap(), enhanced, but finally unused... * New: In Anabatic::AutoSegment::create(), use the PWireWidth when the segment is not in the preferred routing direction (and of minimal width). * New: In Anabatic::Configuration, add new getPWirewidth(), DPHorizontalWidth() and DPVerticalWidth() accessors. * Change: In AnabaticEngine::setupPreRouteds(), skip components in in "cut" material. We are only interested in objects containing some metal (happens in real mode when VIAs cuts are really there). * New: In Katana::PowerRailsPlanes::Rail::doLayout(), add an hard-coded patch that artificially enlarge the *wide wire* so the spacing for wide wire is enforced. For now, two pitches on each side for "FlexLib" gauge. * New: In Katana::Track, add support to find and correct small wire chunks so they respect the minimum area rules. Two helper functions: * ::hasSameLayerTurn(), to find if a a TrackElement as non-zero length perpandicular is same layer connected to it. * ::toFoundryGrid(), to ensure that all coordinates will be on the foundry grid (may move in a more shared location). * ::expandToMinArea(), try to expand, *in the routing direction* the too small wire so it respect the minimal area. Check for the free space in the track. Track::minExpandArea() go through all the TrackElements in the track to look for too small ones and correct them. * Change: In Katana::RoutingPlane, add an accessor to get the tracks. * New: In KatanaEngine::finalizeLayout(), add a post-treatment to find for minimal area violations. * Change: In cumulus/plugins.block.configuration.GaugeConf, add a routingBb attribute that will serve as a common reference to all the functions calculation track positions. We must not have two different reference for the core and the corona. The reference is always the corona when we working on a complete chip. * New: In cumulus/plugins.block.configuration.GaugeConf.getTrack(), Simplified and more reliable way of getting tracks positions. Use the routingBb. * New: In cumulus/plugins.block.configuration.GaugeConf.rpAccess(), Make use of getTrack() to get every metal strap on the right X/Y position. * New: In cumulus/plugins.block.configuration.GaugeConf.expandMinArea(), As those wires are left alone by the router, it is our responsability to abide by the minimal area rule here. Hence the code duplication from the router (bad). Mainly wires made for the clock tree, I mean. * Bug: In cumulus/plugins.chip.configuration.ChipConf.setupICore(), the core instance must be placed on the GCell grid, defined by the slice height (X *and* Y). * Bug: In cumulus/plugins.chip.corona.Builder(), forgot to use bigvia for the corners of the inner ring. * Bug: In cumulus/plugins.chip.pads.corona._createCoreWire(), hard-coded patch for LibreSOCIO, the power/ground connectors toward the core are too wide and can create DRC errors when put side by side. Shrink them by the minimal distance.
2020-11-23 16:07:15 -06:00
<< " t:" << (getFlags() & SegSourceBottom)
<< " b:" << (getFlags() & SegSourceTop)
Basic support for FreePDK 45 completed. * New: In Commons, inspector support for std::pair<T,U>. * New: In Hurricane::Layer, ContactLayer & ViaLayer, support for non square VIAs. The hole (cut) remains square, but the various metal extensions can now be different in X and Y. The ::getEnclosure() method now takes a flag EnclosureH / EnclosureV. * New: In Hurricane::DbU, inspector support for: std::pair<DbU::Unit,DbU::Unit> std::array<DbU::Unit,3> Must be defined here as DbU do not exists yet in Commons.h * Bug: In Hurricane::Interval::getSize(), when the interval is "full span", do not return the difference between min and max, but directly DbU::Max. (the previous result was -1 !) * New: In CRL Core Python/Technology.py, support for non square VIAs in the configuration files. Applied to FreePDK 45. * New: In CRL::RoutingGauge, added a "symbolic" flag to tell if a gauge is for symbolic layout or not. Exported to Python. * New: In Anabatic::AutoHorizontal::updatePosition(), differentiated computation for soure or target taking account of the VIA extension in the right segment metal (due to non-square VIAs). * Change: In Anabatic::AutoHorizontal::_makeDogleg(), the dogleg is UP for HV gauges and DOWN for VH. * New: In Anabatic::AutoSegment::_initialize(), create a cache of the various extension length for each layer (viaToTop, viaToBottom, viaToSame). New implementation of getExtensionCap() using the previous cached extension table. See updatePositions(). New static functions to access the extension cache in the header: getViaTotopCap() ... * Change: In Anabatic::AutoSegment, in various update methods, updateOrient() must always be called *before* updatePositions() as extensions are dependant on source/target. * New: In Anabatic::AutoSegment::getEndAxes() compute the position of the first source and last target position (center/axes) on an *aligned* set of segments. * New: In Anabatic::AutoSegment, add a new state flag SegAxisFixed to signal segments that can be put on only one track. Specific case to VH gauge for a M1 vertical terminal with a M2 vertical segment. The M2 is effectively bound to the M1 axis position. * Bug: In Anabatic::NetBuilderVH::_do_xG_xM1_xM3(), in case of E/W global and only one RoutingPad the connexion to the RoutingPad was duplicated. It was valid, but totally stupid. * Bug: In Anabatic::Session::_canonize(), for an aligned segment set, intersect the user constraints from all segments instead of only considering the canonical one. Issue a warning about too tight constraints only for symbolic gauges. It may be correct for the real ones. * New: In Katata::DataNegociate::update(), more accurate computation of the perpandicular free interval. Use segment extension cap calculation. Create a special case for fixed axis segments allowing them to find alternative free interval, try under source and under target as they are likely to be draggable segments. * Change: In Katana::Manipulator::relax(), use the extension cap value to compute the axis of the perpandicular segemnts. * Change: In Katana::Manipulator::moveUp(), now move up the whole set of aligned segments instead of just the canonical one. * Change: In Katana::NegociateWindow::loadRoutingPads(), more accurate TrackMarkers insertions for fixed terminals. * New: In Katana::RoutingEvent::Key::Compare::operator(), segments with fixed axis are processed prior to any others. * New: In Katana::RoutingEventLoop, store segment pointers instead of ids to generate more accurate error messages. * Change: In Katana::RoutingPlane::create(), perform local track assignment only for HV gauges. * Change: In Katana::SegmentFsm::_slackenLocal(), add a "dragMinimize" step in the automaton. Mutliple states transitions can occurs in a row if an action fails. * New: In Katana::Session::_toIntervalAxis(), normalize interval bounds so they are on track positions (by shrinking the interval). * Bug: In Katana::TrackMarker CTOR, the weigh computation was wrong.
2018-02-17 13:27:38 -06:00
<< endl;
}
if (flags & Flags::Target) {
if (getFlags() & SegTargetTop ) cap = getViaToTopCap (depth);
else if (getFlags() & SegTargetBottom) cap = getViaToBottomCap(depth);
else cap = getViaToSameCap (depth);
}
DRC correct on Arlet6505 / TSMC C180. Integrate new features and bug fixes so the Arlet 6502 benchs successfully passes real DRC with reference industrial tools. Short summary: * Manage minimum area for VIAs in Katana::Tracks. * Allow different wire width for wires perpandicular to the prefered routing direction. * StackedVIAs used in the clock tree no longer assume an uniform routing grid (same offset & pitch all the way up). * Some hard-coded patches in PowerRails for FlexLib. * New: In CRL/symbolic/cmos/kite.py & cmos45/kite.py, update the RoutingLayerGauges by adding the new PWireWidth parameter. Always zero in case of symbolic layout (too fine tuning). * New: In CRL::RoutingGauge, add accessor to PWireWidth parameter. Modify the clone method. * New: In CRL::RoutingLayerGauge, add new parameter "PWireWidth" to give the width of a wire when it not drawn in the prefered routing direction. If it is set to zero, the normal width is used. * New: In CRL::PyRoutingGauge, export the updated constructor interface. It is *not* backward compatible, one must add the PWireWidth parameter in the various kite.py configuration files (in etc/). * Change: In AnabaticEngine::_gutAnabatic(), disable the minimum area detection mechanism, replaced by a more complete one in Katana::Track. Left commented out for now, but will be removed in the future. * Change: In Anabatic::AutoContact::updateLayer(), now systematically calls setLayerAndWidth() to potentially resize the VIAs. This is needed in real mode as VIAs are *not* macro-generated but have their real final size. * Change: In Anabatic::AutoContact::setLayerAndWidth(), select the width and height of the contact using the gauge wire width *and* perpandicular *wire width*. * Change: In Anabatic::AutoSegment::_initialize(), the "VIA to same cap" to PWireWidth/2, this will be the size of the VIA in the non-preferred direction at the end cap (non-square in real mode). * Change: In Anabatic::AutoSegment::getExtensionCap(), makes different cases for symbolic and real. Use raw length in real, add half the wire width in symbolic. Add a flag to get the extension cap *only*, not increased of half the minimal spacing. * Change: In Anabatic::AutoSegment::bloatStackedStrap(), enhanced, but finally unused... * New: In Anabatic::AutoSegment::create(), use the PWireWidth when the segment is not in the preferred routing direction (and of minimal width). * New: In Anabatic::Configuration, add new getPWirewidth(), DPHorizontalWidth() and DPVerticalWidth() accessors. * Change: In AnabaticEngine::setupPreRouteds(), skip components in in "cut" material. We are only interested in objects containing some metal (happens in real mode when VIAs cuts are really there). * New: In Katana::PowerRailsPlanes::Rail::doLayout(), add an hard-coded patch that artificially enlarge the *wide wire* so the spacing for wide wire is enforced. For now, two pitches on each side for "FlexLib" gauge. * New: In Katana::Track, add support to find and correct small wire chunks so they respect the minimum area rules. Two helper functions: * ::hasSameLayerTurn(), to find if a a TrackElement as non-zero length perpandicular is same layer connected to it. * ::toFoundryGrid(), to ensure that all coordinates will be on the foundry grid (may move in a more shared location). * ::expandToMinArea(), try to expand, *in the routing direction* the too small wire so it respect the minimal area. Check for the free space in the track. Track::minExpandArea() go through all the TrackElements in the track to look for too small ones and correct them. * Change: In Katana::RoutingPlane, add an accessor to get the tracks. * New: In KatanaEngine::finalizeLayout(), add a post-treatment to find for minimal area violations. * Change: In cumulus/plugins.block.configuration.GaugeConf, add a routingBb attribute that will serve as a common reference to all the functions calculation track positions. We must not have two different reference for the core and the corona. The reference is always the corona when we working on a complete chip. * New: In cumulus/plugins.block.configuration.GaugeConf.getTrack(), Simplified and more reliable way of getting tracks positions. Use the routingBb. * New: In cumulus/plugins.block.configuration.GaugeConf.rpAccess(), Make use of getTrack() to get every metal strap on the right X/Y position. * New: In cumulus/plugins.block.configuration.GaugeConf.expandMinArea(), As those wires are left alone by the router, it is our responsability to abide by the minimal area rule here. Hence the code duplication from the router (bad). Mainly wires made for the clock tree, I mean. * Bug: In cumulus/plugins.chip.configuration.ChipConf.setupICore(), the core instance must be placed on the GCell grid, defined by the slice height (X *and* Y). * Bug: In cumulus/plugins.chip.corona.Builder(), forgot to use bigvia for the corners of the inner ring. * Bug: In cumulus/plugins.chip.pads.corona._createCoreWire(), hard-coded patch for LibreSOCIO, the power/ground connectors toward the core are too wide and can create DRC errors when put side by side. Shrink them by the minimal distance.
2020-11-23 16:07:15 -06:00
if (getLayer()->isSymbolic() and (cap < getWidth()/2)) cap = getWidth()/2;
if (not (flags & Flags::LayerCapOnly)) cap += getLayer()->getMinimalSpacing()/2;
return cap;
TrackCost clean support for symmetric and wide segments. * New: In Katana::TrackCost, the TrackElement and it's optional symmetric are now kept as attribute of a TrackCost. The cost is completly computed inside the constructor. TrackCost now support any mix of symmetric event and wide segments. The cost is now computed by adding directly to the current one instead of creating secondaries that are merged afterwards. As a consequence, remove all copy construction and merge capabilities. All the various methods used to compute the cost are renamed "addOverlapcost()" in all the various related objects. As a reminder, the overal cost method call is as follow: 1. TrackCost constructor on a TrackElement. 2. Call TrackElement::addOverlapcost() 3. For all Track under the TrackElement, call Track::addOverlapCost() 4. For all other TrackElement intersecting with the overlap interval call: TrackElement::incOverlapCost() 5. The callback overlap function for segments is called (defined in NegociateWidow). Don't confuse: - TrackElement::addOverlapCost(), which compute the cost of inserting the segment inside a track (or a set of). - TrackElement::incOverlapCost(), which compute the cost of overlaping with this already inserted segment. It is the other way around of the previous one. * Change: In Katana::SegmentFsm, use a vector of pointer to TrackCost instead of an object to avoid copy construction.
2017-06-10 05:27:25 -05:00
}
Added support for 2-Metal block routing in Anabatic & Katana. * New: In AnabaticEngine::invalidateRoutingPads() this method is a temporary workaround for a Hurricane problems. When an instance is moved, the RoutingPads that use it must be moved accordingly, but they are not invalidated so they stay in the wrong QuadTree. New method ::_resizeMatrix() to be called when the associated Cell is resized. * Bug: In AutoHorizontal::getConstraints() and AutoVertical::getConstraints(), the *target* constraints where never merged. * Change: In AutoHorizontal::getCells() and AutoVertical::getGCells(), now return a boolean to tell if it was ok (must not encounter a NULL GCell while progessing from source to target). * New: In Anabatic::Configuration and Anabatic:Session, create new methods: - getDHorizontalLayer() - getDhorizontalDepth() - getDHorizontalWidth() - getDHorizontalPitch() And so on for Vertical and Contact. They supply depth-independant informations about the H/V layers to build the initial detailed routing. The AutoSegment::create() methods have been modificated accordingly. * New: In Anabatic::GCell, add two new types "StdCellRow" and "ChannelRow" for implementing 2-Metal blocks. Rename the GCell::setXY() method in GCell::setSouthWestCorner(), move the contents of GCell::updateContactsPosition() into it and suppress it. WARNING: In case of a GCell shrink this may cause problems. But for now we only expand... New method GCell::getNetCount() to count the number of Net going though the GCell. * Change: In Anabatic::Edge, add specific support for capacity of 2-Metal routing channels. * Change: In Anabatic::Dijsktra various methods, replace the "gcell->isMatrix()" calls by "not gcell->isAnalog()". Add more check so that the methods pertaining to the analog routing (GRData) are not called in digital mode. * New: In Anabatic::Dijkstra::materialize(), add support for 2-Metal specific cases. That is, always break in case of vertical pass-through or U-turn. The global routing must always be broken in H-Channel. * New: In Anabatic::GCell & Anabatic::Edge, make use of the Session mechanism to ensure the revalidation. The "::revalidate()" method is then moved as "::materialize()" (overload of Go) and "::_invalidate()" becomes "::invalidate()" * Change: In LoadGlobalRouting, cosmetic rename of SortHkByX in SortHookByX. * New: In GCellTopology, added support for building 2-Metal topologies. * ForkStack is now an object attribute as many methods do need it. * To push segments/hook on the stack, a new method "push()" is available. Perform NULL and fromHook checking. Can also setup _southWestContact or _northEastContact if it is the "from" edge. * N/S/E/W edges are now vector as in digital channel mode there can be more than one. * Added build topological build methods: - doRp_2m_Access() RoutingPad stem access. - _do_2m_1G_1M1() North or south access. - _do_2m_2G_1M1() North AND south access. - _do_2m_xG() H-Channel routing. * New: In Anabatic::Matrix, new ::resize() function, as Cell can be resizeds. * New: In Anabatic::Vertex, new static method ::getValueString() for a friendly text rendering. * New: In Katana::DigitalDistance, support for channel routing. * Change: In KatanaEngine::digitalSetup() and KatanaEngine::runGlobalrouter(), for channel routing, calls to setupPowerRails() and protectRoutingPads() must be called after the core block has been fully dimensionned. ::runGlobalrouter() contains the code tasked with the grid creation and channel sizing. * New: In KatanaEngine: Added support for core block, for 2-Metal routing. May be expanded for over-the-cell routing in the future. Added methods : - isDigitalMode() - isAnalogMode() - isMixedMode() - isChannelMode() - getBlock() / addBlock() - setupChannelMode() - createChannel() * New: In Katana, new class Block to manage core blocks and perform channel routing. * New: In Katana::Session, new convenience method "isOpen()".
2017-08-18 16:56:23 -05:00
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
DbU::Unit AutoSegment::getSlack () const
{
DbU::Unit constraintMin;
DbU::Unit constraintMax;
getConstraints( constraintMin, constraintMax );
return constraintMax - constraintMin;
}
DbU::Unit AutoSegment::getCost ( DbU::Unit axis ) const
{
DbU::Unit optimal = getOptimalMin();
if (axis < optimal) return optimal - axis;
optimal = getOptimalMax();
if (axis > optimal) return axis - optimal;
return 0;
}
AutoSegment* AutoSegment::getCanonical ( DbU::Unit& min, DbU::Unit& max )
{
cdebug_log(145,0) << "AutoSegment::getCanonical() - " << this << endl;
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
min = getSourcePosition();
max = getTargetPosition();
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
if (max < min) swap( min, max );
//cdebug_log(145,0) << "[" << DbU::getValueString(min) << " " << DbU::getValueString(max) << "]" << endl;
AutoSegment* canonical = this;
size_t canonicals = isCanonical();
size_t aligneds = 1;
DbU::Unit collapsedMin;
DbU::Unit collapsedMax;
if (not isNotAligned()) {
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
for ( AutoSegment* segment : getAligneds() ) {
if (segment->isCanonical()) {
canonical = segment;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
canonicals++;
}
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
collapsedMin = segment->getSourcePosition();
collapsedMax = segment->getTargetPosition();
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
if (collapsedMax < collapsedMin) swap( collapsedMin, collapsedMax );
if (collapsedMin < min) min = collapsedMin;
if (collapsedMax > max) max = collapsedMax;
aligneds++;
}
//cdebug_log(145,0) << "[" << DbU::getValueString(min) << " " << DbU::getValueString(max) << "]" << endl;
cdebug_log(145,0) << "Canonical: " << canonical << endl;
if ( (canonicals > 1) or ( not canonicals and (aligneds > 2) ) ) {
cerr << Bug("AutoSegment::getCanonical(): %p:%s"
"\n Bad canonization: %d canonicals out of %d collapseds."
, base(), _getString().c_str(), canonicals, aligneds ) << endl;
int count = 0;
cerr << " " << count++ << ": " << this << endl;
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
for ( AutoSegment* segment : getAligneds() )
cerr << " " << count++ << ": " << segment << endl;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
}
}
return canonical;
}
Basic support for FreePDK 45 completed. * New: In Commons, inspector support for std::pair<T,U>. * New: In Hurricane::Layer, ContactLayer & ViaLayer, support for non square VIAs. The hole (cut) remains square, but the various metal extensions can now be different in X and Y. The ::getEnclosure() method now takes a flag EnclosureH / EnclosureV. * New: In Hurricane::DbU, inspector support for: std::pair<DbU::Unit,DbU::Unit> std::array<DbU::Unit,3> Must be defined here as DbU do not exists yet in Commons.h * Bug: In Hurricane::Interval::getSize(), when the interval is "full span", do not return the difference between min and max, but directly DbU::Max. (the previous result was -1 !) * New: In CRL Core Python/Technology.py, support for non square VIAs in the configuration files. Applied to FreePDK 45. * New: In CRL::RoutingGauge, added a "symbolic" flag to tell if a gauge is for symbolic layout or not. Exported to Python. * New: In Anabatic::AutoHorizontal::updatePosition(), differentiated computation for soure or target taking account of the VIA extension in the right segment metal (due to non-square VIAs). * Change: In Anabatic::AutoHorizontal::_makeDogleg(), the dogleg is UP for HV gauges and DOWN for VH. * New: In Anabatic::AutoSegment::_initialize(), create a cache of the various extension length for each layer (viaToTop, viaToBottom, viaToSame). New implementation of getExtensionCap() using the previous cached extension table. See updatePositions(). New static functions to access the extension cache in the header: getViaTotopCap() ... * Change: In Anabatic::AutoSegment, in various update methods, updateOrient() must always be called *before* updatePositions() as extensions are dependant on source/target. * New: In Anabatic::AutoSegment::getEndAxes() compute the position of the first source and last target position (center/axes) on an *aligned* set of segments. * New: In Anabatic::AutoSegment, add a new state flag SegAxisFixed to signal segments that can be put on only one track. Specific case to VH gauge for a M1 vertical terminal with a M2 vertical segment. The M2 is effectively bound to the M1 axis position. * Bug: In Anabatic::NetBuilderVH::_do_xG_xM1_xM3(), in case of E/W global and only one RoutingPad the connexion to the RoutingPad was duplicated. It was valid, but totally stupid. * Bug: In Anabatic::Session::_canonize(), for an aligned segment set, intersect the user constraints from all segments instead of only considering the canonical one. Issue a warning about too tight constraints only for symbolic gauges. It may be correct for the real ones. * New: In Katata::DataNegociate::update(), more accurate computation of the perpandicular free interval. Use segment extension cap calculation. Create a special case for fixed axis segments allowing them to find alternative free interval, try under source and under target as they are likely to be draggable segments. * Change: In Katana::Manipulator::relax(), use the extension cap value to compute the axis of the perpandicular segemnts. * Change: In Katana::Manipulator::moveUp(), now move up the whole set of aligned segments instead of just the canonical one. * Change: In Katana::NegociateWindow::loadRoutingPads(), more accurate TrackMarkers insertions for fixed terminals. * New: In Katana::RoutingEvent::Key::Compare::operator(), segments with fixed axis are processed prior to any others. * New: In Katana::RoutingEventLoop, store segment pointers instead of ids to generate more accurate error messages. * Change: In Katana::RoutingPlane::create(), perform local track assignment only for HV gauges. * Change: In Katana::SegmentFsm::_slackenLocal(), add a "dragMinimize" step in the automaton. Mutliple states transitions can occurs in a row if an action fails. * New: In Katana::Session::_toIntervalAxis(), normalize interval bounds so they are on track positions (by shrinking the interval). * Bug: In Katana::TrackMarker CTOR, the weigh computation was wrong.
2018-02-17 13:27:38 -06:00
void AutoSegment::getEndAxes ( DbU::Unit& sourceAxis, DbU::Unit& targetAxis ) const
{
cdebug_log(145,0) << "AutoSegment::getEndAxes() - " << this << endl;
sourceAxis = getSourceU();
targetAxis = getTargetU();
if (not isNotAligned()) {
for( AutoSegment* aligned : const_cast<AutoSegment*>(this)->getAligneds() ) {
sourceAxis = std::min( sourceAxis, aligned->getSourceU() );
targetAxis = std::min( targetAxis, aligned->getTargetU() );
}
}
}
Replace "unsigned int" by "Flags" in all AutoSegments collections. * Change: In Anabatic::AutoSegments collections, change the type of all the flags that where in "unsigned int" (32 bits) to Flags (uint64_t) as there is now more than 32 flags for functions. * New: In Ababatic::Constants, added new flag Flags::WithPerpands, which makes the number of flags tip over 32 bits, thus making mandatory to uses Flags and not unsigned int. * New: In Anabatic::AutoSegments_Perpandiculars, manage a new flag Flags::WithDoglegs to allow to propagate through global segments that are connecteds via doglegs on local segments. Meaning that there is a good chance that they could be aligned. Slighly change the way we propagate on aligned segments: no longer check for VTee or HTee, but only for same direction and layer as master. * New: In Anabatic & Katana, replace all the "int", "long" and their variants by the less implementation ambiguous "int32_t", "int64_t" (and variant). This should help to better detect bit trucation in flags. Use the type to give a hint about the flags kind: - Type "Flags", for flags shared among Anabatic & Katana functions/methods (may also appear in some objects states). - Type "uint32_t" for flags belonging to an object internal state of from Hurricane functions flags (those should be grouped in a Flag subclass in a perfect world).
2017-05-16 07:53:33 -05:00
AutoSegments AutoSegment::getOnSourceContact ( Flags direction )
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
{
return AutoSegments_OnContact
( this, getSource() ).getSubSet( AutoSegments_InDirection(direction) );
}
Replace "unsigned int" by "Flags" in all AutoSegments collections. * Change: In Anabatic::AutoSegments collections, change the type of all the flags that where in "unsigned int" (32 bits) to Flags (uint64_t) as there is now more than 32 flags for functions. * New: In Ababatic::Constants, added new flag Flags::WithPerpands, which makes the number of flags tip over 32 bits, thus making mandatory to uses Flags and not unsigned int. * New: In Anabatic::AutoSegments_Perpandiculars, manage a new flag Flags::WithDoglegs to allow to propagate through global segments that are connecteds via doglegs on local segments. Meaning that there is a good chance that they could be aligned. Slighly change the way we propagate on aligned segments: no longer check for VTee or HTee, but only for same direction and layer as master. * New: In Anabatic & Katana, replace all the "int", "long" and their variants by the less implementation ambiguous "int32_t", "int64_t" (and variant). This should help to better detect bit trucation in flags. Use the type to give a hint about the flags kind: - Type "Flags", for flags shared among Anabatic & Katana functions/methods (may also appear in some objects states). - Type "uint32_t" for flags belonging to an object internal state of from Hurricane functions flags (those should be grouped in a Flag subclass in a perfect world).
2017-05-16 07:53:33 -05:00
AutoSegments AutoSegment::getOnTargetContact ( Flags direction )
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
{
return AutoSegments_OnContact
( this, getTarget() ).getSubSet( AutoSegments_InDirection(direction) );
}
Replace "unsigned int" by "Flags" in all AutoSegments collections. * Change: In Anabatic::AutoSegments collections, change the type of all the flags that where in "unsigned int" (32 bits) to Flags (uint64_t) as there is now more than 32 flags for functions. * New: In Ababatic::Constants, added new flag Flags::WithPerpands, which makes the number of flags tip over 32 bits, thus making mandatory to uses Flags and not unsigned int. * New: In Anabatic::AutoSegments_Perpandiculars, manage a new flag Flags::WithDoglegs to allow to propagate through global segments that are connecteds via doglegs on local segments. Meaning that there is a good chance that they could be aligned. Slighly change the way we propagate on aligned segments: no longer check for VTee or HTee, but only for same direction and layer as master. * New: In Anabatic & Katana, replace all the "int", "long" and their variants by the less implementation ambiguous "int32_t", "int64_t" (and variant). This should help to better detect bit trucation in flags. Use the type to give a hint about the flags kind: - Type "Flags", for flags shared among Anabatic & Katana functions/methods (may also appear in some objects states). - Type "uint32_t" for flags belonging to an object internal state of from Hurricane functions flags (those should be grouped in a Flag subclass in a perfect world).
2017-05-16 07:53:33 -05:00
AutoSegments AutoSegment::getCachedOnSourceContact ( Flags direction )
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
{ return AutoSegments_CachedOnContact( getAutoSource(), direction ); }
Replace "unsigned int" by "Flags" in all AutoSegments collections. * Change: In Anabatic::AutoSegments collections, change the type of all the flags that where in "unsigned int" (32 bits) to Flags (uint64_t) as there is now more than 32 flags for functions. * New: In Ababatic::Constants, added new flag Flags::WithPerpands, which makes the number of flags tip over 32 bits, thus making mandatory to uses Flags and not unsigned int. * New: In Anabatic::AutoSegments_Perpandiculars, manage a new flag Flags::WithDoglegs to allow to propagate through global segments that are connecteds via doglegs on local segments. Meaning that there is a good chance that they could be aligned. Slighly change the way we propagate on aligned segments: no longer check for VTee or HTee, but only for same direction and layer as master. * New: In Anabatic & Katana, replace all the "int", "long" and their variants by the less implementation ambiguous "int32_t", "int64_t" (and variant). This should help to better detect bit trucation in flags. Use the type to give a hint about the flags kind: - Type "Flags", for flags shared among Anabatic & Katana functions/methods (may also appear in some objects states). - Type "uint32_t" for flags belonging to an object internal state of from Hurricane functions flags (those should be grouped in a Flag subclass in a perfect world).
2017-05-16 07:53:33 -05:00
AutoSegments AutoSegment::getCachedOnTargetContact ( Flags direction )
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
{ return AutoSegments_CachedOnContact( getAutoTarget(), direction ); }
Replace "unsigned int" by "Flags" in all AutoSegments collections. * Change: In Anabatic::AutoSegments collections, change the type of all the flags that where in "unsigned int" (32 bits) to Flags (uint64_t) as there is now more than 32 flags for functions. * New: In Ababatic::Constants, added new flag Flags::WithPerpands, which makes the number of flags tip over 32 bits, thus making mandatory to uses Flags and not unsigned int. * New: In Anabatic::AutoSegments_Perpandiculars, manage a new flag Flags::WithDoglegs to allow to propagate through global segments that are connecteds via doglegs on local segments. Meaning that there is a good chance that they could be aligned. Slighly change the way we propagate on aligned segments: no longer check for VTee or HTee, but only for same direction and layer as master. * New: In Anabatic & Katana, replace all the "int", "long" and their variants by the less implementation ambiguous "int32_t", "int64_t" (and variant). This should help to better detect bit trucation in flags. Use the type to give a hint about the flags kind: - Type "Flags", for flags shared among Anabatic & Katana functions/methods (may also appear in some objects states). - Type "uint32_t" for flags belonging to an object internal state of from Hurricane functions flags (those should be grouped in a Flag subclass in a perfect world).
2017-05-16 07:53:33 -05:00
AutoSegments AutoSegment::getAligneds ( Flags flags )
Katana routing algorithm adjustements for analog routing. * Bug: In Hurricane::Interval, the ::getSize() method was returning a negative length when the Interval was empty. Now return zero. This was causing slight cost functions side effect when computing the overlap of a segment with another one belonging to the same net (i.e. shareDelta) * Bug: In Katana::AutoSegment::computeOptimal(), checks that the optimal position for the axis is indeed inside the GCell interval. * Bug: In Katana::DataSymmetric::getSymmetrical(Interval&), reverse the shrink applied to GCells sides. Interval are of the type [min:max[ so after symmetric transformation they become ]smax:smin]. Apply a correction so that they are [smax:smin[ and could be merged with the paired interval (the one without symmetry applied). * Bug: In katana::TrackCost::consolidate(), the overlap (now) always positive, and must be *substracted* to the "delta" (not *added*). * Bug: In Katana::_computeNetOptimals(), now skip the already processeds AutoSegments instead of re-processing them. * New: In Katana::AutoSegment::computeOptimal(), improved computation of the axis hint: - For analog net, propagate through the simple doglegs to find attractors. - Restrict the allowed interval to the union of GCell sides and segments constraints that are terminals. - Consider a local as "long" if it's length exceed 20 the P-pitch. (maybe make that a tool parameter).
2017-05-22 17:20:31 -05:00
{ return AutoSegments_Aligneds( this, flags ); }
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
Replace "unsigned int" by "Flags" in all AutoSegments collections. * Change: In Anabatic::AutoSegments collections, change the type of all the flags that where in "unsigned int" (32 bits) to Flags (uint64_t) as there is now more than 32 flags for functions. * New: In Ababatic::Constants, added new flag Flags::WithPerpands, which makes the number of flags tip over 32 bits, thus making mandatory to uses Flags and not unsigned int. * New: In Anabatic::AutoSegments_Perpandiculars, manage a new flag Flags::WithDoglegs to allow to propagate through global segments that are connecteds via doglegs on local segments. Meaning that there is a good chance that they could be aligned. Slighly change the way we propagate on aligned segments: no longer check for VTee or HTee, but only for same direction and layer as master. * New: In Anabatic & Katana, replace all the "int", "long" and their variants by the less implementation ambiguous "int32_t", "int64_t" (and variant). This should help to better detect bit trucation in flags. Use the type to give a hint about the flags kind: - Type "Flags", for flags shared among Anabatic & Katana functions/methods (may also appear in some objects states). - Type "uint32_t" for flags belonging to an object internal state of from Hurricane functions flags (those should be grouped in a Flag subclass in a perfect world).
2017-05-16 07:53:33 -05:00
AutoSegments AutoSegment::getConnecteds ( Flags flags )
Katana routing algorithm adjustements for analog routing. * Bug: In Hurricane::Interval, the ::getSize() method was returning a negative length when the Interval was empty. Now return zero. This was causing slight cost functions side effect when computing the overlap of a segment with another one belonging to the same net (i.e. shareDelta) * Bug: In Katana::AutoSegment::computeOptimal(), checks that the optimal position for the axis is indeed inside the GCell interval. * Bug: In Katana::DataSymmetric::getSymmetrical(Interval&), reverse the shrink applied to GCells sides. Interval are of the type [min:max[ so after symmetric transformation they become ]smax:smin]. Apply a correction so that they are [smax:smin[ and could be merged with the paired interval (the one without symmetry applied). * Bug: In katana::TrackCost::consolidate(), the overlap (now) always positive, and must be *substracted* to the "delta" (not *added*). * Bug: In Katana::_computeNetOptimals(), now skip the already processeds AutoSegments instead of re-processing them. * New: In Katana::AutoSegment::computeOptimal(), improved computation of the axis hint: - For analog net, propagate through the simple doglegs to find attractors. - Restrict the allowed interval to the union of GCell sides and segments constraints that are terminals. - Consider a local as "long" if it's length exceed 20 the P-pitch. (maybe make that a tool parameter).
2017-05-22 17:20:31 -05:00
{ return AutoSegments_Connecteds( this, flags ); }
Apaired segments building for symmetric routing (step 1). * Change: In Hurricane::BaseFlags, store flags in uint64_t instead of unsigned int because we start to need more than 32 different flags in some tools. * New: In ::getString() & ::getRecord() templates, add support for std::array<>. * Change: In CRL::ToolEngine, add support for timer (time & memory measurements) displaced from Katabatic. This way all ToolEngine can use this feature. The _postCreate() method display the memory just after ToolEngine allocation. * Change: In Etesian::EtesianEngine, make use of the ToolEngine builtin timer (remove the local one). Forgot to call the base class _postCreate() and _preDestroy(). * Change: In Anabatic::AnabaticEngine, make use of the ToolEngine builtin timer (remove the local one). * New: In Anabatic, new AutoSegments_Connecteds() collection. This Collection allows a deterministic walkthough *all* the AutoSegments connected either to source or target of one AutoSegment. * New: In Anabatic::AutoContactTerminal::isEndPoint() to check if an AutoContactTerminal is the *only one* anchored on a RoutingPad, thus being a true "end point" and not a kind of feed-through. * New: In Katana::KatanaEngine, added support for symmetric nets. Created new class DataSymmetric to store symmetric information of a net (mainly the paired AutoSegments). Added KatanaEngine::runSymmetricRouter(), for now only build the DataSymmetric informations. More to come... * Change: In Katana::GraphicKatanaEngine::_runTest(), now perform symmetric information building the non-symmetric routing.
2017-03-12 13:34:12 -05:00
Added analog type on segment NetRoutingProperty. * New: In Anabatic_AutoSegments collection, added a Flag to the constructors to allow different behavior between digital and analog modes. For "Aligneds" and "Perpandiculars" collections, now manage a new Flag WithDoglegs to follow aligned globals through local doglegs (for analog nets). Adjust the log level of collections to 144 (formerly 145). * New: In Anabatic::AutoSegment, new flag SegAnalog for segments that are part of an analog net. Note that with this flag, we reach the 32 bits limit... * Change: In Anabatic::Constants, Flags are now declared as BaseFlags objects and *not* uint64_t. This avoids overload resolution problems with arithmetical overload of the operators. The BaseFlags/Flags types are now completly "isolated" from the uint64_t, it has the advantage of showing where unwanted previous implicit conversions where occuring. * Change: In Katana::Constants, Flags values are now of BaseFlags type instead of uint64_t. * Change: In Anabatic::Dijkstra, lots of log cleanup. * Change: In Anabatic::GCell::getSide(), make the "shrink" parameter visible to allow to substract the topmost and rightmost track for axis span computation in AutoSegment::computeOptimal(). Used for analog mode. * Change: In NetRoutingState, added a flag for analog mode. Use uint32_t for the flags type. * New: In Isobar, export the NetRoutingState and NetRoutingExtension objects.
2017-05-20 05:33:12 -05:00
AutoSegments AutoSegment::getPerpandiculars ( Flags flags )
{ return AutoSegments_Perpandiculars( this, flags ); }
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
bool AutoSegment::checkDepthSpin () const
{
bool valid = true;
const Layer* sourceLayer = getAutoSource()->getLayer();
const Layer* targetLayer = getAutoTarget()->getLayer();
In Anabatic/Katana, add support for VH gauges (real technos). * Change: In Anabatic::AutoContactTerminal::getNativeConstraintBox(), when the anchor is a RoutingPad (which must be always the case), perform the true computation of it's position based on the segment occurrence. It is a important change, previously the area was in fact the "center line" of the connector while now it is really an area (mandatory for "half-offgrid" terminals of real technologies). The change is not complete yet, the area should be shrinked by the half size of a VIA, because the area applies to the center coordinate of the VIA (to be done quickly). * Bug: In Anabatic::AutoContactTurn::updateTopology(), when a dogleg is created (restore connexity after a layer change) the layer of the VIA, based on the segments it connects to must be re-computed *after* the dogleg has been made. * Change: In all files of Anabatic, when comparing two layers, no longer use the Layer pointer itself, but the layer mask. This allow a transparent management of both real and symbolic layers (which do share the same mask). Real metal layers (not VIAs) will be BasicLayer and symbolic metal layers will be RegularLayer. * New: Anabatic::Configuration::selectRpComponent(), select the best RoutingPad component for metal1 terminals. Look for the metal1 component with the biggest accessibility on-grid. RoutingPad using other metals are left untoucheds. * New: New function Anabatic::Vertex::getNeighbor(Edge*) to get the neighbor Vertex through an Edge*. This method allows to write clearer code as we no longer need to access the neighbor through the underlying GCell. Also add proxies for GCell methods in Vertex. * Bug: In Anabatic::Dijkstra::_toSources(), in the ripup stage, when a component with multiples vertexes is reached *and* two of it's vertexes are reached *at the same time* (one from which we backtrack and one still in the queue) extraneous edges may be created by _materialize(). Case occurs on snx/c35b4, "abc_5360_n903_1". To solve this, Dijkstra::_toSource() is modificated, the "from" edges of the newly reacheds vertexes are reset to NULL, *except* for the one we will be backtracking from. That is, the one given in the source argument. * Change: In Anabatic::NetBuilder class, put the various Hooks and RoutingPad sorting functions as class ones. * Bug: In AutoSegment::setLayer(), raise the SegInvalidatedFayer flag. This unset flag was causing AutoContactTurn::updateTopology() to not work as expected and making gaps, this was the cause of the last remaining warnings about layer connexity.
2018-01-06 09:55:53 -06:00
if ( (_flags & SegSourceTop)
and (sourceLayer->getBottom()->getMask() != getLayer()->getMask()) ) {
cerr << Error( "%s\n"
" Source is not going above, connected to *top* of %s.\n"
" bottom:%s mask:%s\n"
" layer:%s mask:%s\n"
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
, getString(this).c_str()
, getString(getAutoSource()).c_str()
In Anabatic/Katana, add support for VH gauges (real technos). * Change: In Anabatic::AutoContactTerminal::getNativeConstraintBox(), when the anchor is a RoutingPad (which must be always the case), perform the true computation of it's position based on the segment occurrence. It is a important change, previously the area was in fact the "center line" of the connector while now it is really an area (mandatory for "half-offgrid" terminals of real technologies). The change is not complete yet, the area should be shrinked by the half size of a VIA, because the area applies to the center coordinate of the VIA (to be done quickly). * Bug: In Anabatic::AutoContactTurn::updateTopology(), when a dogleg is created (restore connexity after a layer change) the layer of the VIA, based on the segments it connects to must be re-computed *after* the dogleg has been made. * Change: In all files of Anabatic, when comparing two layers, no longer use the Layer pointer itself, but the layer mask. This allow a transparent management of both real and symbolic layers (which do share the same mask). Real metal layers (not VIAs) will be BasicLayer and symbolic metal layers will be RegularLayer. * New: Anabatic::Configuration::selectRpComponent(), select the best RoutingPad component for metal1 terminals. Look for the metal1 component with the biggest accessibility on-grid. RoutingPad using other metals are left untoucheds. * New: New function Anabatic::Vertex::getNeighbor(Edge*) to get the neighbor Vertex through an Edge*. This method allows to write clearer code as we no longer need to access the neighbor through the underlying GCell. Also add proxies for GCell methods in Vertex. * Bug: In Anabatic::Dijkstra::_toSources(), in the ripup stage, when a component with multiples vertexes is reached *and* two of it's vertexes are reached *at the same time* (one from which we backtrack and one still in the queue) extraneous edges may be created by _materialize(). Case occurs on snx/c35b4, "abc_5360_n903_1". To solve this, Dijkstra::_toSource() is modificated, the "from" edges of the newly reacheds vertexes are reset to NULL, *except* for the one we will be backtracking from. That is, the one given in the source argument. * Change: In Anabatic::NetBuilder class, put the various Hooks and RoutingPad sorting functions as class ones. * Bug: In AutoSegment::setLayer(), raise the SegInvalidatedFayer flag. This unset flag was causing AutoContactTurn::updateTopology() to not work as expected and making gaps, this was the cause of the last remaining warnings about layer connexity.
2018-01-06 09:55:53 -06:00
, getString(sourceLayer->getBottom()).c_str()
, getString(sourceLayer->getBottom()->getMask()).c_str()
, getString(getLayer()).c_str()
, getString(getLayer()->getMask()).c_str()
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
) << endl;
valid = false;
}
In Anabatic/Katana, add support for VH gauges (real technos). * Change: In Anabatic::AutoContactTerminal::getNativeConstraintBox(), when the anchor is a RoutingPad (which must be always the case), perform the true computation of it's position based on the segment occurrence. It is a important change, previously the area was in fact the "center line" of the connector while now it is really an area (mandatory for "half-offgrid" terminals of real technologies). The change is not complete yet, the area should be shrinked by the half size of a VIA, because the area applies to the center coordinate of the VIA (to be done quickly). * Bug: In Anabatic::AutoContactTurn::updateTopology(), when a dogleg is created (restore connexity after a layer change) the layer of the VIA, based on the segments it connects to must be re-computed *after* the dogleg has been made. * Change: In all files of Anabatic, when comparing two layers, no longer use the Layer pointer itself, but the layer mask. This allow a transparent management of both real and symbolic layers (which do share the same mask). Real metal layers (not VIAs) will be BasicLayer and symbolic metal layers will be RegularLayer. * New: Anabatic::Configuration::selectRpComponent(), select the best RoutingPad component for metal1 terminals. Look for the metal1 component with the biggest accessibility on-grid. RoutingPad using other metals are left untoucheds. * New: New function Anabatic::Vertex::getNeighbor(Edge*) to get the neighbor Vertex through an Edge*. This method allows to write clearer code as we no longer need to access the neighbor through the underlying GCell. Also add proxies for GCell methods in Vertex. * Bug: In Anabatic::Dijkstra::_toSources(), in the ripup stage, when a component with multiples vertexes is reached *and* two of it's vertexes are reached *at the same time* (one from which we backtrack and one still in the queue) extraneous edges may be created by _materialize(). Case occurs on snx/c35b4, "abc_5360_n903_1". To solve this, Dijkstra::_toSource() is modificated, the "from" edges of the newly reacheds vertexes are reset to NULL, *except* for the one we will be backtracking from. That is, the one given in the source argument. * Change: In Anabatic::NetBuilder class, put the various Hooks and RoutingPad sorting functions as class ones. * Bug: In AutoSegment::setLayer(), raise the SegInvalidatedFayer flag. This unset flag was causing AutoContactTurn::updateTopology() to not work as expected and making gaps, this was the cause of the last remaining warnings about layer connexity.
2018-01-06 09:55:53 -06:00
if ( (_flags & SegSourceBottom)
and (sourceLayer->getTop()->getMask() != getLayer()->getMask()) ) {
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
cerr << Error("%s\n"
" Source is not going below, connected to *bottom* of %s."
, getString(this).c_str()
, getString(getAutoSource()).c_str()
) << endl;
valid = false;
}
In Anabatic/Katana, add support for VH gauges (real technos). * Change: In Anabatic::AutoContactTerminal::getNativeConstraintBox(), when the anchor is a RoutingPad (which must be always the case), perform the true computation of it's position based on the segment occurrence. It is a important change, previously the area was in fact the "center line" of the connector while now it is really an area (mandatory for "half-offgrid" terminals of real technologies). The change is not complete yet, the area should be shrinked by the half size of a VIA, because the area applies to the center coordinate of the VIA (to be done quickly). * Bug: In Anabatic::AutoContactTurn::updateTopology(), when a dogleg is created (restore connexity after a layer change) the layer of the VIA, based on the segments it connects to must be re-computed *after* the dogleg has been made. * Change: In all files of Anabatic, when comparing two layers, no longer use the Layer pointer itself, but the layer mask. This allow a transparent management of both real and symbolic layers (which do share the same mask). Real metal layers (not VIAs) will be BasicLayer and symbolic metal layers will be RegularLayer. * New: Anabatic::Configuration::selectRpComponent(), select the best RoutingPad component for metal1 terminals. Look for the metal1 component with the biggest accessibility on-grid. RoutingPad using other metals are left untoucheds. * New: New function Anabatic::Vertex::getNeighbor(Edge*) to get the neighbor Vertex through an Edge*. This method allows to write clearer code as we no longer need to access the neighbor through the underlying GCell. Also add proxies for GCell methods in Vertex. * Bug: In Anabatic::Dijkstra::_toSources(), in the ripup stage, when a component with multiples vertexes is reached *and* two of it's vertexes are reached *at the same time* (one from which we backtrack and one still in the queue) extraneous edges may be created by _materialize(). Case occurs on snx/c35b4, "abc_5360_n903_1". To solve this, Dijkstra::_toSource() is modificated, the "from" edges of the newly reacheds vertexes are reset to NULL, *except* for the one we will be backtracking from. That is, the one given in the source argument. * Change: In Anabatic::NetBuilder class, put the various Hooks and RoutingPad sorting functions as class ones. * Bug: In AutoSegment::setLayer(), raise the SegInvalidatedFayer flag. This unset flag was causing AutoContactTurn::updateTopology() to not work as expected and making gaps, this was the cause of the last remaining warnings about layer connexity.
2018-01-06 09:55:53 -06:00
if ( (_flags & SegTargetTop)
and (targetLayer->getBottom()->getMask() != getLayer()->getMask()) ) {
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
cerr << Error("%s\n"
" Target is not going above connected to *top* of %s."
, getString(this).c_str()
, getString(getAutoTarget()).c_str()
) << endl;
valid = false;
}
In Anabatic/Katana, add support for VH gauges (real technos). * Change: In Anabatic::AutoContactTerminal::getNativeConstraintBox(), when the anchor is a RoutingPad (which must be always the case), perform the true computation of it's position based on the segment occurrence. It is a important change, previously the area was in fact the "center line" of the connector while now it is really an area (mandatory for "half-offgrid" terminals of real technologies). The change is not complete yet, the area should be shrinked by the half size of a VIA, because the area applies to the center coordinate of the VIA (to be done quickly). * Bug: In Anabatic::AutoContactTurn::updateTopology(), when a dogleg is created (restore connexity after a layer change) the layer of the VIA, based on the segments it connects to must be re-computed *after* the dogleg has been made. * Change: In all files of Anabatic, when comparing two layers, no longer use the Layer pointer itself, but the layer mask. This allow a transparent management of both real and symbolic layers (which do share the same mask). Real metal layers (not VIAs) will be BasicLayer and symbolic metal layers will be RegularLayer. * New: Anabatic::Configuration::selectRpComponent(), select the best RoutingPad component for metal1 terminals. Look for the metal1 component with the biggest accessibility on-grid. RoutingPad using other metals are left untoucheds. * New: New function Anabatic::Vertex::getNeighbor(Edge*) to get the neighbor Vertex through an Edge*. This method allows to write clearer code as we no longer need to access the neighbor through the underlying GCell. Also add proxies for GCell methods in Vertex. * Bug: In Anabatic::Dijkstra::_toSources(), in the ripup stage, when a component with multiples vertexes is reached *and* two of it's vertexes are reached *at the same time* (one from which we backtrack and one still in the queue) extraneous edges may be created by _materialize(). Case occurs on snx/c35b4, "abc_5360_n903_1". To solve this, Dijkstra::_toSource() is modificated, the "from" edges of the newly reacheds vertexes are reset to NULL, *except* for the one we will be backtracking from. That is, the one given in the source argument. * Change: In Anabatic::NetBuilder class, put the various Hooks and RoutingPad sorting functions as class ones. * Bug: In AutoSegment::setLayer(), raise the SegInvalidatedFayer flag. This unset flag was causing AutoContactTurn::updateTopology() to not work as expected and making gaps, this was the cause of the last remaining warnings about layer connexity.
2018-01-06 09:55:53 -06:00
if ( (_flags & SegTargetBottom)
and (targetLayer->getTop()->getMask() != getLayer()->getMask()) ) {
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
cerr << Error("%s\n"
" Target is not going below, connected to *bottom* of %s."
, getString(this).c_str()
, getString(getAutoTarget()).c_str()
) << endl;
valid = false;
}
return valid;
}
Katana manage wide wires, and they can also be symmetric. * New: In Anabatic::AutoContact and the derived classes, manages wide wires. The contact self dimension itself according to the segments it is connected to. Special case for the AutoContactTerminal which also read the size of the component it is anchored upon. New refresh method "updateSize()" and flag CntInvalidatedWidth. to compute the size. In AutoContactTerminal, compute the constraint box according to the width of the segment. * New: In Anabatic::AutoSegment, flags are now implemented as "static const" attributes of the class. The flags are stored into a uint64_t as they are more than 32. Added new flag "SegWide" and associated predicates. * Change: In GCellTopology::_doHChannel() and GCellTopology::_doVChannel(), uses the simpler overload of AutoSegment::create() in order to detect the wire width automatically. * New: In Katana::Manipulator, split insertToTrack() and forceToTrack() into a one-track method and a segment level method that iterate over the track span of the segment. * New: In Katana::SegmentFsm, for each cost in the table, now allow access to a specific track. So the base functions have now two parameters: "icost" and "itrack" (has a cost can have multiple tracks in the case of wide segments). * Change: In Katana::TrackElement, remove the index of the element inside it's track, as for a wide segment it will not be meaningful for the non-base track. This means that we have to use the Track::find() method each time instead. Remove the wide flag, as it is a duplicate of the one in AutoSegment. Added a getTrackCount() method to tell the number of track the segment is inserted into. Needed in the Track destroy step to delete a segment only when the last track that refers it is destroyed. Added getSymmetricAxis() to correct the computation of the symmetric base track in case of wide segment as the base track is not centered but the the leftmost one. * Change: In Track::insert() insert wide segments in their whole track span. * Change: In TrackCost, create an array of costs according to the segment track span. * Change: In TrackSegment::create(), now activate the factory and create wide segments. * Bug: In Katana::AutoSegments_Perpandicular, correct the debug indentation problem (ever shifting to the right).
2017-07-28 08:30:22 -05:00
void AutoSegment::setFlagsOnAligneds ( uint64_t flags )
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
{
setFlags( flags );
if (not isNotAligned()) {
Anabatic transient commit 18. Port of Kite (Katana), Yeah, Baby! Yeah! * Bug: In Hurricane, in StaticObservable::getObserver(), if the slot pointer is NULL, do not try to access the owner. Returns NULL, so the caller can be aware of the situation... * Change: In Hurricane, in BreakpointWidget & ExceptionWidget some cosmetic changes (fonts and window sizes). * Bug: In Anabatic, In AutoHorizontal::getConstraints(), take into account the constraints from the source AutoContact, as it holds the constraints transmitted by the RoutingPads and sets up by propageConstraintsFromRp(). It is likely to be a bug affecting the original Katabatic as well. * Change: In Anabatic, in RawGCellsUnder(), check that the segment is not completly oustside the cell abutment box and truncate the coordinates to the part that is inside. Use the "shrink" if we reach the east/north border. * Change: In Anabatic, in Configuration, no more decorator because we will use a true derived relationship. Katana *derives* from *Anabatic* and do not *decorate* it, so the Configuration can do the same. It also implies that we directly create a Katana engine, not an Anabatic one. * Change: In Anabatic, in Session, do not allow the opening of the Session in a standalone fashion (with a static method). Instead it must be opened using the relevant method of the Anabatic/Katana engine. This ensure we are opening the right Session type. * Change: In Anabatic, in AutoSegment_Aligneds() collection the seed segment is not part of the collection by default, but will be included if the Flags::WithSelf is set. * Change: In Configuration, all the flags value are now defined in two steps. Declared in the header and initialized in the module. This is to prevent the fact that on some cases, in relation with the Python "extern C" part modules, we need a true allocated variable. It was causing weird linking problems. A side effect is that they can no longer be used as entry is switches, have to replace them by if/else. * New: In Anabatic, new GCell::getNeighborAt() utility function. * Bug: In Anabatic, in GCell::doGrid(), tag all the GCells of the grid with the grid type... Back annote all the edges capacity (north & east) with the reserved local capacity. * New: Complete portage of Kite over Anabatic. The new engine is christened "Katana" for Kite-Analogic. When it's capabilities and performances will be on a part with Kite, it is to completly replace it (and take back the "Kite" name). Preliminary tests seems to show that, contrary to intuition (because built on a more complex/slower grid), it is even slightly faster than Kite 8-).
2016-08-15 09:30:13 -05:00
for( AutoSegment* segment : getAligneds() )
segment->setFlags( flags );
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
}
}
void AutoSegment::sourceDetach ()
{
AutoContact* source = getAutoSource();
if (source) {
if (source->isTurn()) {
AutoSegment* perpandicular = source->getPerpandicular(this);
if (perpandicular and perpandicular->isReduced())
decReduceds();
}
base()->getSourceHook()->detach();
source->cacheDetach( this );
unsetFlags( SegNotSourceAligned );
setFlags( SegInvalidatedSource );
}
}
void AutoSegment::targetDetach ()
{
AutoContact* target = getAutoTarget();
if (target) {
if (target->isTurn()) {
AutoSegment* perpandicular = target->getPerpandicular(this);
if (perpandicular and perpandicular->isReduced())
decReduceds();
}
base()->getTargetHook()->detach();
target->cacheDetach( this );
unsetFlags( SegNotTargetAligned );
setFlags( SegInvalidatedTarget );
}
}
void AutoSegment::sourceAttach ( AutoContact* source )
{
if (source) {
if (not base()->getSourceHook()->isAttached())
base()->getSourceHook()->attach( source->base()->getBodyHook() );
source->cacheAttach( this );
// if (source->isHTee() and isHorizontal()) return;
// else if (source->isVTee() and isVertical ()) return;
// setFlags( SegNotSourceAligned );
}
}
void AutoSegment::targetAttach ( AutoContact* target )
{
if (target) {
if (not base()->getTargetHook()->isAttached())
base()->getTargetHook()->attach( target->base()->getBodyHook() );
target->cacheAttach( this );
// if (target->isHTee() and isHorizontal()) return;
// else if (target->isVTee() and isVertical ()) return;
// setFlags( SegNotTargetAligned );
}
}
void AutoSegment::mergeUserConstraints ( const Interval& constraints )
{
Improved handling of short nets (fully included in one GCell). The short net mode degrade the routing in some cases. This will be fixed in a next batch of commits. * New: In Hurricane::NetRoutingProperty, added "ShortNet" flag for Nets that are completly inside *one* GCell. * Bug: In CRL::BlifParser::Model::staticInit(), when looking for the output of zero and one cell, also skip the blockage net (as well as automatic and supplies). * New: In Anabatic::AutoSegment, added "ShortNet" flag to know if the segment is part of a short net (fully included in *one* GCell). Also add accessor/mutators for the _analogMode flag (was it ever used before?). * New: In Anabatic::NetBuilder::singleGCell(), if a RoutingPad is vertically small, add a vertical segment to give it some slack. * New: In Anabatic::Dijkstra::_materialize(), detect "short net" as they have only one GCell in their source list... * Bug: In AnabaticEngine::_loadGrbyNet(), reset the AutoSegment "short net" and "analog mode" creation flags between two different nets. * New: In Katana::Configuration, added dedicated ripup for short net segmnts. * New: In Katana: partially implemented support for "short dogleg", that is dogleg that are always kept in same metal because they connect neighboring perpandicular tracks. Not finished neither activated yet. * New: In Katana::TreckElement and derived, export the the *short net* support from AutoSegment. * Bug: In Katana::RoutingEvent::_processRepair(), when a segment is successfully inserted, re-process any perpandicular that is in repair state, as it may have a new chance to be placed. * New: In Katana::SegmentFsm::slackenTopology(), always reject short nets. * Bug: In Katana::Track::check(), correctly handle wide segments instead of issuing false check messages.
2018-07-16 04:16:51 -05:00
DebugSession::open( getNet(), 149, 160 );
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
cdebug_log(149,0) << "mergeUserConstraints() " << this << endl;
cdebug_log(149,0) << "| " << constraints << " merged with " << _userConstraints << endl;
_userConstraints.intersection(constraints);
Improved handling of short nets (fully included in one GCell). The short net mode degrade the routing in some cases. This will be fixed in a next batch of commits. * New: In Hurricane::NetRoutingProperty, added "ShortNet" flag for Nets that are completly inside *one* GCell. * Bug: In CRL::BlifParser::Model::staticInit(), when looking for the output of zero and one cell, also skip the blockage net (as well as automatic and supplies). * New: In Anabatic::AutoSegment, added "ShortNet" flag to know if the segment is part of a short net (fully included in *one* GCell). Also add accessor/mutators for the _analogMode flag (was it ever used before?). * New: In Anabatic::NetBuilder::singleGCell(), if a RoutingPad is vertically small, add a vertical segment to give it some slack. * New: In Anabatic::Dijkstra::_materialize(), detect "short net" as they have only one GCell in their source list... * Bug: In AnabaticEngine::_loadGrbyNet(), reset the AutoSegment "short net" and "analog mode" creation flags between two different nets. * New: In Katana::Configuration, added dedicated ripup for short net segmnts. * New: In Katana: partially implemented support for "short dogleg", that is dogleg that are always kept in same metal because they connect neighboring perpandicular tracks. Not finished neither activated yet. * New: In Katana::TreckElement and derived, export the the *short net* support from AutoSegment. * Bug: In Katana::RoutingEvent::_processRepair(), when a segment is successfully inserted, re-process any perpandicular that is in repair state, as it may have a new chance to be placed. * New: In Katana::SegmentFsm::slackenTopology(), always reject short nets. * Bug: In Katana::Track::check(), correctly handle wide segments instead of issuing false check messages.
2018-07-16 04:16:51 -05:00
DebugSession::close();
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
}
Replace "unsigned int" by "Flags" in all AutoSegments collections. * Change: In Anabatic::AutoSegments collections, change the type of all the flags that where in "unsigned int" (32 bits) to Flags (uint64_t) as there is now more than 32 flags for functions. * New: In Ababatic::Constants, added new flag Flags::WithPerpands, which makes the number of flags tip over 32 bits, thus making mandatory to uses Flags and not unsigned int. * New: In Anabatic::AutoSegments_Perpandiculars, manage a new flag Flags::WithDoglegs to allow to propagate through global segments that are connecteds via doglegs on local segments. Meaning that there is a good chance that they could be aligned. Slighly change the way we propagate on aligned segments: no longer check for VTee or HTee, but only for same direction and layer as master. * New: In Anabatic & Katana, replace all the "int", "long" and their variants by the less implementation ambiguous "int32_t", "int64_t" (and variant). This should help to better detect bit trucation in flags. Use the type to give a hint about the flags kind: - Type "Flags", for flags shared among Anabatic & Katana functions/methods (may also appear in some objects states). - Type "uint32_t" for flags belonging to an object internal state of from Hurricane functions flags (those should be grouped in a Flag subclass in a perfect world).
2017-05-16 07:53:33 -05:00
bool AutoSegment::toConstraintAxis ( Flags flags )
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
{
cdebug_log(149,1) << "toConstraintAxis() " << this << endl;
if (not isCanonical()) { cdebug_tabw(149,-1); return false; }
DbU::Unit constraintMin;
DbU::Unit constraintMax;
getConstraints( constraintMin, constraintMax );
// Empty constraint interval: ignore.
if (constraintMin > constraintMax) { cdebug_tabw(149,-1); return false; }
if (isDogleg()) {
DbU::Unit halfSideLength = getAutoSource()->getGCell()->getSide
( isHorizontal() ? Flags::Vertical : Flags::Horizontal ).getHalfSize();
constraintMin -= halfSideLength;
constraintMax += halfSideLength;
}
if (getAxis() < constraintMin) {
setAxis( constraintMin, flags );
cdebug_tabw(149,-1);
return true;
}
if (getAxis() > constraintMax) {
setAxis( constraintMax, flags );
cdebug_tabw(149,-1);
return true;
}
cdebug_tabw(149,-1);
return false;
}
Replace "unsigned int" by "Flags" in all AutoSegments collections. * Change: In Anabatic::AutoSegments collections, change the type of all the flags that where in "unsigned int" (32 bits) to Flags (uint64_t) as there is now more than 32 flags for functions. * New: In Ababatic::Constants, added new flag Flags::WithPerpands, which makes the number of flags tip over 32 bits, thus making mandatory to uses Flags and not unsigned int. * New: In Anabatic::AutoSegments_Perpandiculars, manage a new flag Flags::WithDoglegs to allow to propagate through global segments that are connecteds via doglegs on local segments. Meaning that there is a good chance that they could be aligned. Slighly change the way we propagate on aligned segments: no longer check for VTee or HTee, but only for same direction and layer as master. * New: In Anabatic & Katana, replace all the "int", "long" and their variants by the less implementation ambiguous "int32_t", "int64_t" (and variant). This should help to better detect bit trucation in flags. Use the type to give a hint about the flags kind: - Type "Flags", for flags shared among Anabatic & Katana functions/methods (may also appear in some objects states). - Type "uint32_t" for flags belonging to an object internal state of from Hurricane functions flags (those should be grouped in a Flag subclass in a perfect world).
2017-05-16 07:53:33 -05:00
bool AutoSegment::toOptimalAxis ( Flags flags )
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
{
cdebug_log(149,1) << "toOptimalAxis() " << this << endl;
if (not isCanonical()) { cdebug_tabw(149,-1); return false; }
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
if (not isUnsetAxis()) {
cdebug_tabw(149,-1);
return toConstraintAxis( flags );
}
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
DbU::Unit constraintMin;
DbU::Unit constraintMax;
getConstraints( constraintMin, constraintMax );
DbU::Unit optimalMin = max( min(getOptimalMin(),constraintMax), constraintMin );
DbU::Unit optimalMax = min( max(getOptimalMax(),constraintMin), constraintMax );
First intergration of the Analogic router parts. * New: In Anabatic::AutoSegment, introduce a the kind (associated to a flag) "LongLocal". Analog GCells can be very wide, so at least some carefuly choosen long local segments must be took into account as attractors in the computation of the optimal axis. * New: In Anabatic::AutoSegment::computeOptimal(), take LongLocal into account as attractors. * Change: In ::GCellTopology constructors compare the layers of the RoutingPads using layer masks instead of Layer pointers. Allows to find both "METALx" (symbolic) and "metalX" (real). * Change: In ::GCellTopology::_doHChannel(), _doChannel(), _doStrut() and _doDevice(), tag long locals as "LongLocal". This need to be reviewed as it as bind done a bit too quickly. * Change: In Anabatic::AutoSegment, due too a much bigger span of the analogic GCells the _optimalMin & _optimalMax bitfields must use 16 bits instead of 8 (they where overflowed). * New: In Katana, reorganisation of the initialization procedure to fit both digital and analogic cases. Create an analogInit() method. * Change: In Katana::RoutingEvent, the _tracksNb and _tracksFree bitfields where too short for the Analog GCell size, now uses 16 bits instead of 6. * Bug: In Katana::GraphicKatanEngine::drawGCell(), skip drawing of a GCell if *both* width and height are under 150 pixels. * New: In Katana::Session, add a new isOpen() method.
2016-10-04 10:12:58 -05:00
cdebug_log(149,0) << "optimal:[" << DbU::getValueString(optimalMin)
Support for symmetric routing in Anabatic/Katabatic. * New: In Katana::SegmentFsm, the object is now able to handle two events at the same time. The master and it's symmetric. When there is no symmetric, the corresponing data is just left blank. This makes a bigger object, but as there is only one when running, it is not an issue. Candidates tracks are now an vector of array<2> (pairs), the TrackCost::Compare() functor has to be wrapped through CompareCostArray. The compined TrackCost of the two tracks is accumulated into the first element. Everything related to events gets duplicated: _event is now _event1 and _event2, and so on. As there can be now two Manipulator actions done with SegmentFsm, this class now completly hide the Manipulator level from the RoutingEvent processing. New function ::bindToTrack() to perform the track insertion. * New: In Katana::TrackCost, add a new ::merge() function. * New: In Katana::TrackElement and Katana::TrackSegment, add symmetric management. Allows to know if a TrackElement has a symmetric and to access it. * New: In Katana::DataSymmetric, add new overload for ::getSymmetrical() to handle DbU::Unit and intervals. * Change: In Katana::RoutingEvent, remove all direct uses of Manipulator objects. Now any change to the event associated segment must go through call to Segment Fsm. * Change: In Katana, adjust the debug level so internal informations are put below level 156. * New: In Hurricane::DbU, in ::getValueString(), special display when the value is Min or Max (more helpful than a gigantic number).
2017-04-30 16:46:33 -05:00
<< " " << DbU::getValueString(optimalMax) << "]" << endl;
First intergration of the Analogic router parts. * New: In Anabatic::AutoSegment, introduce a the kind (associated to a flag) "LongLocal". Analog GCells can be very wide, so at least some carefuly choosen long local segments must be took into account as attractors in the computation of the optimal axis. * New: In Anabatic::AutoSegment::computeOptimal(), take LongLocal into account as attractors. * Change: In ::GCellTopology constructors compare the layers of the RoutingPads using layer masks instead of Layer pointers. Allows to find both "METALx" (symbolic) and "metalX" (real). * Change: In ::GCellTopology::_doHChannel(), _doChannel(), _doStrut() and _doDevice(), tag long locals as "LongLocal". This need to be reviewed as it as bind done a bit too quickly. * Change: In Anabatic::AutoSegment, due too a much bigger span of the analogic GCells the _optimalMin & _optimalMax bitfields must use 16 bits instead of 8 (they where overflowed). * New: In Katana, reorganisation of the initialization procedure to fit both digital and analogic cases. Create an analogInit() method. * Change: In Katana::RoutingEvent, the _tracksNb and _tracksFree bitfields where too short for the Analog GCell size, now uses 16 bits instead of 6. * Bug: In Katana::GraphicKatanEngine::drawGCell(), skip drawing of a GCell if *both* width and height are under 150 pixels. * New: In Katana::Session, add a new isOpen() method.
2016-10-04 10:12:58 -05:00
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
if (getAxis() < optimalMin) {
setAxis( optimalMin, flags );
cdebug_tabw(149,-1);
return true;
}
Synchronize priority of TrackSegments connecteds through doglegs. * Bug: In Anabatic::Dijsktra, correct the indentation in cdebug calls (lots of them causing a big shift right). * New: In Anabatic::TrackSegment, new helper structure SideStack to manage a set of aligned GCells and their various sides sizes. * Change: In Anabatic::TrackSegment::computeOptimal(), more accurate computation of attractors from global segments and variable size GCells using SideStack. * Change: In Katana::DataSymmetric::checkPairing(), increase the tolerance for misaligned symmetrics from 2 to 5 tracks (should be enough for our narrow channel routing). * New: In Katana::TrackSegment, add a first flag to enable locking of priority. If it is set, calls to either "computePriority()" or "forcePriority()" will have no effect. Added the uint32_t flags paraphernalia. * New: In ::computeNetPriority(), overall function to control the call of TrackSegment::computeAlignedPriority(). The call is done from NegociateWindow::run(). * New: Katana::TrackSegment::computeAlignedPriority(), order the TrackSegments aligneds through doglegs to the one with the highest priority is routed first and others progressively from him. Done by forcing an ever decreasing priority on the aligneds ones. The amount of decrease is small so the aligned segments got routed (ordered) in close, if not contiguous, sequence. Priority is locked for the order to remain. * Bug: In Katana::TrackSegment::computePriority(), correct computation of the priority when there is more than 10 free tracks (a DbU::toLambda() call was missing, leading to very big priorities). * Change: In katana::TrackCost CTOR, do not compute a distance to fixed in the case of analog segments, this is backfiring. Slight change of the compare function when delta differs. Seems to improve a little.
2017-05-30 15:33:06 -05:00
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
if (getAxis() > optimalMax) {
setAxis( optimalMax, flags );
cdebug_tabw(149,-1);
return true;
}
if (flags & Flags::Realignate) setAxis( getAxis(), flags );
Select S/T in AutoSegments_Aligneds. Priority moved to TrackSegment. * In Katana::AutoSegments_Aligneds, allow the the caller to select if we are propagating through source or target only (or both). Note that if no flag is given, it is assumed that we want to propagate the old way on both source and target. * In Katana::TrackElement, Katana::TrackSegment and Katana::RoutingEvent, move the event priority from RoutingEvent to TrackSegment. Add accessor and mutators associated in TrackElement (virtual methods).
2017-05-27 13:35:28 -05:00
//setAxis( optimalMin, flags );
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
cdebug_tabw(149,-1);
return false;
}
Replace "unsigned int" by "Flags" in all AutoSegments collections. * Change: In Anabatic::AutoSegments collections, change the type of all the flags that where in "unsigned int" (32 bits) to Flags (uint64_t) as there is now more than 32 flags for functions. * New: In Ababatic::Constants, added new flag Flags::WithPerpands, which makes the number of flags tip over 32 bits, thus making mandatory to uses Flags and not unsigned int. * New: In Anabatic::AutoSegments_Perpandiculars, manage a new flag Flags::WithDoglegs to allow to propagate through global segments that are connecteds via doglegs on local segments. Meaning that there is a good chance that they could be aligned. Slighly change the way we propagate on aligned segments: no longer check for VTee or HTee, but only for same direction and layer as master. * New: In Anabatic & Katana, replace all the "int", "long" and their variants by the less implementation ambiguous "int32_t", "int64_t" (and variant). This should help to better detect bit trucation in flags. Use the type to give a hint about the flags kind: - Type "Flags", for flags shared among Anabatic & Katana functions/methods (may also appear in some objects states). - Type "uint32_t" for flags belonging to an object internal state of from Hurricane functions flags (those should be grouped in a Flag subclass in a perfect world).
2017-05-16 07:53:33 -05:00
void AutoSegment::setAxis ( DbU::Unit axis, Flags flags )
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
{
Added core2chip support for Phenitec80. This commit degrades the run success rate of ARMv2a to 87% (40 iters). * New: In CRLcore/etc/.../kite.conf, add configuration parameters: katana.termSatReservedlocal katana.termSatthreshold for the new edge capacity computation system. * New: In CRLcore/etc/symbolic/phenitec06/, add support for N. Shimizu small I/O pads (supplied in phlib80). Tune various parameters of Anabatic/Katana to increase routing success. * Change: In CRLcore/alliance/ap/ApParser, make Pin external components, so RoutingPad will be build upon in global routing. Do not complain when a I/O pad has a physical instance that did not exists in the netlist. Just create it (appeared in phlib80). When no netlist instance exists in a pad, the pad Cell is still considered as terminal. * New: In Etesian::BloatCells, new profile named "3metals" better suited for two routing metals technologies (i.e. Phenitec). * New: In Anabatic::RawGCellsUnder, new CTOR which take only source & target points instead of a segment. Needed to manage wide segment for which the axis to consider is not that of the segment (one axis for each track it intersect). * New: In Anabatic::GCell, add a RoutingPad count attribute, for Edge reservation computation. * New: In AnabaticEngine::computeEdgeCapacities(), instead of decreasing all edges of a fixed amount (hTrackReservedLocal), guess the GCell cluttering from the number of RoutingPads that it contains. For non-saturated GCells, the four edges are decreased by the number of RoutingPads. We use the maximum from the two neigboring GCells. The hTrackReservedLocal parameter is now used only as a *maximum* that the edge reservation can reach. If a GCell is saturated (more than 8 RoutingPads, the saturation is propagated horizontally to 2 neigboring GCells). * Change: In AutoContactTerminal::getNativeConstraintBox(), use a more flexible gauge name matching for terminal vertical extensions correction. Namely, match all "msxlib*" kind of gauges. * Change: In AutoSegment::setAxis(), add the ability to force the axis position, even if it is a non-canonical segment. Maybe needed in the initialisation steo, before the first canonisation is performed. * New: In NetBuilder, added new methods _do_1G_1PinM1() and _do_2G_1PinM1(), to manage coronas for Phenitec designs. To avoid various side effects from segments being too close from the north / east side of the routing area, make those segments fixeds. * Change: In KatanaEngine::annotateGlobalGraph(), the management of wide wires was wrong. The axis to use to find the underlying GCells is the one of the track, not of the segment. This was creating bad edge capacity computation under the power ring of a block and subsequently routing failures. * New: In Kanata::Manipulator, added method reprocessParallels(), not used though, but keep it anyway, might be of use later... * New: In Kanata::Manipulator, added method avoidBlockage() for terminal METAL2 in non-preferred direction, restrict the terminal and turn constraint box at the current position of the perpandicular, so it doesn't create a deadlock in METAL2. * Change: In SegmentFsm::conflictSolveByPlaceds(), if we cannot break using the whole overlap, try the first atomic overlap. * New: In SegmentFsm::_slackenStrap(), manage conflict between a non-prefered segment and a blockage, this when to call avoidBlockage()... * New: In Katana::Configuration, management of the new edge computation parameters: katana.termSatReservedlocal katana.termSatthreshold * New: In Cumulus/plugins/Core2Chip, support for Phenitec I/O pads.
2019-09-17 10:05:54 -05:00
if (not isCanonical() and not (flags & Flags::Force)) return;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
if ( (axis == getAxis()) and not (flags & Flags::Realignate) ) return;
In AutoSegment slackening, bad positioning of the perpandicular.
2016-08-28 12:09:07 -05:00
cdebug_log(159,0) << "setAxis() @"
First intergration of the Analogic router parts. * New: In Anabatic::AutoSegment, introduce a the kind (associated to a flag) "LongLocal". Analog GCells can be very wide, so at least some carefuly choosen long local segments must be took into account as attractors in the computation of the optimal axis. * New: In Anabatic::AutoSegment::computeOptimal(), take LongLocal into account as attractors. * Change: In ::GCellTopology constructors compare the layers of the RoutingPads using layer masks instead of Layer pointers. Allows to find both "METALx" (symbolic) and "metalX" (real). * Change: In ::GCellTopology::_doHChannel(), _doChannel(), _doStrut() and _doDevice(), tag long locals as "LongLocal". This need to be reviewed as it as bind done a bit too quickly. * Change: In Anabatic::AutoSegment, due too a much bigger span of the analogic GCells the _optimalMin & _optimalMax bitfields must use 16 bits instead of 8 (they where overflowed). * New: In Katana, reorganisation of the initialization procedure to fit both digital and analogic cases. Create an analogInit() method. * Change: In Katana::RoutingEvent, the _tracksNb and _tracksFree bitfields where too short for the Analog GCell size, now uses 16 bits instead of 6. * Bug: In Katana::GraphicKatanEngine::drawGCell(), skip drawing of a GCell if *both* width and height are under 150 pixels. * New: In Katana::Session, add a new isOpen() method.
2016-10-04 10:12:58 -05:00
<< ((isHorizontal())?"Y ":"X ") << DbU::getValueString(getAxis())
<< " to " << DbU::getValueString(axis) << " on " << this << endl;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
cdebug_tabw(145,1);
_setAxis( axis );
if (not isNotAligned()) {
Anabatic transient commit 18. Port of Kite (Katana), Yeah, Baby! Yeah! * Bug: In Hurricane, in StaticObservable::getObserver(), if the slot pointer is NULL, do not try to access the owner. Returns NULL, so the caller can be aware of the situation... * Change: In Hurricane, in BreakpointWidget & ExceptionWidget some cosmetic changes (fonts and window sizes). * Bug: In Anabatic, In AutoHorizontal::getConstraints(), take into account the constraints from the source AutoContact, as it holds the constraints transmitted by the RoutingPads and sets up by propageConstraintsFromRp(). It is likely to be a bug affecting the original Katabatic as well. * Change: In Anabatic, in RawGCellsUnder(), check that the segment is not completly oustside the cell abutment box and truncate the coordinates to the part that is inside. Use the "shrink" if we reach the east/north border. * Change: In Anabatic, in Configuration, no more decorator because we will use a true derived relationship. Katana *derives* from *Anabatic* and do not *decorate* it, so the Configuration can do the same. It also implies that we directly create a Katana engine, not an Anabatic one. * Change: In Anabatic, in Session, do not allow the opening of the Session in a standalone fashion (with a static method). Instead it must be opened using the relevant method of the Anabatic/Katana engine. This ensure we are opening the right Session type. * Change: In Anabatic, in AutoSegment_Aligneds() collection the seed segment is not part of the collection by default, but will be included if the Flags::WithSelf is set. * Change: In Configuration, all the flags value are now defined in two steps. Declared in the header and initialized in the module. This is to prevent the fact that on some cases, in relation with the Python "extern C" part modules, we need a true allocated variable. It was causing weird linking problems. A side effect is that they can no longer be used as entry is switches, have to replace them by if/else. * New: In Anabatic, new GCell::getNeighborAt() utility function. * Bug: In Anabatic, in GCell::doGrid(), tag all the GCells of the grid with the grid type... Back annote all the edges capacity (north & east) with the reserved local capacity. * New: Complete portage of Kite over Anabatic. The new engine is christened "Katana" for Kite-Analogic. When it's capabilities and performances will be on a part with Kite, it is to completly replace it (and take back the "Kite" name). Preliminary tests seems to show that, contrary to intuition (because built on a more complex/slower grid), it is even slightly faster than Kite 8-).
2016-08-15 09:30:13 -05:00
for ( AutoSegment* segment : getAligneds() ) {
segment->_setAxis( getAxis() );
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
}
} else {
cdebug_log(149,0) << "No need to process parallels." << endl;
}
cdebug_tabw(145,-1);
}
void AutoSegment::computeTerminal ()
{
Forgot to perform Track re-order after removing zero-length segments. * Change: In Anabatic::Autocontact, replace getMinDepth() and getMaxDepth() by getDepthSpan(). * New: In Anabatic::AutoSegment::canMoveUp(), add an optional check of low up density (Flags::CheckLowUpDensity). Allows to move up a segment if the up density is (very) low, and in this case it's more efficient than breaking it to fit in the lower layer. canMoveUp() is now able to perform the same work as canPivotUp() if *not* supplied the flag Flags::IgnoreContacts. * New: In Katana, in GlobalRouting::DigitalDistance() now take into account the cost of a VIA (currently set to 2.5). Need to known the Net currently routed in the DigitalDistance object itself. * Change: In Katana::Track::Element::canPivotUp(), now also takes a flag parameter. * Change: In Katana::Manipulator, new flag IgnoreContacts to mirror the one of Anabatic. * Change: In Katana::SegmentFsm, allocate once a Manipulator object instead of many times on the fly. In SegmentFsm::_slackenGlobal(), in the slacken state, if the up density is (very) low, bypass to move up instead of slackening. This solve better the routing of the control part of the register file. The register file having a pathological case of terminal placement: many punctual terminals aligneds in METAL2 *and* a grid of METAL2 and METAL3 blockages near below... * Bug: In Katana::Session::_revalidate(), after removing the zero-length segments, forgot to re-order the track, leading to many stranges effects as the indexes where no longer coherent in the Track.
2016-09-20 04:30:45 -05:00
AutoContact* source = getAutoSource();
AutoContact* target = getAutoTarget();
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
if (source->isTerminal()) {
unsetFlags( SegWeakTerminal );
setFlags ( SegSourceTerminal );
Forgot to perform Track re-order after removing zero-length segments. * Change: In Anabatic::Autocontact, replace getMinDepth() and getMaxDepth() by getDepthSpan(). * New: In Anabatic::AutoSegment::canMoveUp(), add an optional check of low up density (Flags::CheckLowUpDensity). Allows to move up a segment if the up density is (very) low, and in this case it's more efficient than breaking it to fit in the lower layer. canMoveUp() is now able to perform the same work as canPivotUp() if *not* supplied the flag Flags::IgnoreContacts. * New: In Katana, in GlobalRouting::DigitalDistance() now take into account the cost of a VIA (currently set to 2.5). Need to known the Net currently routed in the DigitalDistance object itself. * Change: In Katana::Track::Element::canPivotUp(), now also takes a flag parameter. * Change: In Katana::Manipulator, new flag IgnoreContacts to mirror the one of Anabatic. * Change: In Katana::SegmentFsm, allocate once a Manipulator object instead of many times on the fly. In SegmentFsm::_slackenGlobal(), in the slacken state, if the up density is (very) low, bypass to move up instead of slackening. This solve better the routing of the control part of the register file. The register file having a pathological case of terminal placement: many punctual terminals aligneds in METAL2 *and* a grid of METAL2 and METAL3 blockages near below... * Bug: In Katana::Session::_revalidate(), after removing the zero-length segments, forgot to re-order the track, leading to many stranges effects as the indexes where no longer coherent in the Track.
2016-09-20 04:30:45 -05:00
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
if (not target->isTerminal())
target->setFlags( CntWeakTerminal );
} else if (target->isTerminal()) {
unsetFlags( SegWeakTerminal );
setFlags ( SegTargetTerminal );
Forgot to perform Track re-order after removing zero-length segments. * Change: In Anabatic::Autocontact, replace getMinDepth() and getMaxDepth() by getDepthSpan(). * New: In Anabatic::AutoSegment::canMoveUp(), add an optional check of low up density (Flags::CheckLowUpDensity). Allows to move up a segment if the up density is (very) low, and in this case it's more efficient than breaking it to fit in the lower layer. canMoveUp() is now able to perform the same work as canPivotUp() if *not* supplied the flag Flags::IgnoreContacts. * New: In Katana, in GlobalRouting::DigitalDistance() now take into account the cost of a VIA (currently set to 2.5). Need to known the Net currently routed in the DigitalDistance object itself. * Change: In Katana::Track::Element::canPivotUp(), now also takes a flag parameter. * Change: In Katana::Manipulator, new flag IgnoreContacts to mirror the one of Anabatic. * Change: In Katana::SegmentFsm, allocate once a Manipulator object instead of many times on the fly. In SegmentFsm::_slackenGlobal(), in the slacken state, if the up density is (very) low, bypass to move up instead of slackening. This solve better the routing of the control part of the register file. The register file having a pathological case of terminal placement: many punctual terminals aligneds in METAL2 *and* a grid of METAL2 and METAL3 blockages near below... * Bug: In Katana::Session::_revalidate(), after removing the zero-length segments, forgot to re-order the track, leading to many stranges effects as the indexes where no longer coherent in the Track.
2016-09-20 04:30:45 -05:00
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
if (not source->isTerminal())
source->setFlags( CntWeakTerminal );
} else {
Katana manage wide wires, and they can also be symmetric. * New: In Anabatic::AutoContact and the derived classes, manages wide wires. The contact self dimension itself according to the segments it is connected to. Special case for the AutoContactTerminal which also read the size of the component it is anchored upon. New refresh method "updateSize()" and flag CntInvalidatedWidth. to compute the size. In AutoContactTerminal, compute the constraint box according to the width of the segment. * New: In Anabatic::AutoSegment, flags are now implemented as "static const" attributes of the class. The flags are stored into a uint64_t as they are more than 32. Added new flag "SegWide" and associated predicates. * Change: In GCellTopology::_doHChannel() and GCellTopology::_doVChannel(), uses the simpler overload of AutoSegment::create() in order to detect the wire width automatically. * New: In Katana::Manipulator, split insertToTrack() and forceToTrack() into a one-track method and a segment level method that iterate over the track span of the segment. * New: In Katana::SegmentFsm, for each cost in the table, now allow access to a specific track. So the base functions have now two parameters: "icost" and "itrack" (has a cost can have multiple tracks in the case of wide segments). * Change: In Katana::TrackElement, remove the index of the element inside it's track, as for a wide segment it will not be meaningful for the non-base track. This means that we have to use the Track::find() method each time instead. Remove the wide flag, as it is a duplicate of the one in AutoSegment. Added a getTrackCount() method to tell the number of track the segment is inserted into. Needed in the Track destroy step to delete a segment only when the last track that refers it is destroyed. Added getSymmetricAxis() to correct the computation of the symmetric base track in case of wide segment as the base track is not centered but the the leftmost one. * Change: In Track::insert() insert wide segments in their whole track span. * Change: In TrackCost, create an array of costs according to the segment track span. * Change: In TrackSegment::create(), now activate the factory and create wide segments. * Bug: In Katana::AutoSegments_Perpandicular, correct the debug indentation problem (ever shifting to the right).
2017-07-28 08:30:22 -05:00
uint64_t terminalFlag = 0;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
switch ( _getFlags() & SegWeakTerminal ) {
case 0: break;
case SegSourceTerminal|SegTargetTerminal:
case SegSourceTerminal:
case SegTargetTerminal: terminalFlag = SegWeakTerminal1; break;
case SegWeakTerminal1: terminalFlag = SegWeakTerminal1; break;
case SegWeakTerminal2: terminalFlag = SegWeakTerminal2; break;
default:
Forgot to switch terminal flags when orienting AutoSegments (again!).
2016-08-28 09:12:05 -05:00
cerr << Warning("%s has multiple terminal flag sets:%s (%x)."
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
,getString(this).c_str()
Forgot to switch terminal flags when orienting AutoSegments (again!).
2016-08-28 09:12:05 -05:00
,_getStringFlags().c_str()
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
,_flags
) << endl;
terminalFlag = SegWeakTerminal2; break;
}
unsetFlags( SegWeakTerminal );
setFlags ( terminalFlag );
}
Forgot to perform Track re-order after removing zero-length segments. * Change: In Anabatic::Autocontact, replace getMinDepth() and getMaxDepth() by getDepthSpan(). * New: In Anabatic::AutoSegment::canMoveUp(), add an optional check of low up density (Flags::CheckLowUpDensity). Allows to move up a segment if the up density is (very) low, and in this case it's more efficient than breaking it to fit in the lower layer. canMoveUp() is now able to perform the same work as canPivotUp() if *not* supplied the flag Flags::IgnoreContacts. * New: In Katana, in GlobalRouting::DigitalDistance() now take into account the cost of a VIA (currently set to 2.5). Need to known the Net currently routed in the DigitalDistance object itself. * Change: In Katana::Track::Element::canPivotUp(), now also takes a flag parameter. * Change: In Katana::Manipulator, new flag IgnoreContacts to mirror the one of Anabatic. * Change: In Katana::SegmentFsm, allocate once a Manipulator object instead of many times on the fly. In SegmentFsm::_slackenGlobal(), in the slacken state, if the up density is (very) low, bypass to move up instead of slackening. This solve better the routing of the control part of the register file. The register file having a pathological case of terminal placement: many punctual terminals aligneds in METAL2 *and* a grid of METAL2 and METAL3 blockages near below... * Bug: In Katana::Session::_revalidate(), after removing the zero-length segments, forgot to re-order the track, leading to many stranges effects as the indexes where no longer coherent in the Track.
2016-09-20 04:30:45 -05:00
cdebug_log(145,0) << "computeTerminal() S:" << source->isTerminal()
<< " T:" << target->isTerminal()
<< " " << this << endl;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
}
void AutoSegment::computeOptimal ( set<AutoSegment*>& processeds )
{
cdebug_log(145,1) << "computeOptimal() - " << this << endl;
Added analog type on segment NetRoutingProperty. * New: In Anabatic_AutoSegments collection, added a Flag to the constructors to allow different behavior between digital and analog modes. For "Aligneds" and "Perpandiculars" collections, now manage a new Flag WithDoglegs to follow aligned globals through local doglegs (for analog nets). Adjust the log level of collections to 144 (formerly 145). * New: In Anabatic::AutoSegment, new flag SegAnalog for segments that are part of an analog net. Note that with this flag, we reach the 32 bits limit... * Change: In Anabatic::Constants, Flags are now declared as BaseFlags objects and *not* uint64_t. This avoids overload resolution problems with arithmetical overload of the operators. The BaseFlags/Flags types are now completly "isolated" from the uint64_t, it has the advantage of showing where unwanted previous implicit conversions where occuring. * Change: In Katana::Constants, Flags values are now of BaseFlags type instead of uint64_t. * Change: In Anabatic::Dijkstra, lots of log cleanup. * Change: In Anabatic::GCell::getSide(), make the "shrink" parameter visible to allow to substract the topmost and rightmost track for axis span computation in AutoSegment::computeOptimal(). Used for analog mode. * Change: In NetRoutingState, added a flag for analog mode. Use uint32_t for the flags type. * New: In Isobar, export the NetRoutingState and NetRoutingExtension objects.
2017-05-20 05:33:12 -05:00
DbU::Unit optimalMin;
DbU::Unit optimalMax;
DbU::Unit constraintMin;
DbU::Unit constraintMax;
vector<AutoSegment*> aligneds;
Synchronize priority of TrackSegments connecteds through doglegs. * Bug: In Anabatic::Dijsktra, correct the indentation in cdebug calls (lots of them causing a big shift right). * New: In Anabatic::TrackSegment, new helper structure SideStack to manage a set of aligned GCells and their various sides sizes. * Change: In Anabatic::TrackSegment::computeOptimal(), more accurate computation of attractors from global segments and variable size GCells using SideStack. * Change: In Katana::DataSymmetric::checkPairing(), increase the tolerance for misaligned symmetrics from 2 to 5 tracks (should be enough for our narrow channel routing). * New: In Katana::TrackSegment, add a first flag to enable locking of priority. If it is set, calls to either "computePriority()" or "forcePriority()" will have no effect. Added the uint32_t flags paraphernalia. * New: In ::computeNetPriority(), overall function to control the call of TrackSegment::computeAlignedPriority(). The call is done from NegociateWindow::run(). * New: Katana::TrackSegment::computeAlignedPriority(), order the TrackSegments aligneds through doglegs to the one with the highest priority is routed first and others progressively from him. Done by forcing an ever decreasing priority on the aligneds ones. The amount of decrease is small so the aligned segments got routed (ordered) in close, if not contiguous, sequence. Priority is locked for the order to remain. * Bug: In Katana::TrackSegment::computePriority(), correct computation of the priority when there is more than 10 free tracks (a DbU::toLambda() call was missing, leading to very big priorities). * Change: In katana::TrackCost CTOR, do not compute a distance to fixed in the case of analog segments, this is backfiring. Slight change of the compare function when delta differs. Seems to improve a little.
2017-05-30 15:33:06 -05:00
SideStack sideStack ( (isHorizontal() ? Flags::Horizontal : Flags::Vertical), getPitch() );
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
getConstraints( constraintMin, constraintMax );
Katana routing algorithm adjustements for analog routing. * Bug: In Hurricane::Interval, the ::getSize() method was returning a negative length when the Interval was empty. Now return zero. This was causing slight cost functions side effect when computing the overlap of a segment with another one belonging to the same net (i.e. shareDelta) * Bug: In Katana::AutoSegment::computeOptimal(), checks that the optimal position for the axis is indeed inside the GCell interval. * Bug: In Katana::DataSymmetric::getSymmetrical(Interval&), reverse the shrink applied to GCells sides. Interval are of the type [min:max[ so after symmetric transformation they become ]smax:smin]. Apply a correction so that they are [smax:smin[ and could be merged with the paired interval (the one without symmetry applied). * Bug: In katana::TrackCost::consolidate(), the overlap (now) always positive, and must be *substracted* to the "delta" (not *added*). * Bug: In Katana::_computeNetOptimals(), now skip the already processeds AutoSegments instead of re-processing them. * New: In Katana::AutoSegment::computeOptimal(), improved computation of the axis hint: - For analog net, propagate through the simple doglegs to find attractors. - Restrict the allowed interval to the union of GCell sides and segments constraints that are terminals. - Consider a local as "long" if it's length exceed 20 the P-pitch. (maybe make that a tool parameter).
2017-05-22 17:20:31 -05:00
cdebug_log(145,0) << "Constraints: [" << DbU::getValueString(constraintMin)
<< " " << DbU::getValueString(constraintMax) << "]" << endl;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
Basic support for FreePDK 45 completed. * New: In Commons, inspector support for std::pair<T,U>. * New: In Hurricane::Layer, ContactLayer & ViaLayer, support for non square VIAs. The hole (cut) remains square, but the various metal extensions can now be different in X and Y. The ::getEnclosure() method now takes a flag EnclosureH / EnclosureV. * New: In Hurricane::DbU, inspector support for: std::pair<DbU::Unit,DbU::Unit> std::array<DbU::Unit,3> Must be defined here as DbU do not exists yet in Commons.h * Bug: In Hurricane::Interval::getSize(), when the interval is "full span", do not return the difference between min and max, but directly DbU::Max. (the previous result was -1 !) * New: In CRL Core Python/Technology.py, support for non square VIAs in the configuration files. Applied to FreePDK 45. * New: In CRL::RoutingGauge, added a "symbolic" flag to tell if a gauge is for symbolic layout or not. Exported to Python. * New: In Anabatic::AutoHorizontal::updatePosition(), differentiated computation for soure or target taking account of the VIA extension in the right segment metal (due to non-square VIAs). * Change: In Anabatic::AutoHorizontal::_makeDogleg(), the dogleg is UP for HV gauges and DOWN for VH. * New: In Anabatic::AutoSegment::_initialize(), create a cache of the various extension length for each layer (viaToTop, viaToBottom, viaToSame). New implementation of getExtensionCap() using the previous cached extension table. See updatePositions(). New static functions to access the extension cache in the header: getViaTotopCap() ... * Change: In Anabatic::AutoSegment, in various update methods, updateOrient() must always be called *before* updatePositions() as extensions are dependant on source/target. * New: In Anabatic::AutoSegment::getEndAxes() compute the position of the first source and last target position (center/axes) on an *aligned* set of segments. * New: In Anabatic::AutoSegment, add a new state flag SegAxisFixed to signal segments that can be put on only one track. Specific case to VH gauge for a M1 vertical terminal with a M2 vertical segment. The M2 is effectively bound to the M1 axis position. * Bug: In Anabatic::NetBuilderVH::_do_xG_xM1_xM3(), in case of E/W global and only one RoutingPad the connexion to the RoutingPad was duplicated. It was valid, but totally stupid. * Bug: In Anabatic::Session::_canonize(), for an aligned segment set, intersect the user constraints from all segments instead of only considering the canonical one. Issue a warning about too tight constraints only for symbolic gauges. It may be correct for the real ones. * New: In Katata::DataNegociate::update(), more accurate computation of the perpandicular free interval. Use segment extension cap calculation. Create a special case for fixed axis segments allowing them to find alternative free interval, try under source and under target as they are likely to be draggable segments. * Change: In Katana::Manipulator::relax(), use the extension cap value to compute the axis of the perpandicular segemnts. * Change: In Katana::Manipulator::moveUp(), now move up the whole set of aligned segments instead of just the canonical one. * Change: In Katana::NegociateWindow::loadRoutingPads(), more accurate TrackMarkers insertions for fixed terminals. * New: In Katana::RoutingEvent::Key::Compare::operator(), segments with fixed axis are processed prior to any others. * New: In Katana::RoutingEventLoop, store segment pointers instead of ids to generate more accurate error messages. * Change: In Katana::RoutingPlane::create(), perform local track assignment only for HV gauges. * Change: In Katana::SegmentFsm::_slackenLocal(), add a "dragMinimize" step in the automaton. Mutliple states transitions can occurs in a row if an action fails. * New: In Katana::Session::_toIntervalAxis(), normalize interval bounds so they are on track positions (by shrinking the interval). * Bug: In Katana::TrackMarker CTOR, the weigh computation was wrong.
2018-02-17 13:27:38 -06:00
AutoContact* source = getAutoSource();
AutoContact* target = getAutoTarget();
Determinism correction of Hurricane, CRL core & Katana. * Bug: In Hurricane::Cell::uniquify(), a set<Cell*> sorted on pointers was remaining. Now sorted on Entity::Id. * Bug: In CRL::VhdlEntity, in the driver, the components where driven in pointer order (set<> again). Now use ids. * Bug: In CRL/etc/scn6m_deep_09/technology.conf, the symbolic extentions for VIAs and layers were wrong. Have to be multiplied by two. * Change: In AnabaticEngine, AutoContact and AutoSegment LUTs are now sorted on Entity::Id. Should not have had any impact, but better safe than sorry. * Change: In KatanaEngine, Symmetric contraint map<> is now sorted on Entity::id. Idem for TrackSegmentLut.
2018-03-24 18:58:22 -05:00
Basic support for FreePDK 45 completed. * New: In Commons, inspector support for std::pair<T,U>. * New: In Hurricane::Layer, ContactLayer & ViaLayer, support for non square VIAs. The hole (cut) remains square, but the various metal extensions can now be different in X and Y. The ::getEnclosure() method now takes a flag EnclosureH / EnclosureV. * New: In Hurricane::DbU, inspector support for: std::pair<DbU::Unit,DbU::Unit> std::array<DbU::Unit,3> Must be defined here as DbU do not exists yet in Commons.h * Bug: In Hurricane::Interval::getSize(), when the interval is "full span", do not return the difference between min and max, but directly DbU::Max. (the previous result was -1 !) * New: In CRL Core Python/Technology.py, support for non square VIAs in the configuration files. Applied to FreePDK 45. * New: In CRL::RoutingGauge, added a "symbolic" flag to tell if a gauge is for symbolic layout or not. Exported to Python. * New: In Anabatic::AutoHorizontal::updatePosition(), differentiated computation for soure or target taking account of the VIA extension in the right segment metal (due to non-square VIAs). * Change: In Anabatic::AutoHorizontal::_makeDogleg(), the dogleg is UP for HV gauges and DOWN for VH. * New: In Anabatic::AutoSegment::_initialize(), create a cache of the various extension length for each layer (viaToTop, viaToBottom, viaToSame). New implementation of getExtensionCap() using the previous cached extension table. See updatePositions(). New static functions to access the extension cache in the header: getViaTotopCap() ... * Change: In Anabatic::AutoSegment, in various update methods, updateOrient() must always be called *before* updatePositions() as extensions are dependant on source/target. * New: In Anabatic::AutoSegment::getEndAxes() compute the position of the first source and last target position (center/axes) on an *aligned* set of segments. * New: In Anabatic::AutoSegment, add a new state flag SegAxisFixed to signal segments that can be put on only one track. Specific case to VH gauge for a M1 vertical terminal with a M2 vertical segment. The M2 is effectively bound to the M1 axis position. * Bug: In Anabatic::NetBuilderVH::_do_xG_xM1_xM3(), in case of E/W global and only one RoutingPad the connexion to the RoutingPad was duplicated. It was valid, but totally stupid. * Bug: In Anabatic::Session::_canonize(), for an aligned segment set, intersect the user constraints from all segments instead of only considering the canonical one. Issue a warning about too tight constraints only for symbolic gauges. It may be correct for the real ones. * New: In Katata::DataNegociate::update(), more accurate computation of the perpandicular free interval. Use segment extension cap calculation. Create a special case for fixed axis segments allowing them to find alternative free interval, try under source and under target as they are likely to be draggable segments. * Change: In Katana::Manipulator::relax(), use the extension cap value to compute the axis of the perpandicular segemnts. * Change: In Katana::Manipulator::moveUp(), now move up the whole set of aligned segments instead of just the canonical one. * Change: In Katana::NegociateWindow::loadRoutingPads(), more accurate TrackMarkers insertions for fixed terminals. * New: In Katana::RoutingEvent::Key::Compare::operator(), segments with fixed axis are processed prior to any others. * New: In Katana::RoutingEventLoop, store segment pointers instead of ids to generate more accurate error messages. * Change: In Katana::RoutingPlane::create(), perform local track assignment only for HV gauges. * Change: In Katana::SegmentFsm::_slackenLocal(), add a "dragMinimize" step in the automaton. Mutliple states transitions can occurs in a row if an action fails. * New: In Katana::Session::_toIntervalAxis(), normalize interval bounds so they are on track positions (by shrinking the interval). * Bug: In Katana::TrackMarker CTOR, the weigh computation was wrong.
2018-02-17 13:27:38 -06:00
if (isLocal() and source->isTurn() and target->isTurn() and not isUserDefined()) {
AutoSegment* sourcePerpand = source->getPerpandicular(this);
AutoSegment* targetPerpand = target->getPerpandicular(this);
sourcePerpand->updateOrient();
targetPerpand->updateOrient();
if (not ( (sourcePerpand->getAutoSource() == source)
xor (targetPerpand->getAutoSource() == target) ) ) {
// This is a U-Turn.
Validating channel routing mode (two metals) on SNX. * New: In Hurricane::Entity, add an id counter limit and a memory size limit. The two limits are checked only when a new Entity object is created. This should help avoiding massive memory links. * New: In CRL Core, add a "crlcore.groundName" and a "crlcore.powerName" parameter to specify the name of the ground/power signals to be created if they are missing in a Cell. For Alliance libraries it would be "vss" & "vdd" (default values), but for real technologies, it is often "gnd!" & "vdd!". The Blif parser is modificated to make use of it. * Bug: In AnabaticEngine::unify(), set the resulting unified segment in the center of the GCells common side. Gcells under a segment are found by using the edge that cover the segment axis. When we have a "bend" GCell stack and the axis is wrong, they could be ommited. This was causing deleted segments to be not removed from some Edges, then core dump. * Change: In Anabatic::AutoSegment::create(), smarter choosing of the reference contact, select the fixed or terminal one instead of always the source one. * New: In Anabatic::Edge::isEnding(), new function to check if a segment going through an Edge is starting/ending in either source or target GCell of the edge (active only when running in channel mode). * New: In Anabatic::Edge::add(), a segment takes part in the occupancy only if it is not ending in either source or target (channel mode only). The occupancy due to terminal is pre-computed in Katana. * New: In Anabatic::Edge::ripup(), in channel mode, never ripup a segment which is ending in either source or target (we *have* to access this edge to connect to the terminal). * Bug: In Anabatic::GCell::hcut() and vcut(), force the update of the Edge which is on the side that will get splitted by the cut. It's capacity will be reduced to it must be updated. * Change: In Anabatic::GCell::updateGContacts() add a flag to conditionnally update horizontals or verticals only. We may require only a partial update when resizing the GCell in only one direction. This, again, related to the fact that we compute the GCells under a segment thanks to it's axis position, so we need to be very careful when modificating axis. * Change: In Katana::Block::resizeChannels(), only update GContact vertical position. Do not disturb X positions of segments. * Bug: In Katana::GlobalRoute::DigitalDistance, in channel mode, some Edges can have a zero capacity, but still be reachable if the net has a terminal in either source or target. Look for this case and return a distance of zero instead of "unreachable". This was causing the global routing not to complete in channel mode. For computing the edge distance, makes the vertical edges much more long (10 times) than the horizontal ones as the vertical capacity is very limited. Hard coded for now, should make it a parameter in the future. * Change: In KatanaEngine::annotateGlobalGraph(), decrease the capacity of edges with reserveCapacity for each terminal inside a GCell. Both north and south edges are decreased as we a terminal will block both north and south edges. As a counterpart, the Edge capacity is not decreased when the global router connect to a terminal. * Change: In Katana::RoutingEvent::revalidate(), when in repair stage, do not expand the slack for horizontal segments in channel mode. So they may not overlap the standard cell row. * Bug: In Stratus documentation, do not use the french option in babel, the documentation is in english! * New: In Documentation, added Hurricane/Python tutorial, part for drawing layout.
2018-03-16 10:20:04 -05:00
cdebug_log(145,0) << "U-Turn special case." << endl;
Basic support for FreePDK 45 completed. * New: In Commons, inspector support for std::pair<T,U>. * New: In Hurricane::Layer, ContactLayer & ViaLayer, support for non square VIAs. The hole (cut) remains square, but the various metal extensions can now be different in X and Y. The ::getEnclosure() method now takes a flag EnclosureH / EnclosureV. * New: In Hurricane::DbU, inspector support for: std::pair<DbU::Unit,DbU::Unit> std::array<DbU::Unit,3> Must be defined here as DbU do not exists yet in Commons.h * Bug: In Hurricane::Interval::getSize(), when the interval is "full span", do not return the difference between min and max, but directly DbU::Max. (the previous result was -1 !) * New: In CRL Core Python/Technology.py, support for non square VIAs in the configuration files. Applied to FreePDK 45. * New: In CRL::RoutingGauge, added a "symbolic" flag to tell if a gauge is for symbolic layout or not. Exported to Python. * New: In Anabatic::AutoHorizontal::updatePosition(), differentiated computation for soure or target taking account of the VIA extension in the right segment metal (due to non-square VIAs). * Change: In Anabatic::AutoHorizontal::_makeDogleg(), the dogleg is UP for HV gauges and DOWN for VH. * New: In Anabatic::AutoSegment::_initialize(), create a cache of the various extension length for each layer (viaToTop, viaToBottom, viaToSame). New implementation of getExtensionCap() using the previous cached extension table. See updatePositions(). New static functions to access the extension cache in the header: getViaTotopCap() ... * Change: In Anabatic::AutoSegment, in various update methods, updateOrient() must always be called *before* updatePositions() as extensions are dependant on source/target. * New: In Anabatic::AutoSegment::getEndAxes() compute the position of the first source and last target position (center/axes) on an *aligned* set of segments. * New: In Anabatic::AutoSegment, add a new state flag SegAxisFixed to signal segments that can be put on only one track. Specific case to VH gauge for a M1 vertical terminal with a M2 vertical segment. The M2 is effectively bound to the M1 axis position. * Bug: In Anabatic::NetBuilderVH::_do_xG_xM1_xM3(), in case of E/W global and only one RoutingPad the connexion to the RoutingPad was duplicated. It was valid, but totally stupid. * Bug: In Anabatic::Session::_canonize(), for an aligned segment set, intersect the user constraints from all segments instead of only considering the canonical one. Issue a warning about too tight constraints only for symbolic gauges. It may be correct for the real ones. * New: In Katata::DataNegociate::update(), more accurate computation of the perpandicular free interval. Use segment extension cap calculation. Create a special case for fixed axis segments allowing them to find alternative free interval, try under source and under target as they are likely to be draggable segments. * Change: In Katana::Manipulator::relax(), use the extension cap value to compute the axis of the perpandicular segemnts. * Change: In Katana::Manipulator::moveUp(), now move up the whole set of aligned segments instead of just the canonical one. * Change: In Katana::NegociateWindow::loadRoutingPads(), more accurate TrackMarkers insertions for fixed terminals. * New: In Katana::RoutingEvent::Key::Compare::operator(), segments with fixed axis are processed prior to any others. * New: In Katana::RoutingEventLoop, store segment pointers instead of ids to generate more accurate error messages. * Change: In Katana::RoutingPlane::create(), perform local track assignment only for HV gauges. * Change: In Katana::SegmentFsm::_slackenLocal(), add a "dragMinimize" step in the automaton. Mutliple states transitions can occurs in a row if an action fails. * New: In Katana::Session::_toIntervalAxis(), normalize interval bounds so they are on track positions (by shrinking the interval). * Bug: In Katana::TrackMarker CTOR, the weigh computation was wrong.
2018-02-17 13:27:38 -06:00
DbU::Unit optimal = 0;
if (sourcePerpand->getAutoSource() == source) {
optimal = std::min( sourcePerpand->getTargetU(), targetPerpand->getTargetU() );
optimal = std::min( optimal, constraintMax );
} else {
optimal = std::max( sourcePerpand->getSourceU(), targetPerpand->getSourceU() );
optimal = std::max( optimal, constraintMin );
}
cdebug_log(145,0) << "| Source perpandicular: " << sourcePerpand << endl;
cdebug_log(145,0) << "| Target perpandicular: " << targetPerpand << endl;
cdebug_log(145,0) << "Applying constraint (U-Turn) on: " << this << endl;
cdebug_log(145,0) << "optimal: " << DbU::getValueString(optimal) << endl;
setOptimalMin( optimal );
setOptimalMax( optimal );
processeds.insert( this );
cdebug_tabw(145,-1);
return;
}
}
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
if (isUserDefined()) {
optimalMin = optimalMax = getAxis();
Added analog type on segment NetRoutingProperty. * New: In Anabatic_AutoSegments collection, added a Flag to the constructors to allow different behavior between digital and analog modes. For "Aligneds" and "Perpandiculars" collections, now manage a new Flag WithDoglegs to follow aligned globals through local doglegs (for analog nets). Adjust the log level of collections to 144 (formerly 145). * New: In Anabatic::AutoSegment, new flag SegAnalog for segments that are part of an analog net. Note that with this flag, we reach the 32 bits limit... * Change: In Anabatic::Constants, Flags are now declared as BaseFlags objects and *not* uint64_t. This avoids overload resolution problems with arithmetical overload of the operators. The BaseFlags/Flags types are now completly "isolated" from the uint64_t, it has the advantage of showing where unwanted previous implicit conversions where occuring. * Change: In Katana::Constants, Flags values are now of BaseFlags type instead of uint64_t. * Change: In Anabatic::Dijkstra, lots of log cleanup. * Change: In Anabatic::GCell::getSide(), make the "shrink" parameter visible to allow to substract the topmost and rightmost track for axis span computation in AutoSegment::computeOptimal(). Used for analog mode. * Change: In NetRoutingState, added a flag for analog mode. Use uint32_t for the flags type. * New: In Isobar, export the NetRoutingState and NetRoutingExtension objects.
2017-05-20 05:33:12 -05:00
aligneds.push_back( this );
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
} else {
DbU::Unit terminalMin;
DbU::Unit terminalMax;
AttractorsMap attractors;
Added analog type on segment NetRoutingProperty. * New: In Anabatic_AutoSegments collection, added a Flag to the constructors to allow different behavior between digital and analog modes. For "Aligneds" and "Perpandiculars" collections, now manage a new Flag WithDoglegs to follow aligned globals through local doglegs (for analog nets). Adjust the log level of collections to 144 (formerly 145). * New: In Anabatic::AutoSegment, new flag SegAnalog for segments that are part of an analog net. Note that with this flag, we reach the 32 bits limit... * Change: In Anabatic::Constants, Flags are now declared as BaseFlags objects and *not* uint64_t. This avoids overload resolution problems with arithmetical overload of the operators. The BaseFlags/Flags types are now completly "isolated" from the uint64_t, it has the advantage of showing where unwanted previous implicit conversions where occuring. * Change: In Katana::Constants, Flags values are now of BaseFlags type instead of uint64_t. * Change: In Anabatic::Dijkstra, lots of log cleanup. * Change: In Anabatic::GCell::getSide(), make the "shrink" parameter visible to allow to substract the topmost and rightmost track for axis span computation in AutoSegment::computeOptimal(). Used for analog mode. * Change: In NetRoutingState, added a flag for analog mode. Use uint32_t for the flags type. * New: In Isobar, export the NetRoutingState and NetRoutingExtension objects.
2017-05-20 05:33:12 -05:00
Flags flags = (isAnalog() ? Flags::WithDoglegs : Flags::NoFlags);
getAligneds( Flags::WithSelf|flags ).fill( aligneds );
Synchronize priority of TrackSegments connecteds through doglegs. * Bug: In Anabatic::Dijsktra, correct the indentation in cdebug calls (lots of them causing a big shift right). * New: In Anabatic::TrackSegment, new helper structure SideStack to manage a set of aligned GCells and their various sides sizes. * Change: In Anabatic::TrackSegment::computeOptimal(), more accurate computation of attractors from global segments and variable size GCells using SideStack. * Change: In Katana::DataSymmetric::checkPairing(), increase the tolerance for misaligned symmetrics from 2 to 5 tracks (should be enough for our narrow channel routing). * New: In Katana::TrackSegment, add a first flag to enable locking of priority. If it is set, calls to either "computePriority()" or "forcePriority()" will have no effect. Added the uint32_t flags paraphernalia. * New: In ::computeNetPriority(), overall function to control the call of TrackSegment::computeAlignedPriority(). The call is done from NegociateWindow::run(). * New: Katana::TrackSegment::computeAlignedPriority(), order the TrackSegments aligneds through doglegs to the one with the highest priority is routed first and others progressively from him. Done by forcing an ever decreasing priority on the aligneds ones. The amount of decrease is small so the aligned segments got routed (ordered) in close, if not contiguous, sequence. Priority is locked for the order to remain. * Bug: In Katana::TrackSegment::computePriority(), correct computation of the priority when there is more than 10 free tracks (a DbU::toLambda() call was missing, leading to very big priorities). * Change: In katana::TrackCost CTOR, do not compute a distance to fixed in the case of analog segments, this is backfiring. Slight change of the compare function when delta differs. Seems to improve a little.
2017-05-30 15:33:06 -05:00
if (getGCell()->isMatrix()) {
sideStack.addGCell( getGCell() );
} else {
Added analog type on segment NetRoutingProperty. * New: In Anabatic_AutoSegments collection, added a Flag to the constructors to allow different behavior between digital and analog modes. For "Aligneds" and "Perpandiculars" collections, now manage a new Flag WithDoglegs to follow aligned globals through local doglegs (for analog nets). Adjust the log level of collections to 144 (formerly 145). * New: In Anabatic::AutoSegment, new flag SegAnalog for segments that are part of an analog net. Note that with this flag, we reach the 32 bits limit... * Change: In Anabatic::Constants, Flags are now declared as BaseFlags objects and *not* uint64_t. This avoids overload resolution problems with arithmetical overload of the operators. The BaseFlags/Flags types are now completly "isolated" from the uint64_t, it has the advantage of showing where unwanted previous implicit conversions where occuring. * Change: In Katana::Constants, Flags values are now of BaseFlags type instead of uint64_t. * Change: In Anabatic::Dijkstra, lots of log cleanup. * Change: In Anabatic::GCell::getSide(), make the "shrink" parameter visible to allow to substract the topmost and rightmost track for axis span computation in AutoSegment::computeOptimal(). Used for analog mode. * Change: In NetRoutingState, added a flag for analog mode. Use uint32_t for the flags type. * New: In Isobar, export the NetRoutingState and NetRoutingExtension objects.
2017-05-20 05:33:12 -05:00
vector<GCell*> gcells;
Synchronize priority of TrackSegments connecteds through doglegs. * Bug: In Anabatic::Dijsktra, correct the indentation in cdebug calls (lots of them causing a big shift right). * New: In Anabatic::TrackSegment, new helper structure SideStack to manage a set of aligned GCells and their various sides sizes. * Change: In Anabatic::TrackSegment::computeOptimal(), more accurate computation of attractors from global segments and variable size GCells using SideStack. * Change: In Katana::DataSymmetric::checkPairing(), increase the tolerance for misaligned symmetrics from 2 to 5 tracks (should be enough for our narrow channel routing). * New: In Katana::TrackSegment, add a first flag to enable locking of priority. If it is set, calls to either "computePriority()" or "forcePriority()" will have no effect. Added the uint32_t flags paraphernalia. * New: In ::computeNetPriority(), overall function to control the call of TrackSegment::computeAlignedPriority(). The call is done from NegociateWindow::run(). * New: Katana::TrackSegment::computeAlignedPriority(), order the TrackSegments aligneds through doglegs to the one with the highest priority is routed first and others progressively from him. Done by forcing an ever decreasing priority on the aligneds ones. The amount of decrease is small so the aligned segments got routed (ordered) in close, if not contiguous, sequence. Priority is locked for the order to remain. * Bug: In Katana::TrackSegment::computePriority(), correct computation of the priority when there is more than 10 free tracks (a DbU::toLambda() call was missing, leading to very big priorities). * Change: In katana::TrackCost CTOR, do not compute a distance to fixed in the case of analog segments, this is backfiring. Slight change of the compare function when delta differs. Seems to improve a little.
2017-05-30 15:33:06 -05:00
cdebug_log(145,0) << "Using pitch for L/T shrink:" << DbU::getValueString(getPitch()) << endl;
Added analog type on segment NetRoutingProperty. * New: In Anabatic_AutoSegments collection, added a Flag to the constructors to allow different behavior between digital and analog modes. For "Aligneds" and "Perpandiculars" collections, now manage a new Flag WithDoglegs to follow aligned globals through local doglegs (for analog nets). Adjust the log level of collections to 144 (formerly 145). * New: In Anabatic::AutoSegment, new flag SegAnalog for segments that are part of an analog net. Note that with this flag, we reach the 32 bits limit... * Change: In Anabatic::Constants, Flags are now declared as BaseFlags objects and *not* uint64_t. This avoids overload resolution problems with arithmetical overload of the operators. The BaseFlags/Flags types are now completly "isolated" from the uint64_t, it has the advantage of showing where unwanted previous implicit conversions where occuring. * Change: In Katana::Constants, Flags values are now of BaseFlags type instead of uint64_t. * Change: In Anabatic::Dijkstra, lots of log cleanup. * Change: In Anabatic::GCell::getSide(), make the "shrink" parameter visible to allow to substract the topmost and rightmost track for axis span computation in AutoSegment::computeOptimal(). Used for analog mode. * Change: In NetRoutingState, added a flag for analog mode. Use uint32_t for the flags type. * New: In Isobar, export the NetRoutingState and NetRoutingExtension objects.
2017-05-20 05:33:12 -05:00
for ( AutoSegment* aligned : aligneds ) {
Bad computation of optimal position for analog segments. * Bug: In Anabatic::AutoSegment::computeOptimal(), the "side stack" computing the intersection of all the sides of the GCell that the segment is going through was implicitly supposing that the resulting intersection wasn't empty. But for analog segment, we compute the position of all the segments linked though simple doglegs, and as dogleg occurs there can be a sufficient shift so the intersections of all the sides became empty. Now the "side stack" can manage configuration when instead of having an intersection we have a "hole" in the GCell side span (only one hole is supported at the moment). When a hole is present, the optimal position is the center of the hole. Should be refined so that the position is computed according to the longest GCell that we go through. * Change: In Katana::SegmentFsm::_slackenGlobal(), when slackening a global from a ananlog net, allow to make a dog leg once instead of immediately moving up. The break point is the center of the segment. * Change: In Katana::Manipulator::Makedogleg(DbU::Unit) if the break position is over a device, choose the left of right GCell.
2018-02-21 10:03:43 -06:00
cdebug_log(145,0) << "@ " << aligned << endl;
Added analog type on segment NetRoutingProperty. * New: In Anabatic_AutoSegments collection, added a Flag to the constructors to allow different behavior between digital and analog modes. For "Aligneds" and "Perpandiculars" collections, now manage a new Flag WithDoglegs to follow aligned globals through local doglegs (for analog nets). Adjust the log level of collections to 144 (formerly 145). * New: In Anabatic::AutoSegment, new flag SegAnalog for segments that are part of an analog net. Note that with this flag, we reach the 32 bits limit... * Change: In Anabatic::Constants, Flags are now declared as BaseFlags objects and *not* uint64_t. This avoids overload resolution problems with arithmetical overload of the operators. The BaseFlags/Flags types are now completly "isolated" from the uint64_t, it has the advantage of showing where unwanted previous implicit conversions where occuring. * Change: In Katana::Constants, Flags values are now of BaseFlags type instead of uint64_t. * Change: In Anabatic::Dijkstra, lots of log cleanup. * Change: In Anabatic::GCell::getSide(), make the "shrink" parameter visible to allow to substract the topmost and rightmost track for axis span computation in AutoSegment::computeOptimal(). Used for analog mode. * Change: In NetRoutingState, added a flag for analog mode. Use uint32_t for the flags type. * New: In Isobar, export the NetRoutingState and NetRoutingExtension objects.
2017-05-20 05:33:12 -05:00
aligned->getGCells( gcells );
for ( GCell* gcell : gcells ) {
Synchronize priority of TrackSegments connecteds through doglegs. * Bug: In Anabatic::Dijsktra, correct the indentation in cdebug calls (lots of them causing a big shift right). * New: In Anabatic::TrackSegment, new helper structure SideStack to manage a set of aligned GCells and their various sides sizes. * Change: In Anabatic::TrackSegment::computeOptimal(), more accurate computation of attractors from global segments and variable size GCells using SideStack. * Change: In Katana::DataSymmetric::checkPairing(), increase the tolerance for misaligned symmetrics from 2 to 5 tracks (should be enough for our narrow channel routing). * New: In Katana::TrackSegment, add a first flag to enable locking of priority. If it is set, calls to either "computePriority()" or "forcePriority()" will have no effect. Added the uint32_t flags paraphernalia. * New: In ::computeNetPriority(), overall function to control the call of TrackSegment::computeAlignedPriority(). The call is done from NegociateWindow::run(). * New: Katana::TrackSegment::computeAlignedPriority(), order the TrackSegments aligneds through doglegs to the one with the highest priority is routed first and others progressively from him. Done by forcing an ever decreasing priority on the aligneds ones. The amount of decrease is small so the aligned segments got routed (ordered) in close, if not contiguous, sequence. Priority is locked for the order to remain. * Bug: In Katana::TrackSegment::computePriority(), correct computation of the priority when there is more than 10 free tracks (a DbU::toLambda() call was missing, leading to very big priorities). * Change: In katana::TrackCost CTOR, do not compute a distance to fixed in the case of analog segments, this is backfiring. Slight change of the compare function when delta differs. Seems to improve a little.
2017-05-30 15:33:06 -05:00
sideStack.addGCell( gcell );
cdebug_log(145,0) << "| gcellSide:" << sideStack.getGSide() << " (from " << gcell << ")" << endl;
Added analog type on segment NetRoutingProperty. * New: In Anabatic_AutoSegments collection, added a Flag to the constructors to allow different behavior between digital and analog modes. For "Aligneds" and "Perpandiculars" collections, now manage a new Flag WithDoglegs to follow aligned globals through local doglegs (for analog nets). Adjust the log level of collections to 144 (formerly 145). * New: In Anabatic::AutoSegment, new flag SegAnalog for segments that are part of an analog net. Note that with this flag, we reach the 32 bits limit... * Change: In Anabatic::Constants, Flags are now declared as BaseFlags objects and *not* uint64_t. This avoids overload resolution problems with arithmetical overload of the operators. The BaseFlags/Flags types are now completly "isolated" from the uint64_t, it has the advantage of showing where unwanted previous implicit conversions where occuring. * Change: In Katana::Constants, Flags values are now of BaseFlags type instead of uint64_t. * Change: In Anabatic::Dijkstra, lots of log cleanup. * Change: In Anabatic::GCell::getSide(), make the "shrink" parameter visible to allow to substract the topmost and rightmost track for axis span computation in AutoSegment::computeOptimal(). Used for analog mode. * Change: In NetRoutingState, added a flag for analog mode. Use uint32_t for the flags type. * New: In Isobar, export the NetRoutingState and NetRoutingExtension objects.
2017-05-20 05:33:12 -05:00
}
Bad computation of optimal position for analog segments. * Bug: In Anabatic::AutoSegment::computeOptimal(), the "side stack" computing the intersection of all the sides of the GCell that the segment is going through was implicitly supposing that the resulting intersection wasn't empty. But for analog segment, we compute the position of all the segments linked though simple doglegs, and as dogleg occurs there can be a sufficient shift so the intersections of all the sides became empty. Now the "side stack" can manage configuration when instead of having an intersection we have a "hole" in the GCell side span (only one hole is supported at the moment). When a hole is present, the optimal position is the center of the hole. Should be refined so that the position is computed according to the longest GCell that we go through. * Change: In Katana::SegmentFsm::_slackenGlobal(), when slackening a global from a ananlog net, allow to make a dog leg once instead of immediately moving up. The break point is the center of the segment. * Change: In Katana::Manipulator::Makedogleg(DbU::Unit) if the break position is over a device, choose the left of right GCell.
2018-02-21 10:03:43 -06:00
if (aligned->isStrongTerminal() and not sideStack.isHoled()) {
cdebug_log(145,0) << "> Is strong terminal, restrict." << aligned << endl;
Katana routing algorithm adjustements for analog routing. * Bug: In Hurricane::Interval, the ::getSize() method was returning a negative length when the Interval was empty. Now return zero. This was causing slight cost functions side effect when computing the overlap of a segment with another one belonging to the same net (i.e. shareDelta) * Bug: In Katana::AutoSegment::computeOptimal(), checks that the optimal position for the axis is indeed inside the GCell interval. * Bug: In Katana::DataSymmetric::getSymmetrical(Interval&), reverse the shrink applied to GCells sides. Interval are of the type [min:max[ so after symmetric transformation they become ]smax:smin]. Apply a correction so that they are [smax:smin[ and could be merged with the paired interval (the one without symmetry applied). * Bug: In katana::TrackCost::consolidate(), the overlap (now) always positive, and must be *substracted* to the "delta" (not *added*). * Bug: In Katana::_computeNetOptimals(), now skip the already processeds AutoSegments instead of re-processing them. * New: In Katana::AutoSegment::computeOptimal(), improved computation of the axis hint: - For analog net, propagate through the simple doglegs to find attractors. - Restrict the allowed interval to the union of GCell sides and segments constraints that are terminals. - Consider a local as "long" if it's length exceed 20 the P-pitch. (maybe make that a tool parameter).
2017-05-22 17:20:31 -05:00
Interval terminalConstraints;
aligned->getConstraints( terminalConstraints );
Synchronize priority of TrackSegments connecteds through doglegs. * Bug: In Anabatic::Dijsktra, correct the indentation in cdebug calls (lots of them causing a big shift right). * New: In Anabatic::TrackSegment, new helper structure SideStack to manage a set of aligned GCells and their various sides sizes. * Change: In Anabatic::TrackSegment::computeOptimal(), more accurate computation of attractors from global segments and variable size GCells using SideStack. * Change: In Katana::DataSymmetric::checkPairing(), increase the tolerance for misaligned symmetrics from 2 to 5 tracks (should be enough for our narrow channel routing). * New: In Katana::TrackSegment, add a first flag to enable locking of priority. If it is set, calls to either "computePriority()" or "forcePriority()" will have no effect. Added the uint32_t flags paraphernalia. * New: In ::computeNetPriority(), overall function to control the call of TrackSegment::computeAlignedPriority(). The call is done from NegociateWindow::run(). * New: Katana::TrackSegment::computeAlignedPriority(), order the TrackSegments aligneds through doglegs to the one with the highest priority is routed first and others progressively from him. Done by forcing an ever decreasing priority on the aligneds ones. The amount of decrease is small so the aligned segments got routed (ordered) in close, if not contiguous, sequence. Priority is locked for the order to remain. * Bug: In Katana::TrackSegment::computePriority(), correct computation of the priority when there is more than 10 free tracks (a DbU::toLambda() call was missing, leading to very big priorities). * Change: In katana::TrackCost CTOR, do not compute a distance to fixed in the case of analog segments, this is backfiring. Slight change of the compare function when delta differs. Seems to improve a little.
2017-05-30 15:33:06 -05:00
sideStack.restrictGSide( terminalConstraints );
Bad computation of optimal position for analog segments. * Bug: In Anabatic::AutoSegment::computeOptimal(), the "side stack" computing the intersection of all the sides of the GCell that the segment is going through was implicitly supposing that the resulting intersection wasn't empty. But for analog segment, we compute the position of all the segments linked though simple doglegs, and as dogleg occurs there can be a sufficient shift so the intersections of all the sides became empty. Now the "side stack" can manage configuration when instead of having an intersection we have a "hole" in the GCell side span (only one hole is supported at the moment). When a hole is present, the optimal position is the center of the hole. Should be refined so that the position is computed according to the longest GCell that we go through. * Change: In Katana::SegmentFsm::_slackenGlobal(), when slackening a global from a ananlog net, allow to make a dog leg once instead of immediately moving up. The break point is the center of the segment. * Change: In Katana::Manipulator::Makedogleg(DbU::Unit) if the break position is over a device, choose the left of right GCell.
2018-02-21 10:03:43 -06:00
cdebug_log(145,0) << "| " << terminalConstraints.intersection(sideStack.getGSide()) << endl;
Synchronize priority of TrackSegments connecteds through doglegs. * Bug: In Anabatic::Dijsktra, correct the indentation in cdebug calls (lots of them causing a big shift right). * New: In Anabatic::TrackSegment, new helper structure SideStack to manage a set of aligned GCells and their various sides sizes. * Change: In Anabatic::TrackSegment::computeOptimal(), more accurate computation of attractors from global segments and variable size GCells using SideStack. * Change: In Katana::DataSymmetric::checkPairing(), increase the tolerance for misaligned symmetrics from 2 to 5 tracks (should be enough for our narrow channel routing). * New: In Katana::TrackSegment, add a first flag to enable locking of priority. If it is set, calls to either "computePriority()" or "forcePriority()" will have no effect. Added the uint32_t flags paraphernalia. * New: In ::computeNetPriority(), overall function to control the call of TrackSegment::computeAlignedPriority(). The call is done from NegociateWindow::run(). * New: Katana::TrackSegment::computeAlignedPriority(), order the TrackSegments aligneds through doglegs to the one with the highest priority is routed first and others progressively from him. Done by forcing an ever decreasing priority on the aligneds ones. The amount of decrease is small so the aligned segments got routed (ordered) in close, if not contiguous, sequence. Priority is locked for the order to remain. * Bug: In Katana::TrackSegment::computePriority(), correct computation of the priority when there is more than 10 free tracks (a DbU::toLambda() call was missing, leading to very big priorities). * Change: In katana::TrackCost CTOR, do not compute a distance to fixed in the case of analog segments, this is backfiring. Slight change of the compare function when delta differs. Seems to improve a little.
2017-05-30 15:33:06 -05:00
cdebug_log(145,0) << "| gcellSide:" << sideStack.getGSide() << " (from " << aligned << ")" << endl;
Katana routing algorithm adjustements for analog routing. * Bug: In Hurricane::Interval, the ::getSize() method was returning a negative length when the Interval was empty. Now return zero. This was causing slight cost functions side effect when computing the overlap of a segment with another one belonging to the same net (i.e. shareDelta) * Bug: In Katana::AutoSegment::computeOptimal(), checks that the optimal position for the axis is indeed inside the GCell interval. * Bug: In Katana::DataSymmetric::getSymmetrical(Interval&), reverse the shrink applied to GCells sides. Interval are of the type [min:max[ so after symmetric transformation they become ]smax:smin]. Apply a correction so that they are [smax:smin[ and could be merged with the paired interval (the one without symmetry applied). * Bug: In katana::TrackCost::consolidate(), the overlap (now) always positive, and must be *substracted* to the "delta" (not *added*). * Bug: In Katana::_computeNetOptimals(), now skip the already processeds AutoSegments instead of re-processing them. * New: In Katana::AutoSegment::computeOptimal(), improved computation of the axis hint: - For analog net, propagate through the simple doglegs to find attractors. - Restrict the allowed interval to the union of GCell sides and segments constraints that are terminals. - Consider a local as "long" if it's length exceed 20 the P-pitch. (maybe make that a tool parameter).
2017-05-22 17:20:31 -05:00
}
Added analog type on segment NetRoutingProperty. * New: In Anabatic_AutoSegments collection, added a Flag to the constructors to allow different behavior between digital and analog modes. For "Aligneds" and "Perpandiculars" collections, now manage a new Flag WithDoglegs to follow aligned globals through local doglegs (for analog nets). Adjust the log level of collections to 144 (formerly 145). * New: In Anabatic::AutoSegment, new flag SegAnalog for segments that are part of an analog net. Note that with this flag, we reach the 32 bits limit... * Change: In Anabatic::Constants, Flags are now declared as BaseFlags objects and *not* uint64_t. This avoids overload resolution problems with arithmetical overload of the operators. The BaseFlags/Flags types are now completly "isolated" from the uint64_t, it has the advantage of showing where unwanted previous implicit conversions where occuring. * Change: In Katana::Constants, Flags values are now of BaseFlags type instead of uint64_t. * Change: In Anabatic::Dijkstra, lots of log cleanup. * Change: In Anabatic::GCell::getSide(), make the "shrink" parameter visible to allow to substract the topmost and rightmost track for axis span computation in AutoSegment::computeOptimal(). Used for analog mode. * Change: In NetRoutingState, added a flag for analog mode. Use uint32_t for the flags type. * New: In Isobar, export the NetRoutingState and NetRoutingExtension objects.
2017-05-20 05:33:12 -05:00
}
}
Synchronize priority of TrackSegments connecteds through doglegs. * Bug: In Anabatic::Dijsktra, correct the indentation in cdebug calls (lots of them causing a big shift right). * New: In Anabatic::TrackSegment, new helper structure SideStack to manage a set of aligned GCells and their various sides sizes. * Change: In Anabatic::TrackSegment::computeOptimal(), more accurate computation of attractors from global segments and variable size GCells using SideStack. * Change: In Katana::DataSymmetric::checkPairing(), increase the tolerance for misaligned symmetrics from 2 to 5 tracks (should be enough for our narrow channel routing). * New: In Katana::TrackSegment, add a first flag to enable locking of priority. If it is set, calls to either "computePriority()" or "forcePriority()" will have no effect. Added the uint32_t flags paraphernalia. * New: In ::computeNetPriority(), overall function to control the call of TrackSegment::computeAlignedPriority(). The call is done from NegociateWindow::run(). * New: Katana::TrackSegment::computeAlignedPriority(), order the TrackSegments aligneds through doglegs to the one with the highest priority is routed first and others progressively from him. Done by forcing an ever decreasing priority on the aligneds ones. The amount of decrease is small so the aligned segments got routed (ordered) in close, if not contiguous, sequence. Priority is locked for the order to remain. * Bug: In Katana::TrackSegment::computePriority(), correct computation of the priority when there is more than 10 free tracks (a DbU::toLambda() call was missing, leading to very big priorities). * Change: In katana::TrackCost CTOR, do not compute a distance to fixed in the case of analog segments, this is backfiring. Slight change of the compare function when delta differs. Seems to improve a little.
2017-05-30 15:33:06 -05:00
sideStack.show();
Added analog type on segment NetRoutingProperty. * New: In Anabatic_AutoSegments collection, added a Flag to the constructors to allow different behavior between digital and analog modes. For "Aligneds" and "Perpandiculars" collections, now manage a new Flag WithDoglegs to follow aligned globals through local doglegs (for analog nets). Adjust the log level of collections to 144 (formerly 145). * New: In Anabatic::AutoSegment, new flag SegAnalog for segments that are part of an analog net. Note that with this flag, we reach the 32 bits limit... * Change: In Anabatic::Constants, Flags are now declared as BaseFlags objects and *not* uint64_t. This avoids overload resolution problems with arithmetical overload of the operators. The BaseFlags/Flags types are now completly "isolated" from the uint64_t, it has the advantage of showing where unwanted previous implicit conversions where occuring. * Change: In Katana::Constants, Flags values are now of BaseFlags type instead of uint64_t. * Change: In Anabatic::Dijkstra, lots of log cleanup. * Change: In Anabatic::GCell::getSide(), make the "shrink" parameter visible to allow to substract the topmost and rightmost track for axis span computation in AutoSegment::computeOptimal(). Used for analog mode. * Change: In NetRoutingState, added a flag for analog mode. Use uint32_t for the flags type. * New: In Isobar, export the NetRoutingState and NetRoutingExtension objects.
2017-05-20 05:33:12 -05:00
Synchronize priority of TrackSegments connecteds through doglegs. * Bug: In Anabatic::Dijsktra, correct the indentation in cdebug calls (lots of them causing a big shift right). * New: In Anabatic::TrackSegment, new helper structure SideStack to manage a set of aligned GCells and their various sides sizes. * Change: In Anabatic::TrackSegment::computeOptimal(), more accurate computation of attractors from global segments and variable size GCells using SideStack. * Change: In Katana::DataSymmetric::checkPairing(), increase the tolerance for misaligned symmetrics from 2 to 5 tracks (should be enough for our narrow channel routing). * New: In Katana::TrackSegment, add a first flag to enable locking of priority. If it is set, calls to either "computePriority()" or "forcePriority()" will have no effect. Added the uint32_t flags paraphernalia. * New: In ::computeNetPriority(), overall function to control the call of TrackSegment::computeAlignedPriority(). The call is done from NegociateWindow::run(). * New: Katana::TrackSegment::computeAlignedPriority(), order the TrackSegments aligneds through doglegs to the one with the highest priority is routed first and others progressively from him. Done by forcing an ever decreasing priority on the aligneds ones. The amount of decrease is small so the aligned segments got routed (ordered) in close, if not contiguous, sequence. Priority is locked for the order to remain. * Bug: In Katana::TrackSegment::computePriority(), correct computation of the priority when there is more than 10 free tracks (a DbU::toLambda() call was missing, leading to very big priorities). * Change: In katana::TrackCost CTOR, do not compute a distance to fixed in the case of analog segments, this is backfiring. Slight change of the compare function when delta differs. Seems to improve a little.
2017-05-30 15:33:06 -05:00
cdebug_log(145,0) << "GCell interval " << sideStack.getGSide() << endl;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
AutoContact* anchor = getAutoSource();
if (anchor->isTerminal()) {
Box constraintBox = anchor->getConstraintBox();
if ( isHorizontal() ) {
terminalMin = constraintBox.getYMin();
terminalMax = constraintBox.getYMax();
} else {
terminalMin = constraintBox.getXMin();
terminalMax = constraintBox.getXMax();
}
attractors.addAttractor( terminalMin );
if (terminalMin != terminalMax)
attractors.addAttractor( terminalMax );
}
anchor = getAutoTarget();
if (anchor->isTerminal()) {
Box constraintBox = anchor->getConstraintBox();
if (isHorizontal()) {
terminalMin = constraintBox.getYMin();
terminalMax = constraintBox.getYMax();
} else {
terminalMin = constraintBox.getXMin();
terminalMax = constraintBox.getXMax();
}
attractors.addAttractor( terminalMin );
if (terminalMin != terminalMax)
attractors.addAttractor( terminalMax );
}
Added analog type on segment NetRoutingProperty. * New: In Anabatic_AutoSegments collection, added a Flag to the constructors to allow different behavior between digital and analog modes. For "Aligneds" and "Perpandiculars" collections, now manage a new Flag WithDoglegs to follow aligned globals through local doglegs (for analog nets). Adjust the log level of collections to 144 (formerly 145). * New: In Anabatic::AutoSegment, new flag SegAnalog for segments that are part of an analog net. Note that with this flag, we reach the 32 bits limit... * Change: In Anabatic::Constants, Flags are now declared as BaseFlags objects and *not* uint64_t. This avoids overload resolution problems with arithmetical overload of the operators. The BaseFlags/Flags types are now completly "isolated" from the uint64_t, it has the advantage of showing where unwanted previous implicit conversions where occuring. * Change: In Katana::Constants, Flags values are now of BaseFlags type instead of uint64_t. * Change: In Anabatic::Dijkstra, lots of log cleanup. * Change: In Anabatic::GCell::getSide(), make the "shrink" parameter visible to allow to substract the topmost and rightmost track for axis span computation in AutoSegment::computeOptimal(). Used for analog mode. * Change: In NetRoutingState, added a flag for analog mode. Use uint32_t for the flags type. * New: In Isobar, export the NetRoutingState and NetRoutingExtension objects.
2017-05-20 05:33:12 -05:00
for ( AutoSegment* autoSegment : getPerpandiculars(flags) ) {
cdebug_log(145,1) << "| Perpandicular " << autoSegment << endl;
First intergration of the Analogic router parts. * New: In Anabatic::AutoSegment, introduce a the kind (associated to a flag) "LongLocal". Analog GCells can be very wide, so at least some carefuly choosen long local segments must be took into account as attractors in the computation of the optimal axis. * New: In Anabatic::AutoSegment::computeOptimal(), take LongLocal into account as attractors. * Change: In ::GCellTopology constructors compare the layers of the RoutingPads using layer masks instead of Layer pointers. Allows to find both "METALx" (symbolic) and "metalX" (real). * Change: In ::GCellTopology::_doHChannel(), _doChannel(), _doStrut() and _doDevice(), tag long locals as "LongLocal". This need to be reviewed as it as bind done a bit too quickly. * Change: In Anabatic::AutoSegment, due too a much bigger span of the analogic GCells the _optimalMin & _optimalMax bitfields must use 16 bits instead of 8 (they where overflowed). * New: In Katana, reorganisation of the initialization procedure to fit both digital and analogic cases. Create an analogInit() method. * Change: In Katana::RoutingEvent, the _tracksNb and _tracksFree bitfields where too short for the Analog GCell size, now uses 16 bits instead of 6. * Bug: In Katana::GraphicKatanEngine::drawGCell(), skip drawing of a GCell if *both* width and height are under 150 pixels. * New: In Katana::Session, add a new isOpen() method.
2016-10-04 10:12:58 -05:00
if (autoSegment->isGlobal()) {
Added analog type on segment NetRoutingProperty. * New: In Anabatic_AutoSegments collection, added a Flag to the constructors to allow different behavior between digital and analog modes. For "Aligneds" and "Perpandiculars" collections, now manage a new Flag WithDoglegs to follow aligned globals through local doglegs (for analog nets). Adjust the log level of collections to 144 (formerly 145). * New: In Anabatic::AutoSegment, new flag SegAnalog for segments that are part of an analog net. Note that with this flag, we reach the 32 bits limit... * Change: In Anabatic::Constants, Flags are now declared as BaseFlags objects and *not* uint64_t. This avoids overload resolution problems with arithmetical overload of the operators. The BaseFlags/Flags types are now completly "isolated" from the uint64_t, it has the advantage of showing where unwanted previous implicit conversions where occuring. * Change: In Katana::Constants, Flags values are now of BaseFlags type instead of uint64_t. * Change: In Anabatic::Dijkstra, lots of log cleanup. * Change: In Anabatic::GCell::getSide(), make the "shrink" parameter visible to allow to substract the topmost and rightmost track for axis span computation in AutoSegment::computeOptimal(). Used for analog mode. * Change: In NetRoutingState, added a flag for analog mode. Use uint32_t for the flags type. * New: In Isobar, export the NetRoutingState and NetRoutingExtension objects.
2017-05-20 05:33:12 -05:00
cdebug_log(145,0) << "Used as global." << endl;
Synchronize priority of TrackSegments connecteds through doglegs. * Bug: In Anabatic::Dijsktra, correct the indentation in cdebug calls (lots of them causing a big shift right). * New: In Anabatic::TrackSegment, new helper structure SideStack to manage a set of aligned GCells and their various sides sizes. * Change: In Anabatic::TrackSegment::computeOptimal(), more accurate computation of attractors from global segments and variable size GCells using SideStack. * Change: In Katana::DataSymmetric::checkPairing(), increase the tolerance for misaligned symmetrics from 2 to 5 tracks (should be enough for our narrow channel routing). * New: In Katana::TrackSegment, add a first flag to enable locking of priority. If it is set, calls to either "computePriority()" or "forcePriority()" will have no effect. Added the uint32_t flags paraphernalia. * New: In ::computeNetPriority(), overall function to control the call of TrackSegment::computeAlignedPriority(). The call is done from NegociateWindow::run(). * New: Katana::TrackSegment::computeAlignedPriority(), order the TrackSegments aligneds through doglegs to the one with the highest priority is routed first and others progressively from him. Done by forcing an ever decreasing priority on the aligneds ones. The amount of decrease is small so the aligned segments got routed (ordered) in close, if not contiguous, sequence. Priority is locked for the order to remain. * Bug: In Katana::TrackSegment::computePriority(), correct computation of the priority when there is more than 10 free tracks (a DbU::toLambda() call was missing, leading to very big priorities). * Change: In katana::TrackCost CTOR, do not compute a distance to fixed in the case of analog segments, this is backfiring. Slight change of the compare function when delta differs. Seems to improve a little.
2017-05-30 15:33:06 -05:00
const Interval& side = sideStack.getSideAt( autoSegment->getAxis() );
cdebug_log(145,0) << "Side @" << DbU::getValueString(autoSegment->getAxis())
<< " " << side << endl;
if (autoSegment->getSourceU() < side.getVMin()) attractors.addAttractor( sideStack.getGSideMin() );
if (autoSegment->getTargetU() > side.getVMax()) attractors.addAttractor( sideStack.getGSideMax() );
// // Sloppy implentation.
// DbU::Unit perpandMin = autoSegment->getSourceU();
// DbU::Unit perpandMax = autoSegment->getTargetU();
First intergration of the Analogic router parts. * New: In Anabatic::AutoSegment, introduce a the kind (associated to a flag) "LongLocal". Analog GCells can be very wide, so at least some carefuly choosen long local segments must be took into account as attractors in the computation of the optimal axis. * New: In Anabatic::AutoSegment::computeOptimal(), take LongLocal into account as attractors. * Change: In ::GCellTopology constructors compare the layers of the RoutingPads using layer masks instead of Layer pointers. Allows to find both "METALx" (symbolic) and "metalX" (real). * Change: In ::GCellTopology::_doHChannel(), _doChannel(), _doStrut() and _doDevice(), tag long locals as "LongLocal". This need to be reviewed as it as bind done a bit too quickly. * Change: In Anabatic::AutoSegment, due too a much bigger span of the analogic GCells the _optimalMin & _optimalMax bitfields must use 16 bits instead of 8 (they where overflowed). * New: In Katana, reorganisation of the initialization procedure to fit both digital and analogic cases. Create an analogInit() method. * Change: In Katana::RoutingEvent, the _tracksNb and _tracksFree bitfields where too short for the Analog GCell size, now uses 16 bits instead of 6. * Bug: In Katana::GraphicKatanEngine::drawGCell(), skip drawing of a GCell if *both* width and height are under 150 pixels. * New: In Katana::Session, add a new isOpen() method.
2016-10-04 10:12:58 -05:00
Synchronize priority of TrackSegments connecteds through doglegs. * Bug: In Anabatic::Dijsktra, correct the indentation in cdebug calls (lots of them causing a big shift right). * New: In Anabatic::TrackSegment, new helper structure SideStack to manage a set of aligned GCells and their various sides sizes. * Change: In Anabatic::TrackSegment::computeOptimal(), more accurate computation of attractors from global segments and variable size GCells using SideStack. * Change: In Katana::DataSymmetric::checkPairing(), increase the tolerance for misaligned symmetrics from 2 to 5 tracks (should be enough for our narrow channel routing). * New: In Katana::TrackSegment, add a first flag to enable locking of priority. If it is set, calls to either "computePriority()" or "forcePriority()" will have no effect. Added the uint32_t flags paraphernalia. * New: In ::computeNetPriority(), overall function to control the call of TrackSegment::computeAlignedPriority(). The call is done from NegociateWindow::run(). * New: Katana::TrackSegment::computeAlignedPriority(), order the TrackSegments aligneds through doglegs to the one with the highest priority is routed first and others progressively from him. Done by forcing an ever decreasing priority on the aligneds ones. The amount of decrease is small so the aligned segments got routed (ordered) in close, if not contiguous, sequence. Priority is locked for the order to remain. * Bug: In Katana::TrackSegment::computePriority(), correct computation of the priority when there is more than 10 free tracks (a DbU::toLambda() call was missing, leading to very big priorities). * Change: In katana::TrackCost CTOR, do not compute a distance to fixed in the case of analog segments, this is backfiring. Slight change of the compare function when delta differs. Seems to improve a little.
2017-05-30 15:33:06 -05:00
// if (perpandMin < minGCell) attractors.addAttractor( minGCell );
// if (perpandMax > maxGCell) attractors.addAttractor( maxGCell );
First intergration of the Analogic router parts. * New: In Anabatic::AutoSegment, introduce a the kind (associated to a flag) "LongLocal". Analog GCells can be very wide, so at least some carefuly choosen long local segments must be took into account as attractors in the computation of the optimal axis. * New: In Anabatic::AutoSegment::computeOptimal(), take LongLocal into account as attractors. * Change: In ::GCellTopology constructors compare the layers of the RoutingPads using layer masks instead of Layer pointers. Allows to find both "METALx" (symbolic) and "metalX" (real). * Change: In ::GCellTopology::_doHChannel(), _doChannel(), _doStrut() and _doDevice(), tag long locals as "LongLocal". This need to be reviewed as it as bind done a bit too quickly. * Change: In Anabatic::AutoSegment, due too a much bigger span of the analogic GCells the _optimalMin & _optimalMax bitfields must use 16 bits instead of 8 (they where overflowed). * New: In Katana, reorganisation of the initialization procedure to fit both digital and analogic cases. Create an analogInit() method. * Change: In Katana::RoutingEvent, the _tracksNb and _tracksFree bitfields where too short for the Analog GCell size, now uses 16 bits instead of 6. * Bug: In Katana::GraphicKatanEngine::drawGCell(), skip drawing of a GCell if *both* width and height are under 150 pixels. * New: In Katana::Session, add a new isOpen() method.
2016-10-04 10:12:58 -05:00
} else if (autoSegment->isLocal()) {
Support for symmetric routing in Anabatic/Katabatic. * New: In Katana::SegmentFsm, the object is now able to handle two events at the same time. The master and it's symmetric. When there is no symmetric, the corresponing data is just left blank. This makes a bigger object, but as there is only one when running, it is not an issue. Candidates tracks are now an vector of array<2> (pairs), the TrackCost::Compare() functor has to be wrapped through CompareCostArray. The compined TrackCost of the two tracks is accumulated into the first element. Everything related to events gets duplicated: _event is now _event1 and _event2, and so on. As there can be now two Manipulator actions done with SegmentFsm, this class now completly hide the Manipulator level from the RoutingEvent processing. New function ::bindToTrack() to perform the track insertion. * New: In Katana::TrackCost, add a new ::merge() function. * New: In Katana::TrackElement and Katana::TrackSegment, add symmetric management. Allows to know if a TrackElement has a symmetric and to access it. * New: In Katana::DataSymmetric, add new overload for ::getSymmetrical() to handle DbU::Unit and intervals. * Change: In Katana::RoutingEvent, remove all direct uses of Manipulator objects. Now any change to the event associated segment must go through call to Segment Fsm. * Change: In Katana, adjust the debug level so internal informations are put below level 156. * New: In Hurricane::DbU, in ::getValueString(), special display when the value is Min or Max (more helpful than a gigantic number).
2017-04-30 16:46:33 -05:00
if (autoSegment->isStrongTerminal()) {
Replace "unsigned int" by "Flags" in all AutoSegments collections. * Change: In Anabatic::AutoSegments collections, change the type of all the flags that where in "unsigned int" (32 bits) to Flags (uint64_t) as there is now more than 32 flags for functions. * New: In Ababatic::Constants, added new flag Flags::WithPerpands, which makes the number of flags tip over 32 bits, thus making mandatory to uses Flags and not unsigned int. * New: In Anabatic::AutoSegments_Perpandiculars, manage a new flag Flags::WithDoglegs to allow to propagate through global segments that are connecteds via doglegs on local segments. Meaning that there is a good chance that they could be aligned. Slighly change the way we propagate on aligned segments: no longer check for VTee or HTee, but only for same direction and layer as master. * New: In Anabatic & Katana, replace all the "int", "long" and their variants by the less implementation ambiguous "int32_t", "int64_t" (and variant). This should help to better detect bit trucation in flags. Use the type to give a hint about the flags kind: - Type "Flags", for flags shared among Anabatic & Katana functions/methods (may also appear in some objects states). - Type "uint32_t" for flags belonging to an object internal state of from Hurricane functions flags (those should be grouped in a Flag subclass in a perfect world).
2017-05-16 07:53:33 -05:00
cdebug_log(145,0) << "Used as strong terminal." << endl;
Support for symmetric routing in Anabatic/Katabatic. * New: In Katana::SegmentFsm, the object is now able to handle two events at the same time. The master and it's symmetric. When there is no symmetric, the corresponing data is just left blank. This makes a bigger object, but as there is only one when running, it is not an issue. Candidates tracks are now an vector of array<2> (pairs), the TrackCost::Compare() functor has to be wrapped through CompareCostArray. The compined TrackCost of the two tracks is accumulated into the first element. Everything related to events gets duplicated: _event is now _event1 and _event2, and so on. As there can be now two Manipulator actions done with SegmentFsm, this class now completly hide the Manipulator level from the RoutingEvent processing. New function ::bindToTrack() to perform the track insertion. * New: In Katana::TrackCost, add a new ::merge() function. * New: In Katana::TrackElement and Katana::TrackSegment, add symmetric management. Allows to know if a TrackElement has a symmetric and to access it. * New: In Katana::DataSymmetric, add new overload for ::getSymmetrical() to handle DbU::Unit and intervals. * Change: In Katana::RoutingEvent, remove all direct uses of Manipulator objects. Now any change to the event associated segment must go through call to Segment Fsm. * Change: In Katana, adjust the debug level so internal informations are put below level 156. * New: In Hurricane::DbU, in ::getValueString(), special display when the value is Min or Max (more helpful than a gigantic number).
2017-04-30 16:46:33 -05:00
DbU::Unit terminalMin;
DbU::Unit terminalMax;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
Added analog type on segment NetRoutingProperty. * New: In Anabatic_AutoSegments collection, added a Flag to the constructors to allow different behavior between digital and analog modes. For "Aligneds" and "Perpandiculars" collections, now manage a new Flag WithDoglegs to follow aligned globals through local doglegs (for analog nets). Adjust the log level of collections to 144 (formerly 145). * New: In Anabatic::AutoSegment, new flag SegAnalog for segments that are part of an analog net. Note that with this flag, we reach the 32 bits limit... * Change: In Anabatic::Constants, Flags are now declared as BaseFlags objects and *not* uint64_t. This avoids overload resolution problems with arithmetical overload of the operators. The BaseFlags/Flags types are now completly "isolated" from the uint64_t, it has the advantage of showing where unwanted previous implicit conversions where occuring. * Change: In Katana::Constants, Flags values are now of BaseFlags type instead of uint64_t. * Change: In Anabatic::Dijkstra, lots of log cleanup. * Change: In Anabatic::GCell::getSide(), make the "shrink" parameter visible to allow to substract the topmost and rightmost track for axis span computation in AutoSegment::computeOptimal(). Used for analog mode. * Change: In NetRoutingState, added a flag for analog mode. Use uint32_t for the flags type. * New: In Isobar, export the NetRoutingState and NetRoutingExtension objects.
2017-05-20 05:33:12 -05:00
if (getTerminalInterval( autoSegment
Support for symmetric routing in Anabatic/Katabatic. * New: In Katana::SegmentFsm, the object is now able to handle two events at the same time. The master and it's symmetric. When there is no symmetric, the corresponing data is just left blank. This makes a bigger object, but as there is only one when running, it is not an issue. Candidates tracks are now an vector of array<2> (pairs), the TrackCost::Compare() functor has to be wrapped through CompareCostArray. The compined TrackCost of the two tracks is accumulated into the first element. Everything related to events gets duplicated: _event is now _event1 and _event2, and so on. As there can be now two Manipulator actions done with SegmentFsm, this class now completly hide the Manipulator level from the RoutingEvent processing. New function ::bindToTrack() to perform the track insertion. * New: In Katana::TrackCost, add a new ::merge() function. * New: In Katana::TrackElement and Katana::TrackSegment, add symmetric management. Allows to know if a TrackElement has a symmetric and to access it. * New: In Katana::DataSymmetric, add new overload for ::getSymmetrical() to handle DbU::Unit and intervals. * Change: In Katana::RoutingEvent, remove all direct uses of Manipulator objects. Now any change to the event associated segment must go through call to Segment Fsm. * Change: In Katana, adjust the debug level so internal informations are put below level 156. * New: In Hurricane::DbU, in ::getValueString(), special display when the value is Min or Max (more helpful than a gigantic number).
2017-04-30 16:46:33 -05:00
, NULL
, isHorizontal()
, terminalMin
, terminalMax )) {
attractors.addAttractor( terminalMin );
if (terminalMin != terminalMax)
attractors.addAttractor( terminalMax );
}
Katana routing algorithm adjustements for analog routing. * Bug: In Hurricane::Interval, the ::getSize() method was returning a negative length when the Interval was empty. Now return zero. This was causing slight cost functions side effect when computing the overlap of a segment with another one belonging to the same net (i.e. shareDelta) * Bug: In Katana::AutoSegment::computeOptimal(), checks that the optimal position for the axis is indeed inside the GCell interval. * Bug: In Katana::DataSymmetric::getSymmetrical(Interval&), reverse the shrink applied to GCells sides. Interval are of the type [min:max[ so after symmetric transformation they become ]smax:smin]. Apply a correction so that they are [smax:smin[ and could be merged with the paired interval (the one without symmetry applied). * Bug: In katana::TrackCost::consolidate(), the overlap (now) always positive, and must be *substracted* to the "delta" (not *added*). * Bug: In Katana::_computeNetOptimals(), now skip the already processeds AutoSegments instead of re-processing them. * New: In Katana::AutoSegment::computeOptimal(), improved computation of the axis hint: - For analog net, propagate through the simple doglegs to find attractors. - Restrict the allowed interval to the union of GCell sides and segments constraints that are terminals. - Consider a local as "long" if it's length exceed 20 the P-pitch. (maybe make that a tool parameter).
2017-05-22 17:20:31 -05:00
} else if (autoSegment->isLongLocal() or (autoSegment->getLength() > getPPitch()*20)) {
Replace "unsigned int" by "Flags" in all AutoSegments collections. * Change: In Anabatic::AutoSegments collections, change the type of all the flags that where in "unsigned int" (32 bits) to Flags (uint64_t) as there is now more than 32 flags for functions. * New: In Ababatic::Constants, added new flag Flags::WithPerpands, which makes the number of flags tip over 32 bits, thus making mandatory to uses Flags and not unsigned int. * New: In Anabatic::AutoSegments_Perpandiculars, manage a new flag Flags::WithDoglegs to allow to propagate through global segments that are connecteds via doglegs on local segments. Meaning that there is a good chance that they could be aligned. Slighly change the way we propagate on aligned segments: no longer check for VTee or HTee, but only for same direction and layer as master. * New: In Anabatic & Katana, replace all the "int", "long" and their variants by the less implementation ambiguous "int32_t", "int64_t" (and variant). This should help to better detect bit trucation in flags. Use the type to give a hint about the flags kind: - Type "Flags", for flags shared among Anabatic & Katana functions/methods (may also appear in some objects states). - Type "uint32_t" for flags belonging to an object internal state of from Hurricane functions flags (those should be grouped in a Flag subclass in a perfect world).
2017-05-16 07:53:33 -05:00
cdebug_log(145,0) << "Used as long global attractor." << endl;
Support for symmetric routing in Anabatic/Katabatic. * New: In Katana::SegmentFsm, the object is now able to handle two events at the same time. The master and it's symmetric. When there is no symmetric, the corresponing data is just left blank. This makes a bigger object, but as there is only one when running, it is not an issue. Candidates tracks are now an vector of array<2> (pairs), the TrackCost::Compare() functor has to be wrapped through CompareCostArray. The compined TrackCost of the two tracks is accumulated into the first element. Everything related to events gets duplicated: _event is now _event1 and _event2, and so on. As there can be now two Manipulator actions done with SegmentFsm, this class now completly hide the Manipulator level from the RoutingEvent processing. New function ::bindToTrack() to perform the track insertion. * New: In Katana::TrackCost, add a new ::merge() function. * New: In Katana::TrackElement and Katana::TrackSegment, add symmetric management. Allows to know if a TrackElement has a symmetric and to access it. * New: In Katana::DataSymmetric, add new overload for ::getSymmetrical() to handle DbU::Unit and intervals. * Change: In Katana::RoutingEvent, remove all direct uses of Manipulator objects. Now any change to the event associated segment must go through call to Segment Fsm. * Change: In Katana, adjust the debug level so internal informations are put below level 156. * New: In Hurricane::DbU, in ::getValueString(), special display when the value is Min or Max (more helpful than a gigantic number).
2017-04-30 16:46:33 -05:00
DbU::Unit perpandMin = autoSegment->getSourceU();
DbU::Unit perpandMax = autoSegment->getTargetU();
if (perpandMin != perpandMax) {
if (perpandMin == getAxis()) attractors.addAttractor( perpandMax );
if (perpandMax == getAxis()) attractors.addAttractor( perpandMin );
}
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
}
}
cdebug_tabw(145,-1);
}
Bad computation of optimal position for analog segments. * Bug: In Anabatic::AutoSegment::computeOptimal(), the "side stack" computing the intersection of all the sides of the GCell that the segment is going through was implicitly supposing that the resulting intersection wasn't empty. But for analog segment, we compute the position of all the segments linked though simple doglegs, and as dogleg occurs there can be a sufficient shift so the intersections of all the sides became empty. Now the "side stack" can manage configuration when instead of having an intersection we have a "hole" in the GCell side span (only one hole is supported at the moment). When a hole is present, the optimal position is the center of the hole. Should be refined so that the position is computed according to the longest GCell that we go through. * Change: In Katana::SegmentFsm::_slackenGlobal(), when slackening a global from a ananlog net, allow to make a dog leg once instead of immediately moving up. The break point is the center of the segment. * Change: In Katana::Manipulator::Makedogleg(DbU::Unit) if the break position is over a device, choose the left of right GCell.
2018-02-21 10:03:43 -06:00
if (sideStack.isHoled()) {
optimalMin = optimalMax = sideStack.getGSideCenter();
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
} else {
Bad computation of optimal position for analog segments. * Bug: In Anabatic::AutoSegment::computeOptimal(), the "side stack" computing the intersection of all the sides of the GCell that the segment is going through was implicitly supposing that the resulting intersection wasn't empty. But for analog segment, we compute the position of all the segments linked though simple doglegs, and as dogleg occurs there can be a sufficient shift so the intersections of all the sides became empty. Now the "side stack" can manage configuration when instead of having an intersection we have a "hole" in the GCell side span (only one hole is supported at the moment). When a hole is present, the optimal position is the center of the hole. Should be refined so that the position is computed according to the longest GCell that we go through. * Change: In Katana::SegmentFsm::_slackenGlobal(), when slackening a global from a ananlog net, allow to make a dog leg once instead of immediately moving up. The break point is the center of the segment. * Change: In Katana::Manipulator::Makedogleg(DbU::Unit) if the break position is over a device, choose the left of right GCell.
2018-02-21 10:03:43 -06:00
if (attractors.getAttractorsCount()) {
optimalMin = attractors.getLowerMedian();
optimalMax = attractors.getUpperMedian();
} else {
cdebug_log(145,0) << "No attractors, reverting to GCell bounding box" << endl;
optimalMin = 0;
optimalMax = (isHorizontal()) ? _gcell->getBoundingBox().getYMax()
: _gcell->getBoundingBox().getXMax();
}
Replace "unsigned int" by "Flags" in all AutoSegments collections. * Change: In Anabatic::AutoSegments collections, change the type of all the flags that where in "unsigned int" (32 bits) to Flags (uint64_t) as there is now more than 32 flags for functions. * New: In Ababatic::Constants, added new flag Flags::WithPerpands, which makes the number of flags tip over 32 bits, thus making mandatory to uses Flags and not unsigned int. * New: In Anabatic::AutoSegments_Perpandiculars, manage a new flag Flags::WithDoglegs to allow to propagate through global segments that are connecteds via doglegs on local segments. Meaning that there is a good chance that they could be aligned. Slighly change the way we propagate on aligned segments: no longer check for VTee or HTee, but only for same direction and layer as master. * New: In Anabatic & Katana, replace all the "int", "long" and their variants by the less implementation ambiguous "int32_t", "int64_t" (and variant). This should help to better detect bit trucation in flags. Use the type to give a hint about the flags kind: - Type "Flags", for flags shared among Anabatic & Katana functions/methods (may also appear in some objects states). - Type "uint32_t" for flags belonging to an object internal state of from Hurricane functions flags (those should be grouped in a Flag subclass in a perfect world).
2017-05-16 07:53:33 -05:00
Bad computation of optimal position for analog segments. * Bug: In Anabatic::AutoSegment::computeOptimal(), the "side stack" computing the intersection of all the sides of the GCell that the segment is going through was implicitly supposing that the resulting intersection wasn't empty. But for analog segment, we compute the position of all the segments linked though simple doglegs, and as dogleg occurs there can be a sufficient shift so the intersections of all the sides became empty. Now the "side stack" can manage configuration when instead of having an intersection we have a "hole" in the GCell side span (only one hole is supported at the moment). When a hole is present, the optimal position is the center of the hole. Should be refined so that the position is computed according to the longest GCell that we go through. * Change: In Katana::SegmentFsm::_slackenGlobal(), when slackening a global from a ananlog net, allow to make a dog leg once instead of immediately moving up. The break point is the center of the segment. * Change: In Katana::Manipulator::Makedogleg(DbU::Unit) if the break position is over a device, choose the left of right GCell.
2018-02-21 10:03:43 -06:00
setInBound( sideStack.getGSideMin(), sideStack.getGSideMax(), optimalMin );
setInBound( sideStack.getGSideMin(), sideStack.getGSideMax(), optimalMax );
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
}
Katana routing algorithm adjustements for analog routing. * Bug: In Hurricane::Interval, the ::getSize() method was returning a negative length when the Interval was empty. Now return zero. This was causing slight cost functions side effect when computing the overlap of a segment with another one belonging to the same net (i.e. shareDelta) * Bug: In Katana::AutoSegment::computeOptimal(), checks that the optimal position for the axis is indeed inside the GCell interval. * Bug: In Katana::DataSymmetric::getSymmetrical(Interval&), reverse the shrink applied to GCells sides. Interval are of the type [min:max[ so after symmetric transformation they become ]smax:smin]. Apply a correction so that they are [smax:smin[ and could be merged with the paired interval (the one without symmetry applied). * Bug: In katana::TrackCost::consolidate(), the overlap (now) always positive, and must be *substracted* to the "delta" (not *added*). * Bug: In Katana::_computeNetOptimals(), now skip the already processeds AutoSegments instead of re-processing them. * New: In Katana::AutoSegment::computeOptimal(), improved computation of the axis hint: - For analog net, propagate through the simple doglegs to find attractors. - Restrict the allowed interval to the union of GCell sides and segments constraints that are terminals. - Consider a local as "long" if it's length exceed 20 the P-pitch. (maybe make that a tool parameter).
2017-05-22 17:20:31 -05:00
cdebug_log(145,0) << "optimalMin: " << DbU::getValueString(optimalMin) << endl;
cdebug_log(145,0) << "optimalMax: " << DbU::getValueString(optimalMax) << endl;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
}
setInBound( constraintMin, constraintMax, optimalMin );
setInBound( constraintMin, constraintMax, optimalMax );
Added analog type on segment NetRoutingProperty. * New: In Anabatic_AutoSegments collection, added a Flag to the constructors to allow different behavior between digital and analog modes. For "Aligneds" and "Perpandiculars" collections, now manage a new Flag WithDoglegs to follow aligned globals through local doglegs (for analog nets). Adjust the log level of collections to 144 (formerly 145). * New: In Anabatic::AutoSegment, new flag SegAnalog for segments that are part of an analog net. Note that with this flag, we reach the 32 bits limit... * Change: In Anabatic::Constants, Flags are now declared as BaseFlags objects and *not* uint64_t. This avoids overload resolution problems with arithmetical overload of the operators. The BaseFlags/Flags types are now completly "isolated" from the uint64_t, it has the advantage of showing where unwanted previous implicit conversions where occuring. * Change: In Katana::Constants, Flags values are now of BaseFlags type instead of uint64_t. * Change: In Anabatic::Dijkstra, lots of log cleanup. * Change: In Anabatic::GCell::getSide(), make the "shrink" parameter visible to allow to substract the topmost and rightmost track for axis span computation in AutoSegment::computeOptimal(). Used for analog mode. * Change: In NetRoutingState, added a flag for analog mode. Use uint32_t for the flags type. * New: In Isobar, export the NetRoutingState and NetRoutingExtension objects.
2017-05-20 05:33:12 -05:00
for ( AutoSegment* aligned : aligneds ) {
cdebug_log(145,0) << "Applying constraint on: " << aligned << endl;
aligned->setOptimalMin( optimalMin );
aligned->setOptimalMax( optimalMax );
processeds.insert( aligned );
}
// cdebug_log(145,0) << "Applying constraint on: " << this << endl;
// setOptimalMin( optimalMin );
// setOptimalMax( optimalMax );
// processeds.insert( this );
// if (not isNotAligned()) {
// for ( AutoSegment* autoSegment : getAligneds() ) {
// cdebug_log(145,0) << "Applying constraint on: " << autoSegment << endl;
// autoSegment->setOptimalMin( optimalMin );
// autoSegment->setOptimalMax( optimalMax );
// processeds.insert( autoSegment );
// }
// }
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
cdebug_tabw(145,-1);
}
Replace "unsigned int" by "Flags" in all AutoSegments collections. * Change: In Anabatic::AutoSegments collections, change the type of all the flags that where in "unsigned int" (32 bits) to Flags (uint64_t) as there is now more than 32 flags for functions. * New: In Ababatic::Constants, added new flag Flags::WithPerpands, which makes the number of flags tip over 32 bits, thus making mandatory to uses Flags and not unsigned int. * New: In Anabatic::AutoSegments_Perpandiculars, manage a new flag Flags::WithDoglegs to allow to propagate through global segments that are connecteds via doglegs on local segments. Meaning that there is a good chance that they could be aligned. Slighly change the way we propagate on aligned segments: no longer check for VTee or HTee, but only for same direction and layer as master. * New: In Anabatic & Katana, replace all the "int", "long" and their variants by the less implementation ambiguous "int32_t", "int64_t" (and variant). This should help to better detect bit trucation in flags. Use the type to give a hint about the flags kind: - Type "Flags", for flags shared among Anabatic & Katana functions/methods (may also appear in some objects states). - Type "uint32_t" for flags belonging to an object internal state of from Hurricane functions flags (those should be grouped in a Flag subclass in a perfect world).
2017-05-16 07:53:33 -05:00
AutoSegment* AutoSegment::canonize ( Flags flags )
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
{
cdebug_log(149,0) << "canonize() - " << this << endl;
// if (isCanonical() and isGlobal()) {
// cdebug_log(149,0) << "* " << this << " canonical" << endl;
// return this;
// }
vector<AutoSegment*> segments;
AutoSegment* canonical = this;
bool hasCanonical = isCanonical();
bool hasGlobal = isGlobal();
if (not isNotAligned()) {
forEach( AutoSegment*, isegment, getAligneds(flags) ) {
if (isegment->isFixed()) continue;
hasGlobal = hasGlobal or isegment->isGlobal();
segments.push_back( *isegment );
if (not hasCanonical) {
if (isegment->isCanonical()) {
cdebug_log(149,0) << "* " << *isegment << " canonical already set" << endl;
canonical = *isegment;
hasCanonical = true;
}
if (CompareId()(*isegment,canonical)) canonical = *isegment;
}
}
canonical->setFlags( SegCanonical );
if (hasGlobal) {
for ( size_t i=0 ; i<segments.size() ; ++i ) {
if (not segments[i]->isGlobal())
segments[i]->setFlags( SegWeakGlobal );
}
} else {
for ( size_t i=0 ; i<segments.size() ; ++i )
segments[i]->unsetFlags( SegWeakGlobal );
}
if (segments.empty()) setFlags( SegNotAligned );
if (isCanonical()) { cdebug_log(149,0) << "* " << this << " canonical" << endl; }
else {
cdebug_log(149,0) << "* " << this << " not canonical" << endl;
cdebug_log(149,0) << "* " << canonical << " *is* the canonical" << endl;
}
} else {
setFlags ( SegCanonical );
unsetFlags( SegWeakGlobal );
}
return canonical;
}
size_t AutoSegment::getAlignedContacts ( map<AutoContact*,int>& innerContacts ) const
{
map<AutoContact*,int>::iterator icontact;
innerContacts.clear();
innerContacts.insert( make_pair(getAutoSource(),0x1) );
innerContacts.insert( make_pair(getAutoTarget(),0x4) );
if (not isNotAligned()) {
forEach ( AutoSegment*, isegment, const_cast<AutoSegment*>(this)->getAligneds() ) {
if ( (icontact = innerContacts.find(isegment->getAutoSource())) != innerContacts.end() ) {
if (icontact->second & 0x1) icontact->second |= 0x2;
else icontact->second |= 0x1;
} else
innerContacts.insert( make_pair(getAutoSource(),0x1) );
if ( (icontact = innerContacts.find(isegment->getAutoTarget())) != innerContacts.end() ) {
if (icontact->second & 0x4) icontact->second |= 0x8;
else icontact->second |= 0x4;
} else
innerContacts.insert( make_pair(getAutoTarget(),0x4) );
}
}
return innerContacts.size();
}
Interval AutoSegment::getMinSpanU () const
{
map<AutoContact*,int> contacts;
map<AutoContact*,int>::iterator icontact;
getAlignedContacts( contacts );
Replace "unsigned int" by "Flags" in all AutoSegments collections. * Change: In Anabatic::AutoSegments collections, change the type of all the flags that where in "unsigned int" (32 bits) to Flags (uint64_t) as there is now more than 32 flags for functions. * New: In Ababatic::Constants, added new flag Flags::WithPerpands, which makes the number of flags tip over 32 bits, thus making mandatory to uses Flags and not unsigned int. * New: In Anabatic::AutoSegments_Perpandiculars, manage a new flag Flags::WithDoglegs to allow to propagate through global segments that are connecteds via doglegs on local segments. Meaning that there is a good chance that they could be aligned. Slighly change the way we propagate on aligned segments: no longer check for VTee or HTee, but only for same direction and layer as master. * New: In Anabatic & Katana, replace all the "int", "long" and their variants by the less implementation ambiguous "int32_t", "int64_t" (and variant). This should help to better detect bit trucation in flags. Use the type to give a hint about the flags kind: - Type "Flags", for flags shared among Anabatic & Katana functions/methods (may also appear in some objects states). - Type "uint32_t" for flags belonging to an object internal state of from Hurricane functions flags (those should be grouped in a Flag subclass in a perfect world).
2017-05-16 07:53:33 -05:00
DbU::Unit spanMin = DbU::Min;
DbU::Unit spanMax = DbU::Max;
Interval constraints;
Flags direction = getDirection();
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
for ( icontact=contacts.begin() ; icontact != contacts.end() ; icontact++ ) {
constraints = icontact->first->getUConstraints( direction );
if (icontact->second == 0x1) {
spanMin = max( spanMin, constraints.getVMax() );
}
if (icontact->second == 0x4) {
spanMax = min( spanMax, constraints.getVMin() );
}
}
return Interval(spanMin,spanMax);
}
size_t AutoSegment::getPerpandicularsBound ( set<AutoSegment*>& bounds )
{
map<AutoContact*,int> contacts;
map<AutoContact*,int>::iterator icontact;
getAlignedContacts( contacts );
for ( icontact=contacts.begin() ; icontact != contacts.end() ; icontact++ ) {
if ( (icontact->second == 0x1) or (icontact->second == 0x4) ) {
forEach ( Segment*, isegment, icontact->first->getSlaveComponents().getSubSet<Segment*>() ) {
AutoSegment* autoSegment = Session::lookup ( *isegment );
if (not autoSegment) continue;
if (autoSegment->getDirection() == getDirection()) continue;
bounds.insert( autoSegment );
}
}
}
return bounds.size();
}
bool AutoSegment::isUTurn () const
{
if (isGlobal()) return false;
AutoContact* source = getAutoSource();
AutoContact* target = getAutoTarget();
cerr << "AutoSegment::isUTurn():" << endl;
if (not source->isTurn() or not target->isTurn()) return false;
cerr << " Turn connected" << endl;
AutoSegment* perpandicular = source->getPerpandicular( this );
bool onPSourceSource = (perpandicular->getAutoSource() == source);
perpandicular = target->getPerpandicular( this );
bool onPTargetSource = (perpandicular->getAutoSource() == target);
cerr << " PSource:" << onPSourceSource << " PTarget:" << onPTargetSource << endl;
return not (onPSourceSource xor onPTargetSource);
}
bool AutoSegment::isReduceCandidate () const
{
if (isGlobal()) return false;
if (not isSpinTopOrBottom()) return false;
if (_reduceds) return false;
AutoContact* source = getAutoSource();
AutoContact* target = getAutoTarget();
if (not source->isTurn() or not target->isTurn()) return false;
return true;
}
bool AutoSegment::canReduce () const
{
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
cdebug_log(159,0) << "AutoSegment::canReduce():" << this << endl;
cdebug_log(159,0) << " _reduceds:" << _reduceds << endl;
if (isGlobal() or isDrag() or isFixed()) return false;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
if (not isSpinTopOrBottom()) return false;
if (_reduceds) return false;
AutoContact* source = getAutoSource();
AutoContact* target = getAutoTarget();
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
cdebug_log(159,0) << " source:" << source->isHTee() << "+" << source->isVTee() << endl;
cdebug_log(159,0) << " target:" << target->isHTee() << "+" << target->isVTee() << endl;
if ( ((source->isHTee() or target->isHTee()) and isHorizontal())
or ((source->isVTee() or target->isVTee()) and isVertical ()) ) return false;
// if ( source->isHTee() or source->isVTee()
// or target->isHTee() or target->isVTee() ) return false;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
unsigned int perpandicularDepth = getDepth();
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
if (isSpinBottom()) {
if (perpandicularDepth > 0) --perpandicularDepth;
} else if (isSpinTop()) {
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
++perpandicularDepth;
if (perpandicularDepth >= Session::getDepth()) return false;
} else
return false;
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
cdebug_log(159,0) << " length:" << DbU::getValueString(getLength()) << endl;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
if (getLength() >= (Session::getPitch(perpandicularDepth) * 2)) return false;
return true;
}
bool AutoSegment::reduce ()
{
Fix segment AutoSegment::reduce() multiple-call (must run once). * Bug: In Anabatic::AutoSegment::reduce(), no longer reduce *again* an already reduced segment. This bug was seen through a side effect, it was causing the "reduced perpandicular" counter to be incremented too much and so it cannot get back to zero, blocking the layer reduction (AutoSegment::reduceDoglegLayer()) and some bad VIAs where left over. More debug information in reduce & raise functions. * Change: In Katana::Session::_revalidate(), process to be raised segments *before* segments that can be reduceds.
2020-11-17 08:02:50 -06:00
if (isReduced()) return false;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
if (not canReduce()) return false;
Fix segment AutoSegment::reduce() multiple-call (must run once). * Bug: In Anabatic::AutoSegment::reduce(), no longer reduce *again* an already reduced segment. This bug was seen through a side effect, it was causing the "reduced perpandicular" counter to be incremented too much and so it cannot get back to zero, blocking the layer reduction (AutoSegment::reduceDoglegLayer()) and some bad VIAs where left over. More debug information in reduce & raise functions. * Change: In Katana::Session::_revalidate(), process to be raised segments *before* segments that can be reduceds.
2020-11-17 08:02:50 -06:00
cdebug_log(159,0) << "AutoSegment::reduce():" << this << endl;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
AutoContact* source = getAutoSource();
AutoContact* target = getAutoTarget();
_flags |= SegIsReduced;
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
for ( AutoSegment* perpandicular : source->getAutoSegments() ) {
if (perpandicular == this) continue;
perpandicular->incReduceds();
}
for ( AutoSegment* perpandicular : target->getAutoSegments() ) {
if (perpandicular == this) continue;
perpandicular->incReduceds();
}
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
return true;
}
bool AutoSegment::mustRaise () const
{
if (not (_flags & SegIsReduced)) return false;
unsigned int perpandicularDepth = getDepth();
if (isSpinBottom()) --perpandicularDepth;
else if (isSpinTop ()) ++perpandicularDepth;
else return true;
return (getLength() >= (Session::getPitch(perpandicularDepth) * 2));
}
bool AutoSegment::raise ()
{
if (not (_flags & SegIsReduced)) return false;
Fix segment AutoSegment::reduce() multiple-call (must run once). * Bug: In Anabatic::AutoSegment::reduce(), no longer reduce *again* an already reduced segment. This bug was seen through a side effect, it was causing the "reduced perpandicular" counter to be incremented too much and so it cannot get back to zero, blocking the layer reduction (AutoSegment::reduceDoglegLayer()) and some bad VIAs where left over. More debug information in reduce & raise functions. * Change: In Katana::Session::_revalidate(), process to be raised segments *before* segments that can be reduceds.
2020-11-17 08:02:50 -06:00
cdebug_log(159,0) << "AutoSegment::raise():" << this << endl;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
AutoContact* source = getAutoSource();
AutoContact* target = getAutoTarget();
_flags &= ~SegIsReduced;
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
for ( AutoSegment* perpandicular : source->getAutoSegments() ) {
if (perpandicular == this) continue;
Fix segment AutoSegment::reduce() multiple-call (must run once). * Bug: In Anabatic::AutoSegment::reduce(), no longer reduce *again* an already reduced segment. This bug was seen through a side effect, it was causing the "reduced perpandicular" counter to be incremented too much and so it cannot get back to zero, blocking the layer reduction (AutoSegment::reduceDoglegLayer()) and some bad VIAs where left over. More debug information in reduce & raise functions. * Change: In Katana::Session::_revalidate(), process to be raised segments *before* segments that can be reduceds.
2020-11-17 08:02:50 -06:00
cdebug_log(159,0) << "dec PP:" << perpandicular << endl;
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
perpandicular->decReduceds();
}
for ( AutoSegment* perpandicular : target->getAutoSegments() ) {
if (perpandicular == this) continue;
Fix segment AutoSegment::reduce() multiple-call (must run once). * Bug: In Anabatic::AutoSegment::reduce(), no longer reduce *again* an already reduced segment. This bug was seen through a side effect, it was causing the "reduced perpandicular" counter to be incremented too much and so it cannot get back to zero, blocking the layer reduction (AutoSegment::reduceDoglegLayer()) and some bad VIAs where left over. More debug information in reduce & raise functions. * Change: In Katana::Session::_revalidate(), process to be raised segments *before* segments that can be reduceds.
2020-11-17 08:02:50 -06:00
cdebug_log(159,0) << "dec PP:" << perpandicular << endl;
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
perpandicular->decReduceds();
}
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
return true;
}
Replace "unsigned int" by "Flags" in all AutoSegments collections. * Change: In Anabatic::AutoSegments collections, change the type of all the flags that where in "unsigned int" (32 bits) to Flags (uint64_t) as there is now more than 32 flags for functions. * New: In Ababatic::Constants, added new flag Flags::WithPerpands, which makes the number of flags tip over 32 bits, thus making mandatory to uses Flags and not unsigned int. * New: In Anabatic::AutoSegments_Perpandiculars, manage a new flag Flags::WithDoglegs to allow to propagate through global segments that are connecteds via doglegs on local segments. Meaning that there is a good chance that they could be aligned. Slighly change the way we propagate on aligned segments: no longer check for VTee or HTee, but only for same direction and layer as master. * New: In Anabatic & Katana, replace all the "int", "long" and their variants by the less implementation ambiguous "int32_t", "int64_t" (and variant). This should help to better detect bit trucation in flags. Use the type to give a hint about the flags kind: - Type "Flags", for flags shared among Anabatic & Katana functions/methods (may also appear in some objects states). - Type "uint32_t" for flags belonging to an object internal state of from Hurricane functions flags (those should be grouped in a Flag subclass in a perfect world).
2017-05-16 07:53:33 -05:00
void AutoSegment::changeDepth ( unsigned int depth, Flags flags )
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
{
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
DebugSession::open( getNet(), 145, 150 );
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
cdebug_log(149,1) << "changeDepth() " << depth << " - " << this << endl;
Session::invalidate( getNet() );
_changeDepth( depth, flags & ~Flags::Propagate );
if ((flags & Flags::Propagate) and not isNotAligned()) {
Basic support for FreePDK 45 completed. * New: In Commons, inspector support for std::pair<T,U>. * New: In Hurricane::Layer, ContactLayer & ViaLayer, support for non square VIAs. The hole (cut) remains square, but the various metal extensions can now be different in X and Y. The ::getEnclosure() method now takes a flag EnclosureH / EnclosureV. * New: In Hurricane::DbU, inspector support for: std::pair<DbU::Unit,DbU::Unit> std::array<DbU::Unit,3> Must be defined here as DbU do not exists yet in Commons.h * Bug: In Hurricane::Interval::getSize(), when the interval is "full span", do not return the difference between min and max, but directly DbU::Max. (the previous result was -1 !) * New: In CRL Core Python/Technology.py, support for non square VIAs in the configuration files. Applied to FreePDK 45. * New: In CRL::RoutingGauge, added a "symbolic" flag to tell if a gauge is for symbolic layout or not. Exported to Python. * New: In Anabatic::AutoHorizontal::updatePosition(), differentiated computation for soure or target taking account of the VIA extension in the right segment metal (due to non-square VIAs). * Change: In Anabatic::AutoHorizontal::_makeDogleg(), the dogleg is UP for HV gauges and DOWN for VH. * New: In Anabatic::AutoSegment::_initialize(), create a cache of the various extension length for each layer (viaToTop, viaToBottom, viaToSame). New implementation of getExtensionCap() using the previous cached extension table. See updatePositions(). New static functions to access the extension cache in the header: getViaTotopCap() ... * Change: In Anabatic::AutoSegment, in various update methods, updateOrient() must always be called *before* updatePositions() as extensions are dependant on source/target. * New: In Anabatic::AutoSegment::getEndAxes() compute the position of the first source and last target position (center/axes) on an *aligned* set of segments. * New: In Anabatic::AutoSegment, add a new state flag SegAxisFixed to signal segments that can be put on only one track. Specific case to VH gauge for a M1 vertical terminal with a M2 vertical segment. The M2 is effectively bound to the M1 axis position. * Bug: In Anabatic::NetBuilderVH::_do_xG_xM1_xM3(), in case of E/W global and only one RoutingPad the connexion to the RoutingPad was duplicated. It was valid, but totally stupid. * Bug: In Anabatic::Session::_canonize(), for an aligned segment set, intersect the user constraints from all segments instead of only considering the canonical one. Issue a warning about too tight constraints only for symbolic gauges. It may be correct for the real ones. * New: In Katata::DataNegociate::update(), more accurate computation of the perpandicular free interval. Use segment extension cap calculation. Create a special case for fixed axis segments allowing them to find alternative free interval, try under source and under target as they are likely to be draggable segments. * Change: In Katana::Manipulator::relax(), use the extension cap value to compute the axis of the perpandicular segemnts. * Change: In Katana::Manipulator::moveUp(), now move up the whole set of aligned segments instead of just the canonical one. * Change: In Katana::NegociateWindow::loadRoutingPads(), more accurate TrackMarkers insertions for fixed terminals. * New: In Katana::RoutingEvent::Key::Compare::operator(), segments with fixed axis are processed prior to any others. * New: In Katana::RoutingEventLoop, store segment pointers instead of ids to generate more accurate error messages. * Change: In Katana::RoutingPlane::create(), perform local track assignment only for HV gauges. * Change: In Katana::SegmentFsm::_slackenLocal(), add a "dragMinimize" step in the automaton. Mutliple states transitions can occurs in a row if an action fails. * New: In Katana::Session::_toIntervalAxis(), normalize interval bounds so they are on track positions (by shrinking the interval). * Bug: In Katana::TrackMarker CTOR, the weigh computation was wrong.
2018-02-17 13:27:38 -06:00
cdebug_log(149,0) << "Propagate to aligneds." << endl;
for ( AutoSegment* segment : getAligneds(Flags::NoCheckLayer) ) {
segment->_changeDepth( depth, flags & ~Flags::Propagate );
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
}
}
cdebug_tabw(149,-1);
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
DebugSession::close();
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
}
Replace "unsigned int" by "Flags" in all AutoSegments collections. * Change: In Anabatic::AutoSegments collections, change the type of all the flags that where in "unsigned int" (32 bits) to Flags (uint64_t) as there is now more than 32 flags for functions. * New: In Ababatic::Constants, added new flag Flags::WithPerpands, which makes the number of flags tip over 32 bits, thus making mandatory to uses Flags and not unsigned int. * New: In Anabatic::AutoSegments_Perpandiculars, manage a new flag Flags::WithDoglegs to allow to propagate through global segments that are connecteds via doglegs on local segments. Meaning that there is a good chance that they could be aligned. Slighly change the way we propagate on aligned segments: no longer check for VTee or HTee, but only for same direction and layer as master. * New: In Anabatic & Katana, replace all the "int", "long" and their variants by the less implementation ambiguous "int32_t", "int64_t" (and variant). This should help to better detect bit trucation in flags. Use the type to give a hint about the flags kind: - Type "Flags", for flags shared among Anabatic & Katana functions/methods (may also appear in some objects states). - Type "uint32_t" for flags belonging to an object internal state of from Hurricane functions flags (those should be grouped in a Flag subclass in a perfect world).
2017-05-16 07:53:33 -05:00
void AutoSegment::_changeDepth ( unsigned int depth, Flags flags )
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
{
cdebug_log(149,1) << "_changeDepth() - " << this << endl;
Topology update was not correctly done after layer change. * Bug: In Anabatic::AutoSegment, Anabatic::AutoContact and derived classes, the recursive invalidation mechanism did have cross-call troubles. - The various AutoContact::_invalidate() now take the flag into account instead of resetting it. This is for Flags::Topology to be passed along from segments to contacts. - In AutoSegment::invalidate(), do not invalidate S/T contacts already invalidateds. In the same way, do not invalidate aligned segments already invalidateds (save some recursion). - In AutoSegment::_changeDepth(), call invalidate with Flags::Topology (to force contact update) and Flags::NoCheckLayer to avoid disgraceful warnings. * New: Anabatic::NetBuilder::_do_xG_1M1() with (x >= 2) new set of optimized topologies when there is multiple globals but only one terminal. The default one was making two connections to the same RoutingPad, which was making clutter.
2018-03-27 11:03:51 -05:00
invalidate( Flags::Topology|Flags::NoCheckLayer );
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
setFlags( SegInvalidatedLayer|SegInvalidatedSource|SegInvalidatedTarget );
const Layer* newLayer = Session::getRoutingGauge()->getRoutingLayer(depth);
if (getLayer() != newLayer) {
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
cdebug_log(149,0) << "Effective layer change to " << depth << "/" << newLayer << endl;
In Anabatic, correctly set the wire width according to the routing gauge. * New: Anabatic::AutoContact::setLayerAndWidth() to set both layer and VIA width/side according to the RoutingGauge. If the delta in zero, use the metal gauge, and the VIA gauge otherwise. * Bug: In Anabatic::AutoContactTerminal, Anabatic::AutoContactTurn, Anabatic::AutoContactHTee & Anabatic::AutoContactVTee, in the updateTopology() method, set both the layer and the VIA with when there is a change of layer. Note that this default size may be overriden later by updateSeize() in the case of non-default width segments. * New: In Anabatic::AutoSegment, new overload setLayer(size_t) to set both layer and segment with according to the routing gauge. * Bug: In Anabatic::AutoHorizontal and Anabatic::AutoVertical, in method _makeDogleg(), make use of the new setLayer() to correctly set up the wire width. Idem for Anabatic::AutoSegment::changeDepth() and ::makeDogleg(), and in Anabatic::LayerAssign.
2019-02-12 05:43:09 -06:00
setLayer( depth );
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
getAutoSource()->invalidate( Flags::Topology|Flags::NoCheckLayer );
getAutoTarget()->invalidate( Flags::Topology|Flags::NoCheckLayer );
}
In AutoSegment::_changeDepth(), correct invalidation of GCells. * Bug: In Anabatic::AutoSegment::_changeDepth(), GCells under the segment must be invalidated in order for the density to be recomputed. But we were using GCell::invalidate() which is meant for the quad-tree and graphic system, so the density almost never updated in case of a layer change, leading to over-congested GCells in the top layers (110% !). Now we directly set the Invalidated flag.
2018-06-08 05:20:37 -05:00
vector<GCell*> gcells;
getGCells( gcells );
for ( size_t i=0 ; i<gcells.size() ; ++i ) {
gcells[i]->flags() |= Flags::Invalidated;
Tuning of Anabatic & Katana for AMS 350nm (c35b4). * New: In Anabatic::LayerAssign, a new mode "LayerAssignNoGlobalM2V" has been added to manage the 3 routing metal technologies like AMS 350nm. The standard cells have all their connectors punctual and aligned on an horizontal line in the middle of the Cell. This is a design inherited from the channel routing times that makes global routing in metal2 *through* a standard cell almost impossible (except for directly neighboring cells). Thus, the layer assignment must move up all the metal2 that span more than two GCells. * Change: In Katana::Manipulator::ripupPerpandiculars(), when a perpandicular has a fixed axis and it's (sole) underlying track is taken, ripup the other segment to ensure the perpandicular interval will be free. We need to do that, because in DataNegociate::update(), if the perpandicular is taken, the perpandicular interval will always been empty preventing the segment to have a valid track span (so directly failing). * Change: In Katana::SegmentFsm::conflictSolveByPlaceds() the fixed axis segments must also been taken into account as conflicts. Also correct a small bug, the first conflicting segment in the interval was not taken into account correctly.
2018-06-11 09:44:26 -05:00
cdebug_log(149,0) << "changeDepth() " << gcells[i] << this << " " << endl;
In AutoSegment::_changeDepth(), correct invalidation of GCells. * Bug: In Anabatic::AutoSegment::_changeDepth(), GCells under the segment must be invalidated in order for the density to be recomputed. But we were using GCell::invalidate() which is meant for the quad-tree and graphic system, so the density almost never updated in case of a layer change, leading to over-congested GCells in the top layers (110% !). Now we directly set the Invalidated flag.
2018-06-08 05:20:37 -05:00
}
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
if (not (flags & Flags::WithNeighbors)) {
Tuning of Anabatic & Katana for AMS 350nm (c35b4). * New: In Anabatic::LayerAssign, a new mode "LayerAssignNoGlobalM2V" has been added to manage the 3 routing metal technologies like AMS 350nm. The standard cells have all their connectors punctual and aligned on an horizontal line in the middle of the Cell. This is a design inherited from the channel routing times that makes global routing in metal2 *through* a standard cell almost impossible (except for directly neighboring cells). Thus, the layer assignment must move up all the metal2 that span more than two GCells. * Change: In Katana::Manipulator::ripupPerpandiculars(), when a perpandicular has a fixed axis and it's (sole) underlying track is taken, ripup the other segment to ensure the perpandicular interval will be free. We need to do that, because in DataNegociate::update(), if the perpandicular is taken, the perpandicular interval will always been empty preventing the segment to have a valid track span (so directly failing). * Change: In Katana::SegmentFsm::conflictSolveByPlaceds() the fixed axis segments must also been taken into account as conflicts. Also correct a small bug, the first conflicting segment in the interval was not taken into account correctly.
2018-06-11 09:44:26 -05:00
cdebug_tabw(149,-1);
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
return;
}
Topology update was not correctly done after layer change. * Bug: In Anabatic::AutoSegment, Anabatic::AutoContact and derived classes, the recursive invalidation mechanism did have cross-call troubles. - The various AutoContact::_invalidate() now take the flag into account instead of resetting it. This is for Flags::Topology to be passed along from segments to contacts. - In AutoSegment::invalidate(), do not invalidate S/T contacts already invalidateds. In the same way, do not invalidate aligned segments already invalidateds (save some recursion). - In AutoSegment::_changeDepth(), call invalidate with Flags::Topology (to force contact update) and Flags::NoCheckLayer to avoid disgraceful warnings. * New: Anabatic::NetBuilder::_do_xG_1M1() with (x >= 2) new set of optimized topologies when there is multiple globals but only one terminal. The default one was making two connections to the same RoutingPad, which was making clutter.
2018-03-27 11:03:51 -05:00
for ( AutoSegment* segment : getCachedOnSourceContact(Flags::DirectionMask) ) {
if (segment == this) continue;
if (segment->isGlobal ()) continue;
if (segment->isTerminal()) continue;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
Topology update was not correctly done after layer change. * Bug: In Anabatic::AutoSegment, Anabatic::AutoContact and derived classes, the recursive invalidation mechanism did have cross-call troubles. - The various AutoContact::_invalidate() now take the flag into account instead of resetting it. This is for Flags::Topology to be passed along from segments to contacts. - In AutoSegment::invalidate(), do not invalidate S/T contacts already invalidateds. In the same way, do not invalidate aligned segments already invalidateds (save some recursion). - In AutoSegment::_changeDepth(), call invalidate with Flags::Topology (to force contact update) and Flags::NoCheckLayer to avoid disgraceful warnings. * New: Anabatic::NetBuilder::_do_xG_1M1() with (x >= 2) new set of optimized topologies when there is multiple globals but only one terminal. The default one was making two connections to the same RoutingPad, which was making clutter.
2018-03-27 11:03:51 -05:00
if (not (segment->isHorizontal() xor isHorizontal()))
segment->_changeDepth( depth , Flags::NoFlags );
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
else
Topology update was not correctly done after layer change. * Bug: In Anabatic::AutoSegment, Anabatic::AutoContact and derived classes, the recursive invalidation mechanism did have cross-call troubles. - The various AutoContact::_invalidate() now take the flag into account instead of resetting it. This is for Flags::Topology to be passed along from segments to contacts. - In AutoSegment::invalidate(), do not invalidate S/T contacts already invalidateds. In the same way, do not invalidate aligned segments already invalidateds (save some recursion). - In AutoSegment::_changeDepth(), call invalidate with Flags::Topology (to force contact update) and Flags::NoCheckLayer to avoid disgraceful warnings. * New: Anabatic::NetBuilder::_do_xG_1M1() with (x >= 2) new set of optimized topologies when there is multiple globals but only one terminal. The default one was making two connections to the same RoutingPad, which was making clutter.
2018-03-27 11:03:51 -05:00
segment->_changeDepth( depth-1, Flags::NoFlags );
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
}
Topology update was not correctly done after layer change. * Bug: In Anabatic::AutoSegment, Anabatic::AutoContact and derived classes, the recursive invalidation mechanism did have cross-call troubles. - The various AutoContact::_invalidate() now take the flag into account instead of resetting it. This is for Flags::Topology to be passed along from segments to contacts. - In AutoSegment::invalidate(), do not invalidate S/T contacts already invalidateds. In the same way, do not invalidate aligned segments already invalidateds (save some recursion). - In AutoSegment::_changeDepth(), call invalidate with Flags::Topology (to force contact update) and Flags::NoCheckLayer to avoid disgraceful warnings. * New: Anabatic::NetBuilder::_do_xG_1M1() with (x >= 2) new set of optimized topologies when there is multiple globals but only one terminal. The default one was making two connections to the same RoutingPad, which was making clutter.
2018-03-27 11:03:51 -05:00
for ( AutoSegment* segment : getCachedOnTargetContact(Flags::DirectionMask) ) {
if (segment == this) continue;
if (segment->isGlobal ()) continue;
if (segment->isTerminal()) continue;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
Topology update was not correctly done after layer change. * Bug: In Anabatic::AutoSegment, Anabatic::AutoContact and derived classes, the recursive invalidation mechanism did have cross-call troubles. - The various AutoContact::_invalidate() now take the flag into account instead of resetting it. This is for Flags::Topology to be passed along from segments to contacts. - In AutoSegment::invalidate(), do not invalidate S/T contacts already invalidateds. In the same way, do not invalidate aligned segments already invalidateds (save some recursion). - In AutoSegment::_changeDepth(), call invalidate with Flags::Topology (to force contact update) and Flags::NoCheckLayer to avoid disgraceful warnings. * New: Anabatic::NetBuilder::_do_xG_1M1() with (x >= 2) new set of optimized topologies when there is multiple globals but only one terminal. The default one was making two connections to the same RoutingPad, which was making clutter.
2018-03-27 11:03:51 -05:00
if (not (segment->isHorizontal() xor isHorizontal()))
segment->_changeDepth( depth , Flags::NoFlags );
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
else
Topology update was not correctly done after layer change. * Bug: In Anabatic::AutoSegment, Anabatic::AutoContact and derived classes, the recursive invalidation mechanism did have cross-call troubles. - The various AutoContact::_invalidate() now take the flag into account instead of resetting it. This is for Flags::Topology to be passed along from segments to contacts. - In AutoSegment::invalidate(), do not invalidate S/T contacts already invalidateds. In the same way, do not invalidate aligned segments already invalidateds (save some recursion). - In AutoSegment::_changeDepth(), call invalidate with Flags::Topology (to force contact update) and Flags::NoCheckLayer to avoid disgraceful warnings. * New: Anabatic::NetBuilder::_do_xG_1M1() with (x >= 2) new set of optimized topologies when there is multiple globals but only one terminal. The default one was making two connections to the same RoutingPad, which was making clutter.
2018-03-27 11:03:51 -05:00
segment->_changeDepth( depth-1, Flags::NoFlags );
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
}
cdebug_tabw(149,-1);
}
Replace "unsigned int" by "Flags" in all AutoSegments collections. * Change: In Anabatic::AutoSegments collections, change the type of all the flags that where in "unsigned int" (32 bits) to Flags (uint64_t) as there is now more than 32 flags for functions. * New: In Ababatic::Constants, added new flag Flags::WithPerpands, which makes the number of flags tip over 32 bits, thus making mandatory to uses Flags and not unsigned int. * New: In Anabatic::AutoSegments_Perpandiculars, manage a new flag Flags::WithDoglegs to allow to propagate through global segments that are connecteds via doglegs on local segments. Meaning that there is a good chance that they could be aligned. Slighly change the way we propagate on aligned segments: no longer check for VTee or HTee, but only for same direction and layer as master. * New: In Anabatic & Katana, replace all the "int", "long" and their variants by the less implementation ambiguous "int32_t", "int64_t" (and variant). This should help to better detect bit trucation in flags. Use the type to give a hint about the flags kind: - Type "Flags", for flags shared among Anabatic & Katana functions/methods (may also appear in some objects states). - Type "uint32_t" for flags belonging to an object internal state of from Hurricane functions flags (those should be grouped in a Flag subclass in a perfect world).
2017-05-16 07:53:33 -05:00
bool AutoSegment::canSlacken ( Flags flags ) const
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
{
cdebug_log(149,0) << "AutoSegment::canSlacken()" << endl;
if (not isGlobal() and not (flags & Flags::Propagate)) return false;
if (_canSlacken()) return true;
if ((flags & Flags::Propagate) and not isNotAligned()) {
Bug fix: Never slacken a segment from a Pin. * New: In Anabatic::AutoContactTerminal, new method "isOnPin()" to know if the terminal contact is on a Pin. Introduced in base class. * Bug: In Anabatic::AutoHorizontal::_canSlacken() & AutoVertical, forbid slackening from a Pin. As Pin are all on the side of the cell (aligneds) and the perpandicular segment will be locked on the Pin, it generates intractable overlaps for the router.
2020-08-02 11:13:47 -05:00
for ( AutoSegment* segment : const_cast<AutoSegment*>(this)->getAligneds() ) {
if (segment->_canSlacken()) return true;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
}
}
return false;
}
Replace "unsigned int" by "Flags" in all AutoSegments collections. * Change: In Anabatic::AutoSegments collections, change the type of all the flags that where in "unsigned int" (32 bits) to Flags (uint64_t) as there is now more than 32 flags for functions. * New: In Ababatic::Constants, added new flag Flags::WithPerpands, which makes the number of flags tip over 32 bits, thus making mandatory to uses Flags and not unsigned int. * New: In Anabatic::AutoSegments_Perpandiculars, manage a new flag Flags::WithDoglegs to allow to propagate through global segments that are connecteds via doglegs on local segments. Meaning that there is a good chance that they could be aligned. Slighly change the way we propagate on aligned segments: no longer check for VTee or HTee, but only for same direction and layer as master. * New: In Anabatic & Katana, replace all the "int", "long" and their variants by the less implementation ambiguous "int32_t", "int64_t" (and variant). This should help to better detect bit trucation in flags. Use the type to give a hint about the flags kind: - Type "Flags", for flags shared among Anabatic & Katana functions/methods (may also appear in some objects states). - Type "uint32_t" for flags belonging to an object internal state of from Hurricane functions flags (those should be grouped in a Flag subclass in a perfect world).
2017-05-16 07:53:33 -05:00
bool AutoSegment::slacken ( Flags flags )
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
{
bool success = false;
success = success or _slacken( flags );
if ((flags & Flags::Propagate) and not isNotAligned()) {
Basic support for FreePDK 45 completed. * New: In Commons, inspector support for std::pair<T,U>. * New: In Hurricane::Layer, ContactLayer & ViaLayer, support for non square VIAs. The hole (cut) remains square, but the various metal extensions can now be different in X and Y. The ::getEnclosure() method now takes a flag EnclosureH / EnclosureV. * New: In Hurricane::DbU, inspector support for: std::pair<DbU::Unit,DbU::Unit> std::array<DbU::Unit,3> Must be defined here as DbU do not exists yet in Commons.h * Bug: In Hurricane::Interval::getSize(), when the interval is "full span", do not return the difference between min and max, but directly DbU::Max. (the previous result was -1 !) * New: In CRL Core Python/Technology.py, support for non square VIAs in the configuration files. Applied to FreePDK 45. * New: In CRL::RoutingGauge, added a "symbolic" flag to tell if a gauge is for symbolic layout or not. Exported to Python. * New: In Anabatic::AutoHorizontal::updatePosition(), differentiated computation for soure or target taking account of the VIA extension in the right segment metal (due to non-square VIAs). * Change: In Anabatic::AutoHorizontal::_makeDogleg(), the dogleg is UP for HV gauges and DOWN for VH. * New: In Anabatic::AutoSegment::_initialize(), create a cache of the various extension length for each layer (viaToTop, viaToBottom, viaToSame). New implementation of getExtensionCap() using the previous cached extension table. See updatePositions(). New static functions to access the extension cache in the header: getViaTotopCap() ... * Change: In Anabatic::AutoSegment, in various update methods, updateOrient() must always be called *before* updatePositions() as extensions are dependant on source/target. * New: In Anabatic::AutoSegment::getEndAxes() compute the position of the first source and last target position (center/axes) on an *aligned* set of segments. * New: In Anabatic::AutoSegment, add a new state flag SegAxisFixed to signal segments that can be put on only one track. Specific case to VH gauge for a M1 vertical terminal with a M2 vertical segment. The M2 is effectively bound to the M1 axis position. * Bug: In Anabatic::NetBuilderVH::_do_xG_xM1_xM3(), in case of E/W global and only one RoutingPad the connexion to the RoutingPad was duplicated. It was valid, but totally stupid. * Bug: In Anabatic::Session::_canonize(), for an aligned segment set, intersect the user constraints from all segments instead of only considering the canonical one. Issue a warning about too tight constraints only for symbolic gauges. It may be correct for the real ones. * New: In Katata::DataNegociate::update(), more accurate computation of the perpandicular free interval. Use segment extension cap calculation. Create a special case for fixed axis segments allowing them to find alternative free interval, try under source and under target as they are likely to be draggable segments. * Change: In Katana::Manipulator::relax(), use the extension cap value to compute the axis of the perpandicular segemnts. * Change: In Katana::Manipulator::moveUp(), now move up the whole set of aligned segments instead of just the canonical one. * Change: In Katana::NegociateWindow::loadRoutingPads(), more accurate TrackMarkers insertions for fixed terminals. * New: In Katana::RoutingEvent::Key::Compare::operator(), segments with fixed axis are processed prior to any others. * New: In Katana::RoutingEventLoop, store segment pointers instead of ids to generate more accurate error messages. * Change: In Katana::RoutingPlane::create(), perform local track assignment only for HV gauges. * Change: In Katana::SegmentFsm::_slackenLocal(), add a "dragMinimize" step in the automaton. Mutliple states transitions can occurs in a row if an action fails. * New: In Katana::Session::_toIntervalAxis(), normalize interval bounds so they are on track positions (by shrinking the interval). * Bug: In Katana::TrackMarker CTOR, the weigh computation was wrong.
2018-02-17 13:27:38 -06:00
for ( AutoSegment* segment : getAligneds() ) {
success = success or segment->_slacken( flags );
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
}
}
return success;
}
Replace "unsigned int" by "Flags" in all AutoSegments collections. * Change: In Anabatic::AutoSegments collections, change the type of all the flags that where in "unsigned int" (32 bits) to Flags (uint64_t) as there is now more than 32 flags for functions. * New: In Ababatic::Constants, added new flag Flags::WithPerpands, which makes the number of flags tip over 32 bits, thus making mandatory to uses Flags and not unsigned int. * New: In Anabatic::AutoSegments_Perpandiculars, manage a new flag Flags::WithDoglegs to allow to propagate through global segments that are connecteds via doglegs on local segments. Meaning that there is a good chance that they could be aligned. Slighly change the way we propagate on aligned segments: no longer check for VTee or HTee, but only for same direction and layer as master. * New: In Anabatic & Katana, replace all the "int", "long" and their variants by the less implementation ambiguous "int32_t", "int64_t" (and variant). This should help to better detect bit trucation in flags. Use the type to give a hint about the flags kind: - Type "Flags", for flags shared among Anabatic & Katana functions/methods (may also appear in some objects states). - Type "uint32_t" for flags belonging to an object internal state of from Hurricane functions flags (those should be grouped in a Flag subclass in a perfect world).
2017-05-16 07:53:33 -05:00
float AutoSegment::getMaxUnderDensity ( Flags flags )
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
{
cdebug_log(149,0) << "AutoSegment::getMaxUnderDensity() " << endl;
size_t depth = Session::getRoutingGauge()->getLayerDepth(getLayer());
vector<GCell*> gcells;
getGCells( gcells );
float maxDensity = 0.0;
for ( size_t i=0 ; i<gcells.size() ; ++i ) {
maxDensity = std::max( maxDensity, gcells[i]->getFeedthroughs(depth) );
}
if ((flags & Flags::Propagate) and not isNotAligned()) {
forEach ( AutoSegment*, isegment, getAligneds() ) {
isegment->getGCells( gcells );
for ( size_t i=0 ; i<gcells.size() ; ++i ) {
maxDensity = std::max( maxDensity, gcells[i]->getFeedthroughs(depth) );
}
}
}
return maxDensity;
}
Replace "unsigned int" by "Flags" in all AutoSegments collections. * Change: In Anabatic::AutoSegments collections, change the type of all the flags that where in "unsigned int" (32 bits) to Flags (uint64_t) as there is now more than 32 flags for functions. * New: In Ababatic::Constants, added new flag Flags::WithPerpands, which makes the number of flags tip over 32 bits, thus making mandatory to uses Flags and not unsigned int. * New: In Anabatic::AutoSegments_Perpandiculars, manage a new flag Flags::WithDoglegs to allow to propagate through global segments that are connecteds via doglegs on local segments. Meaning that there is a good chance that they could be aligned. Slighly change the way we propagate on aligned segments: no longer check for VTee or HTee, but only for same direction and layer as master. * New: In Anabatic & Katana, replace all the "int", "long" and their variants by the less implementation ambiguous "int32_t", "int64_t" (and variant). This should help to better detect bit trucation in flags. Use the type to give a hint about the flags kind: - Type "Flags", for flags shared among Anabatic & Katana functions/methods (may also appear in some objects states). - Type "uint32_t" for flags belonging to an object internal state of from Hurricane functions flags (those should be grouped in a Flag subclass in a perfect world).
2017-05-16 07:53:33 -05:00
bool AutoSegment::canPivotUp ( float reserve, Flags flags ) const
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
{
cdebug_log(149,0) << "AutoSegment::canPivotUp() - " << flags
<< " (reserve:" << reserve << ")" << endl;
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
if ( isLayerChange() or isFixed() or isUnbreakable() ) return false;
Forgot to perform Track re-order after removing zero-length segments. * Change: In Anabatic::Autocontact, replace getMinDepth() and getMaxDepth() by getDepthSpan(). * New: In Anabatic::AutoSegment::canMoveUp(), add an optional check of low up density (Flags::CheckLowUpDensity). Allows to move up a segment if the up density is (very) low, and in this case it's more efficient than breaking it to fit in the lower layer. canMoveUp() is now able to perform the same work as canPivotUp() if *not* supplied the flag Flags::IgnoreContacts. * New: In Katana, in GlobalRouting::DigitalDistance() now take into account the cost of a VIA (currently set to 2.5). Need to known the Net currently routed in the DigitalDistance object itself. * Change: In Katana::Track::Element::canPivotUp(), now also takes a flag parameter. * Change: In Katana::Manipulator, new flag IgnoreContacts to mirror the one of Anabatic. * Change: In Katana::SegmentFsm, allocate once a Manipulator object instead of many times on the fly. In SegmentFsm::_slackenGlobal(), in the slacken state, if the up density is (very) low, bypass to move up instead of slackening. This solve better the routing of the control part of the register file. The register file having a pathological case of terminal placement: many punctual terminals aligneds in METAL2 *and* a grid of METAL2 and METAL3 blockages near below... * Bug: In Katana::Session::_revalidate(), after removing the zero-length segments, forgot to re-order the track, leading to many stranges effects as the indexes where no longer coherent in the Track.
2016-09-20 04:30:45 -05:00
if ( isStrongTerminal() and (not (flags & Flags::AllowTerminal)) ) return false;
if ( isLocal() and (not (flags & Flags::AllowLocal )) ) return false;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
size_t depth = Session::getRoutingGauge()->getLayerDepth( getLayer() );
if (depth+2 >= Session::getRoutingGauge()->getDepth()) return false;
vector<GCell*> gcells;
getGCells( gcells );
for ( size_t i=0 ; i<gcells.size() ; i++ ) {
if (not gcells[i]->hasFreeTrack(depth+2,reserve)) return false;
}
if ( not (flags&Flags::IgnoreContacts) ) {
cdebug_log(149,0) << getAutoSource() << endl;
cdebug_log(149,0) << getAutoTarget() << endl;
cdebug_log(149,0) << "min depths, Segment:" << depth
Forgot to perform Track re-order after removing zero-length segments. * Change: In Anabatic::Autocontact, replace getMinDepth() and getMaxDepth() by getDepthSpan(). * New: In Anabatic::AutoSegment::canMoveUp(), add an optional check of low up density (Flags::CheckLowUpDensity). Allows to move up a segment if the up density is (very) low, and in this case it's more efficient than breaking it to fit in the lower layer. canMoveUp() is now able to perform the same work as canPivotUp() if *not* supplied the flag Flags::IgnoreContacts. * New: In Katana, in GlobalRouting::DigitalDistance() now take into account the cost of a VIA (currently set to 2.5). Need to known the Net currently routed in the DigitalDistance object itself. * Change: In Katana::Track::Element::canPivotUp(), now also takes a flag parameter. * Change: In Katana::Manipulator, new flag IgnoreContacts to mirror the one of Anabatic. * Change: In Katana::SegmentFsm, allocate once a Manipulator object instead of many times on the fly. In SegmentFsm::_slackenGlobal(), in the slacken state, if the up density is (very) low, bypass to move up instead of slackening. This solve better the routing of the control part of the register file. The register file having a pathological case of terminal placement: many punctual terminals aligneds in METAL2 *and* a grid of METAL2 and METAL3 blockages near below... * Bug: In Katana::Session::_revalidate(), after removing the zero-length segments, forgot to re-order the track, leading to many stranges effects as the indexes where no longer coherent in the Track.
2016-09-20 04:30:45 -05:00
<< " S:" << getAutoSource()->getMinDepth()
<< " T:" << getAutoTarget()->getMinDepth() << endl;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
if (getAutoSource()->getMinDepth() < depth) return false;
if (getAutoTarget()->getMinDepth() < depth) return false;
}
if ((flags & Flags::Propagate) and not isNotAligned()) {
forEach ( AutoSegment*, isegment, const_cast<AutoSegment*>(this)->getAligneds(flags) ) {
isegment->getGCells( gcells );
for ( size_t i=0 ; i<gcells.size() ; i++ ) {
if (not gcells[i]->hasFreeTrack(depth+2,reserve)) return false;
}
if (isegment->getAutoSource()->getMinDepth() < depth) return false;
if (isegment->getAutoTarget()->getMinDepth() < depth) return false;
}
} else {
cdebug_log(149,0) << "AutoSegment::canPivotUp() - true [no propagate]" << endl;
return true;
}
cdebug_log(149,0) << "AutoSegment::canPivotUp() - true [propagate]" << endl;
return true;
}
Replace "unsigned int" by "Flags" in all AutoSegments collections. * Change: In Anabatic::AutoSegments collections, change the type of all the flags that where in "unsigned int" (32 bits) to Flags (uint64_t) as there is now more than 32 flags for functions. * New: In Ababatic::Constants, added new flag Flags::WithPerpands, which makes the number of flags tip over 32 bits, thus making mandatory to uses Flags and not unsigned int. * New: In Anabatic::AutoSegments_Perpandiculars, manage a new flag Flags::WithDoglegs to allow to propagate through global segments that are connecteds via doglegs on local segments. Meaning that there is a good chance that they could be aligned. Slighly change the way we propagate on aligned segments: no longer check for VTee or HTee, but only for same direction and layer as master. * New: In Anabatic & Katana, replace all the "int", "long" and their variants by the less implementation ambiguous "int32_t", "int64_t" (and variant). This should help to better detect bit trucation in flags. Use the type to give a hint about the flags kind: - Type "Flags", for flags shared among Anabatic & Katana functions/methods (may also appear in some objects states). - Type "uint32_t" for flags belonging to an object internal state of from Hurricane functions flags (those should be grouped in a Flag subclass in a perfect world).
2017-05-16 07:53:33 -05:00
bool AutoSegment::canPivotDown ( float reserve, Flags flags ) const
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
{
cdebug_log(149,0) << "AutoSegment::canPivotDown()"
<< " (reserve:" << reserve << ")" << endl;
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
if ( isLayerChange() or isFixed() or isUnbreakable() ) return false;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
if ( isStrongTerminal() or isLocal() ) return false;
size_t depth = Session::getRoutingGauge()->getLayerDepth( getLayer() );
if (depth < 3) return false;
vector<GCell*> gcells;
getGCells( gcells );
for ( size_t i=0 ; i<gcells.size() ; i++ ) {
if (not gcells[i]->hasFreeTrack(depth-2,reserve)) return false;
}
cdebug_log(149,0) << getAutoSource() << endl;
cdebug_log(149,0) << getAutoTarget() << endl;
cdebug_log(149,0) << "max depths, Segment:" << depth
<< " S:" << getAutoSource()->getMaxDepth()
<< " T:" << getAutoTarget()->getMaxDepth() << endl;
if (getAutoSource()->getMaxDepth() > depth) return false;
if (getAutoTarget()->getMaxDepth() > depth) return false;
if (not (flags & Flags::Propagate)) {
cdebug_log(149,0) << "AutoSegment::canPivotDown() - true [no propagate]" << endl;
return true;
}
if ((flags & Flags::Propagate) and not isNotAligned()) {
forEach ( AutoSegment*, isegment, const_cast<AutoSegment*>(this)->getAligneds() ) {
isegment->getGCells( gcells );
for ( size_t i=0 ; i<gcells.size() ; i++ ) {
if (not gcells[i]->hasFreeTrack(depth-2,reserve)) return false;
}
if (isegment->getAutoSource()->getMaxDepth() < depth) return false;
if (isegment->getAutoTarget()->getMaxDepth() < depth) return false;
}
}
cdebug_log(149,0) << "AutoSegment::canPivotDown() - true [propagate]" << endl;
return true;
}
Replace "unsigned int" by "Flags" in all AutoSegments collections. * Change: In Anabatic::AutoSegments collections, change the type of all the flags that where in "unsigned int" (32 bits) to Flags (uint64_t) as there is now more than 32 flags for functions. * New: In Ababatic::Constants, added new flag Flags::WithPerpands, which makes the number of flags tip over 32 bits, thus making mandatory to uses Flags and not unsigned int. * New: In Anabatic::AutoSegments_Perpandiculars, manage a new flag Flags::WithDoglegs to allow to propagate through global segments that are connecteds via doglegs on local segments. Meaning that there is a good chance that they could be aligned. Slighly change the way we propagate on aligned segments: no longer check for VTee or HTee, but only for same direction and layer as master. * New: In Anabatic & Katana, replace all the "int", "long" and their variants by the less implementation ambiguous "int32_t", "int64_t" (and variant). This should help to better detect bit trucation in flags. Use the type to give a hint about the flags kind: - Type "Flags", for flags shared among Anabatic & Katana functions/methods (may also appear in some objects states). - Type "uint32_t" for flags belonging to an object internal state of from Hurricane functions flags (those should be grouped in a Flag subclass in a perfect world).
2017-05-16 07:53:33 -05:00
bool AutoSegment::canMoveUp ( float reserve, Flags flags ) const
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
{
Forgot to perform Track re-order after removing zero-length segments. * Change: In Anabatic::Autocontact, replace getMinDepth() and getMaxDepth() by getDepthSpan(). * New: In Anabatic::AutoSegment::canMoveUp(), add an optional check of low up density (Flags::CheckLowUpDensity). Allows to move up a segment if the up density is (very) low, and in this case it's more efficient than breaking it to fit in the lower layer. canMoveUp() is now able to perform the same work as canPivotUp() if *not* supplied the flag Flags::IgnoreContacts. * New: In Katana, in GlobalRouting::DigitalDistance() now take into account the cost of a VIA (currently set to 2.5). Need to known the Net currently routed in the DigitalDistance object itself. * Change: In Katana::Track::Element::canPivotUp(), now also takes a flag parameter. * Change: In Katana::Manipulator, new flag IgnoreContacts to mirror the one of Anabatic. * Change: In Katana::SegmentFsm, allocate once a Manipulator object instead of many times on the fly. In SegmentFsm::_slackenGlobal(), in the slacken state, if the up density is (very) low, bypass to move up instead of slackening. This solve better the routing of the control part of the register file. The register file having a pathological case of terminal placement: many punctual terminals aligneds in METAL2 *and* a grid of METAL2 and METAL3 blockages near below... * Bug: In Katana::Session::_revalidate(), after removing the zero-length segments, forgot to re-order the track, leading to many stranges effects as the indexes where no longer coherent in the Track.
2016-09-20 04:30:45 -05:00
cdebug_log(159,0) << "AutoSegment::canMoveUp() " << flags
<< " (reserve:" << reserve << ") " << this << endl;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
Forgot to perform Track re-order after removing zero-length segments. * Change: In Anabatic::Autocontact, replace getMinDepth() and getMaxDepth() by getDepthSpan(). * New: In Anabatic::AutoSegment::canMoveUp(), add an optional check of low up density (Flags::CheckLowUpDensity). Allows to move up a segment if the up density is (very) low, and in this case it's more efficient than breaking it to fit in the lower layer. canMoveUp() is now able to perform the same work as canPivotUp() if *not* supplied the flag Flags::IgnoreContacts. * New: In Katana, in GlobalRouting::DigitalDistance() now take into account the cost of a VIA (currently set to 2.5). Need to known the Net currently routed in the DigitalDistance object itself. * Change: In Katana::Track::Element::canPivotUp(), now also takes a flag parameter. * Change: In Katana::Manipulator, new flag IgnoreContacts to mirror the one of Anabatic. * Change: In Katana::SegmentFsm, allocate once a Manipulator object instead of many times on the fly. In SegmentFsm::_slackenGlobal(), in the slacken state, if the up density is (very) low, bypass to move up instead of slackening. This solve better the routing of the control part of the register file. The register file having a pathological case of terminal placement: many punctual terminals aligneds in METAL2 *and* a grid of METAL2 and METAL3 blockages near below... * Bug: In Katana::Session::_revalidate(), after removing the zero-length segments, forgot to re-order the track, leading to many stranges effects as the indexes where no longer coherent in the Track.
2016-09-20 04:30:45 -05:00
bool nLowDensity = true;
bool nLowUpDensity = true;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
if ( isLayerChange() or isFixed() or isUnbreakable() ) return false;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
if ( isStrongTerminal() and (not (flags & Flags::AllowTerminal)) ) return false;
if ( isLocal() and (not (flags & Flags::AllowLocal )) ) return false;
size_t depth = Session::getRoutingGauge()->getLayerDepth(getLayer()) + 2;
if (depth > Session::getConfiguration()->getAllowedDepth()) return false;
vector<GCell*> gcells;
getGCells( gcells );
for ( size_t i=0 ; i<gcells.size() ; i++ ) {
Forgot to perform Track re-order after removing zero-length segments. * Change: In Anabatic::Autocontact, replace getMinDepth() and getMaxDepth() by getDepthSpan(). * New: In Anabatic::AutoSegment::canMoveUp(), add an optional check of low up density (Flags::CheckLowUpDensity). Allows to move up a segment if the up density is (very) low, and in this case it's more efficient than breaking it to fit in the lower layer. canMoveUp() is now able to perform the same work as canPivotUp() if *not* supplied the flag Flags::IgnoreContacts. * New: In Katana, in GlobalRouting::DigitalDistance() now take into account the cost of a VIA (currently set to 2.5). Need to known the Net currently routed in the DigitalDistance object itself. * Change: In Katana::Track::Element::canPivotUp(), now also takes a flag parameter. * Change: In Katana::Manipulator, new flag IgnoreContacts to mirror the one of Anabatic. * Change: In Katana::SegmentFsm, allocate once a Manipulator object instead of many times on the fly. In SegmentFsm::_slackenGlobal(), in the slacken state, if the up density is (very) low, bypass to move up instead of slackening. This solve better the routing of the control part of the register file. The register file having a pathological case of terminal placement: many punctual terminals aligneds in METAL2 *and* a grid of METAL2 and METAL3 blockages near below... * Bug: In Katana::Session::_revalidate(), after removing the zero-length segments, forgot to re-order the track, leading to many stranges effects as the indexes where no longer coherent in the Track.
2016-09-20 04:30:45 -05:00
if ( nLowDensity and (gcells[i]->getWDensity(depth-2) > 0.5) ) nLowDensity = false;
if ( nLowUpDensity and (gcells[i]->getWDensity(depth) > 0.2) ) nLowUpDensity = false;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
if (not gcells[i]->hasFreeTrack(depth,reserve)) {
Forgot to perform Track re-order after removing zero-length segments. * Change: In Anabatic::Autocontact, replace getMinDepth() and getMaxDepth() by getDepthSpan(). * New: In Anabatic::AutoSegment::canMoveUp(), add an optional check of low up density (Flags::CheckLowUpDensity). Allows to move up a segment if the up density is (very) low, and in this case it's more efficient than breaking it to fit in the lower layer. canMoveUp() is now able to perform the same work as canPivotUp() if *not* supplied the flag Flags::IgnoreContacts. * New: In Katana, in GlobalRouting::DigitalDistance() now take into account the cost of a VIA (currently set to 2.5). Need to known the Net currently routed in the DigitalDistance object itself. * Change: In Katana::Track::Element::canPivotUp(), now also takes a flag parameter. * Change: In Katana::Manipulator, new flag IgnoreContacts to mirror the one of Anabatic. * Change: In Katana::SegmentFsm, allocate once a Manipulator object instead of many times on the fly. In SegmentFsm::_slackenGlobal(), in the slacken state, if the up density is (very) low, bypass to move up instead of slackening. This solve better the routing of the control part of the register file. The register file having a pathological case of terminal placement: many punctual terminals aligneds in METAL2 *and* a grid of METAL2 and METAL3 blockages near below... * Bug: In Katana::Session::_revalidate(), after removing the zero-length segments, forgot to re-order the track, leading to many stranges effects as the indexes where no longer coherent in the Track.
2016-09-20 04:30:45 -05:00
cdebug_log(159,0) << "Not enough free track in " << gcells[i] << endl;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
return false;
}
}
Forgot to perform Track re-order after removing zero-length segments. * Change: In Anabatic::Autocontact, replace getMinDepth() and getMaxDepth() by getDepthSpan(). * New: In Anabatic::AutoSegment::canMoveUp(), add an optional check of low up density (Flags::CheckLowUpDensity). Allows to move up a segment if the up density is (very) low, and in this case it's more efficient than breaking it to fit in the lower layer. canMoveUp() is now able to perform the same work as canPivotUp() if *not* supplied the flag Flags::IgnoreContacts. * New: In Katana, in GlobalRouting::DigitalDistance() now take into account the cost of a VIA (currently set to 2.5). Need to known the Net currently routed in the DigitalDistance object itself. * Change: In Katana::Track::Element::canPivotUp(), now also takes a flag parameter. * Change: In Katana::Manipulator, new flag IgnoreContacts to mirror the one of Anabatic. * Change: In Katana::SegmentFsm, allocate once a Manipulator object instead of many times on the fly. In SegmentFsm::_slackenGlobal(), in the slacken state, if the up density is (very) low, bypass to move up instead of slackening. This solve better the routing of the control part of the register file. The register file having a pathological case of terminal placement: many punctual terminals aligneds in METAL2 *and* a grid of METAL2 and METAL3 blockages near below... * Bug: In Katana::Session::_revalidate(), after removing the zero-length segments, forgot to re-order the track, leading to many stranges effects as the indexes where no longer coherent in the Track.
2016-09-20 04:30:45 -05:00
cdebug_log(159,0) << "Enough free track under canonical segment." << endl;
if (not (flags & Flags::IgnoreContacts)) {
if (getAutoSource()->getMinDepth() < depth-2) return false;
if (getAutoTarget()->getMinDepth() < depth-2) return false;
}
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
if ( isLocal() and not (flags & Flags::Propagate) ) {
if (not getAutoSource()->canMoveUp(this)) return false;
if (not getAutoTarget()->canMoveUp(this)) return false;
return true;
}
Improvements in routing quality for ARMv2a * Change: In Katana::SegmentFsm::_slackenStrap(), make unbreakable segments pass through the LocalVsGlobal state so when other try to ripup them they are prioritary. Otherwise some could go through Unimplemented directly, without allowing other to attempt to make a detour. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when a global, which is about to be slackened wants to use a track where there is a segment directly connected to a terminal with a significant ripup count, mark it as "AtRipupLimit" so it tries to avoid it. The idea is that globals with high ripup count must avoid terminal segments because it is likely they will be riped up again so they better find another track. This was leading to unsolvable configuration when two segments always want the same track. In this cas, the global would loose. * Bug: In Katana::TrackSegment::canSlacken(), never slacken a segments in non-preferred direction. * Change: In Anabatic::AutoHorizontal::_slacken(), allow slackening of segments which perpandiculars are in non-preferred direction, and not only directly attached to terminals. * Change: In Anabatic::AutoSegment::canMoveUp(), re-allow segments which perpandiculars are in non-preferred direction to be moved up. * Bug: In Katana::Manipulator::moveUp(), when moving up, do not forget to ripup and reschedule said segment.
2019-08-23 08:43:13 -05:00
// if (getAutoSource()->isTurn() and (getAutoSource()->getPerpandicular(this)->getLayer() == getLayer())) return false;
// if (getAutoTarget()->isTurn() and (getAutoTarget()->getPerpandicular(this)->getLayer() == getLayer())) return false;
Forgot to perform Track re-order after removing zero-length segments. * Change: In Anabatic::Autocontact, replace getMinDepth() and getMaxDepth() by getDepthSpan(). * New: In Anabatic::AutoSegment::canMoveUp(), add an optional check of low up density (Flags::CheckLowUpDensity). Allows to move up a segment if the up density is (very) low, and in this case it's more efficient than breaking it to fit in the lower layer. canMoveUp() is now able to perform the same work as canPivotUp() if *not* supplied the flag Flags::IgnoreContacts. * New: In Katana, in GlobalRouting::DigitalDistance() now take into account the cost of a VIA (currently set to 2.5). Need to known the Net currently routed in the DigitalDistance object itself. * Change: In Katana::Track::Element::canPivotUp(), now also takes a flag parameter. * Change: In Katana::Manipulator, new flag IgnoreContacts to mirror the one of Anabatic. * Change: In Katana::SegmentFsm, allocate once a Manipulator object instead of many times on the fly. In SegmentFsm::_slackenGlobal(), in the slacken state, if the up density is (very) low, bypass to move up instead of slackening. This solve better the routing of the control part of the register file. The register file having a pathological case of terminal placement: many punctual terminals aligneds in METAL2 *and* a grid of METAL2 and METAL3 blockages near below... * Bug: In Katana::Session::_revalidate(), after removing the zero-length segments, forgot to re-order the track, leading to many stranges effects as the indexes where no longer coherent in the Track.
2016-09-20 04:30:45 -05:00
cdebug_log(159,0) << "Both source & target Contacts can move up." << endl;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
//bool hasGlobalSegment = false;
if ((flags & Flags::Propagate) and not isNotAligned()) {
Forgot to perform Track re-order after removing zero-length segments. * Change: In Anabatic::Autocontact, replace getMinDepth() and getMaxDepth() by getDepthSpan(). * New: In Anabatic::AutoSegment::canMoveUp(), add an optional check of low up density (Flags::CheckLowUpDensity). Allows to move up a segment if the up density is (very) low, and in this case it's more efficient than breaking it to fit in the lower layer. canMoveUp() is now able to perform the same work as canPivotUp() if *not* supplied the flag Flags::IgnoreContacts. * New: In Katana, in GlobalRouting::DigitalDistance() now take into account the cost of a VIA (currently set to 2.5). Need to known the Net currently routed in the DigitalDistance object itself. * Change: In Katana::Track::Element::canPivotUp(), now also takes a flag parameter. * Change: In Katana::Manipulator, new flag IgnoreContacts to mirror the one of Anabatic. * Change: In Katana::SegmentFsm, allocate once a Manipulator object instead of many times on the fly. In SegmentFsm::_slackenGlobal(), in the slacken state, if the up density is (very) low, bypass to move up instead of slackening. This solve better the routing of the control part of the register file. The register file having a pathological case of terminal placement: many punctual terminals aligneds in METAL2 *and* a grid of METAL2 and METAL3 blockages near below... * Bug: In Katana::Session::_revalidate(), after removing the zero-length segments, forgot to re-order the track, leading to many stranges effects as the indexes where no longer coherent in the Track.
2016-09-20 04:30:45 -05:00
for ( AutoSegment* segment : const_cast<AutoSegment*>(this)->getAligneds(flags) ) {
if (segment->isFixed ()) return false;
//if (segment->isGlobal()) hasGlobalSegment = true;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
Forgot to perform Track re-order after removing zero-length segments. * Change: In Anabatic::Autocontact, replace getMinDepth() and getMaxDepth() by getDepthSpan(). * New: In Anabatic::AutoSegment::canMoveUp(), add an optional check of low up density (Flags::CheckLowUpDensity). Allows to move up a segment if the up density is (very) low, and in this case it's more efficient than breaking it to fit in the lower layer. canMoveUp() is now able to perform the same work as canPivotUp() if *not* supplied the flag Flags::IgnoreContacts. * New: In Katana, in GlobalRouting::DigitalDistance() now take into account the cost of a VIA (currently set to 2.5). Need to known the Net currently routed in the DigitalDistance object itself. * Change: In Katana::Track::Element::canPivotUp(), now also takes a flag parameter. * Change: In Katana::Manipulator, new flag IgnoreContacts to mirror the one of Anabatic. * Change: In Katana::SegmentFsm, allocate once a Manipulator object instead of many times on the fly. In SegmentFsm::_slackenGlobal(), in the slacken state, if the up density is (very) low, bypass to move up instead of slackening. This solve better the routing of the control part of the register file. The register file having a pathological case of terminal placement: many punctual terminals aligneds in METAL2 *and* a grid of METAL2 and METAL3 blockages near below... * Bug: In Katana::Session::_revalidate(), after removing the zero-length segments, forgot to re-order the track, leading to many stranges effects as the indexes where no longer coherent in the Track.
2016-09-20 04:30:45 -05:00
if (not (flags & Flags::IgnoreContacts)) {
if (segment->getAutoSource()->getMinDepth() < depth-2) return false;
if (segment->getAutoTarget()->getMinDepth() < depth-2) return false;
}
segment->getGCells( gcells );
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
for ( size_t i=0 ; i<gcells.size() ; i++ ) {
Forgot to perform Track re-order after removing zero-length segments. * Change: In Anabatic::Autocontact, replace getMinDepth() and getMaxDepth() by getDepthSpan(). * New: In Anabatic::AutoSegment::canMoveUp(), add an optional check of low up density (Flags::CheckLowUpDensity). Allows to move up a segment if the up density is (very) low, and in this case it's more efficient than breaking it to fit in the lower layer. canMoveUp() is now able to perform the same work as canPivotUp() if *not* supplied the flag Flags::IgnoreContacts. * New: In Katana, in GlobalRouting::DigitalDistance() now take into account the cost of a VIA (currently set to 2.5). Need to known the Net currently routed in the DigitalDistance object itself. * Change: In Katana::Track::Element::canPivotUp(), now also takes a flag parameter. * Change: In Katana::Manipulator, new flag IgnoreContacts to mirror the one of Anabatic. * Change: In Katana::SegmentFsm, allocate once a Manipulator object instead of many times on the fly. In SegmentFsm::_slackenGlobal(), in the slacken state, if the up density is (very) low, bypass to move up instead of slackening. This solve better the routing of the control part of the register file. The register file having a pathological case of terminal placement: many punctual terminals aligneds in METAL2 *and* a grid of METAL2 and METAL3 blockages near below... * Bug: In Katana::Session::_revalidate(), after removing the zero-length segments, forgot to re-order the track, leading to many stranges effects as the indexes where no longer coherent in the Track.
2016-09-20 04:30:45 -05:00
if ( nLowDensity and (gcells[i]->getWDensity(depth-2) > 0.6) ) nLowDensity = false;
if ( nLowUpDensity and (gcells[i]->getWDensity(depth) > 0.2) ) {
cdebug_log(159,0) << "lowUpDensity false in " << gcells[i]
<< "d:" << gcells[i]->getWDensity(depth) << endl;
nLowUpDensity = false;
}
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
if (not gcells[i]->hasFreeTrack(depth,reserve)) {
Forgot to perform Track re-order after removing zero-length segments. * Change: In Anabatic::Autocontact, replace getMinDepth() and getMaxDepth() by getDepthSpan(). * New: In Anabatic::AutoSegment::canMoveUp(), add an optional check of low up density (Flags::CheckLowUpDensity). Allows to move up a segment if the up density is (very) low, and in this case it's more efficient than breaking it to fit in the lower layer. canMoveUp() is now able to perform the same work as canPivotUp() if *not* supplied the flag Flags::IgnoreContacts. * New: In Katana, in GlobalRouting::DigitalDistance() now take into account the cost of a VIA (currently set to 2.5). Need to known the Net currently routed in the DigitalDistance object itself. * Change: In Katana::Track::Element::canPivotUp(), now also takes a flag parameter. * Change: In Katana::Manipulator, new flag IgnoreContacts to mirror the one of Anabatic. * Change: In Katana::SegmentFsm, allocate once a Manipulator object instead of many times on the fly. In SegmentFsm::_slackenGlobal(), in the slacken state, if the up density is (very) low, bypass to move up instead of slackening. This solve better the routing of the control part of the register file. The register file having a pathological case of terminal placement: many punctual terminals aligneds in METAL2 *and* a grid of METAL2 and METAL3 blockages near below... * Bug: In Katana::Session::_revalidate(), after removing the zero-length segments, forgot to re-order the track, leading to many stranges effects as the indexes where no longer coherent in the Track.
2016-09-20 04:30:45 -05:00
cdebug_log(159,0) << "Not enough free track in " << gcells[i] << endl;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
return false;
}
}
}
}
Forgot to perform Track re-order after removing zero-length segments. * Change: In Anabatic::Autocontact, replace getMinDepth() and getMaxDepth() by getDepthSpan(). * New: In Anabatic::AutoSegment::canMoveUp(), add an optional check of low up density (Flags::CheckLowUpDensity). Allows to move up a segment if the up density is (very) low, and in this case it's more efficient than breaking it to fit in the lower layer. canMoveUp() is now able to perform the same work as canPivotUp() if *not* supplied the flag Flags::IgnoreContacts. * New: In Katana, in GlobalRouting::DigitalDistance() now take into account the cost of a VIA (currently set to 2.5). Need to known the Net currently routed in the DigitalDistance object itself. * Change: In Katana::Track::Element::canPivotUp(), now also takes a flag parameter. * Change: In Katana::Manipulator, new flag IgnoreContacts to mirror the one of Anabatic. * Change: In Katana::SegmentFsm, allocate once a Manipulator object instead of many times on the fly. In SegmentFsm::_slackenGlobal(), in the slacken state, if the up density is (very) low, bypass to move up instead of slackening. This solve better the routing of the control part of the register file. The register file having a pathological case of terminal placement: many punctual terminals aligneds in METAL2 *and* a grid of METAL2 and METAL3 blockages near below... * Bug: In Katana::Session::_revalidate(), after removing the zero-length segments, forgot to re-order the track, leading to many stranges effects as the indexes where no longer coherent in the Track.
2016-09-20 04:30:45 -05:00
if ( nLowDensity and (flags & Flags::CheckLowDensity )) return false;
if (not nLowUpDensity and (flags & Flags::CheckLowUpDensity)) return false;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
if ( (depth >= 4) and (flags & Flags::WithPerpands) ) {
Forgot to perform Track re-order after removing zero-length segments. * Change: In Anabatic::Autocontact, replace getMinDepth() and getMaxDepth() by getDepthSpan(). * New: In Anabatic::AutoSegment::canMoveUp(), add an optional check of low up density (Flags::CheckLowUpDensity). Allows to move up a segment if the up density is (very) low, and in this case it's more efficient than breaking it to fit in the lower layer. canMoveUp() is now able to perform the same work as canPivotUp() if *not* supplied the flag Flags::IgnoreContacts. * New: In Katana, in GlobalRouting::DigitalDistance() now take into account the cost of a VIA (currently set to 2.5). Need to known the Net currently routed in the DigitalDistance object itself. * Change: In Katana::Track::Element::canPivotUp(), now also takes a flag parameter. * Change: In Katana::Manipulator, new flag IgnoreContacts to mirror the one of Anabatic. * Change: In Katana::SegmentFsm, allocate once a Manipulator object instead of many times on the fly. In SegmentFsm::_slackenGlobal(), in the slacken state, if the up density is (very) low, bypass to move up instead of slackening. This solve better the routing of the control part of the register file. The register file having a pathological case of terminal placement: many punctual terminals aligneds in METAL2 *and* a grid of METAL2 and METAL3 blockages near below... * Bug: In Katana::Session::_revalidate(), after removing the zero-length segments, forgot to re-order the track, leading to many stranges effects as the indexes where no longer coherent in the Track.
2016-09-20 04:30:45 -05:00
float fragmentation = (*gcells.begin())->getFragmentation( depth-1 );
cdebug_log(159,0) << "Check begin GCell perpandicular fragmentation: " << fragmentation << endl;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
if (fragmentation < 0.5) {
Forgot to perform Track re-order after removing zero-length segments. * Change: In Anabatic::Autocontact, replace getMinDepth() and getMaxDepth() by getDepthSpan(). * New: In Anabatic::AutoSegment::canMoveUp(), add an optional check of low up density (Flags::CheckLowUpDensity). Allows to move up a segment if the up density is (very) low, and in this case it's more efficient than breaking it to fit in the lower layer. canMoveUp() is now able to perform the same work as canPivotUp() if *not* supplied the flag Flags::IgnoreContacts. * New: In Katana, in GlobalRouting::DigitalDistance() now take into account the cost of a VIA (currently set to 2.5). Need to known the Net currently routed in the DigitalDistance object itself. * Change: In Katana::Track::Element::canPivotUp(), now also takes a flag parameter. * Change: In Katana::Manipulator, new flag IgnoreContacts to mirror the one of Anabatic. * Change: In Katana::SegmentFsm, allocate once a Manipulator object instead of many times on the fly. In SegmentFsm::_slackenGlobal(), in the slacken state, if the up density is (very) low, bypass to move up instead of slackening. This solve better the routing of the control part of the register file. The register file having a pathological case of terminal placement: many punctual terminals aligneds in METAL2 *and* a grid of METAL2 and METAL3 blockages near below... * Bug: In Katana::Session::_revalidate(), after removing the zero-length segments, forgot to re-order the track, leading to many stranges effects as the indexes where no longer coherent in the Track.
2016-09-20 04:30:45 -05:00
cdebug_log(159,0) << "Not enough free track for perpandicular in begin GCell "
<< "(frag:" << fragmentation << ")."
<< endl;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
return false;
}
Forgot to perform Track re-order after removing zero-length segments. * Change: In Anabatic::Autocontact, replace getMinDepth() and getMaxDepth() by getDepthSpan(). * New: In Anabatic::AutoSegment::canMoveUp(), add an optional check of low up density (Flags::CheckLowUpDensity). Allows to move up a segment if the up density is (very) low, and in this case it's more efficient than breaking it to fit in the lower layer. canMoveUp() is now able to perform the same work as canPivotUp() if *not* supplied the flag Flags::IgnoreContacts. * New: In Katana, in GlobalRouting::DigitalDistance() now take into account the cost of a VIA (currently set to 2.5). Need to known the Net currently routed in the DigitalDistance object itself. * Change: In Katana::Track::Element::canPivotUp(), now also takes a flag parameter. * Change: In Katana::Manipulator, new flag IgnoreContacts to mirror the one of Anabatic. * Change: In Katana::SegmentFsm, allocate once a Manipulator object instead of many times on the fly. In SegmentFsm::_slackenGlobal(), in the slacken state, if the up density is (very) low, bypass to move up instead of slackening. This solve better the routing of the control part of the register file. The register file having a pathological case of terminal placement: many punctual terminals aligneds in METAL2 *and* a grid of METAL2 and METAL3 blockages near below... * Bug: In Katana::Session::_revalidate(), after removing the zero-length segments, forgot to re-order the track, leading to many stranges effects as the indexes where no longer coherent in the Track.
2016-09-20 04:30:45 -05:00
fragmentation = (*gcells.rbegin())->getFragmentation( depth-1 );
cdebug_log(159,0) << "Check end GCell perpandicular fragmentation: " << fragmentation << endl;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
if (fragmentation < 0.5) {
Forgot to perform Track re-order after removing zero-length segments. * Change: In Anabatic::Autocontact, replace getMinDepth() and getMaxDepth() by getDepthSpan(). * New: In Anabatic::AutoSegment::canMoveUp(), add an optional check of low up density (Flags::CheckLowUpDensity). Allows to move up a segment if the up density is (very) low, and in this case it's more efficient than breaking it to fit in the lower layer. canMoveUp() is now able to perform the same work as canPivotUp() if *not* supplied the flag Flags::IgnoreContacts. * New: In Katana, in GlobalRouting::DigitalDistance() now take into account the cost of a VIA (currently set to 2.5). Need to known the Net currently routed in the DigitalDistance object itself. * Change: In Katana::Track::Element::canPivotUp(), now also takes a flag parameter. * Change: In Katana::Manipulator, new flag IgnoreContacts to mirror the one of Anabatic. * Change: In Katana::SegmentFsm, allocate once a Manipulator object instead of many times on the fly. In SegmentFsm::_slackenGlobal(), in the slacken state, if the up density is (very) low, bypass to move up instead of slackening. This solve better the routing of the control part of the register file. The register file having a pathological case of terminal placement: many punctual terminals aligneds in METAL2 *and* a grid of METAL2 and METAL3 blockages near below... * Bug: In Katana::Session::_revalidate(), after removing the zero-length segments, forgot to re-order the track, leading to many stranges effects as the indexes where no longer coherent in the Track.
2016-09-20 04:30:45 -05:00
cdebug_log(159,0) << "Not enough free track for perpandicular in end GCell "
<< "(frag:" << fragmentation << ")."
<< endl;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
return false;
}
}
return true;
}
Replace "unsigned int" by "Flags" in all AutoSegments collections. * Change: In Anabatic::AutoSegments collections, change the type of all the flags that where in "unsigned int" (32 bits) to Flags (uint64_t) as there is now more than 32 flags for functions. * New: In Ababatic::Constants, added new flag Flags::WithPerpands, which makes the number of flags tip over 32 bits, thus making mandatory to uses Flags and not unsigned int. * New: In Anabatic::AutoSegments_Perpandiculars, manage a new flag Flags::WithDoglegs to allow to propagate through global segments that are connecteds via doglegs on local segments. Meaning that there is a good chance that they could be aligned. Slighly change the way we propagate on aligned segments: no longer check for VTee or HTee, but only for same direction and layer as master. * New: In Anabatic & Katana, replace all the "int", "long" and their variants by the less implementation ambiguous "int32_t", "int64_t" (and variant). This should help to better detect bit trucation in flags. Use the type to give a hint about the flags kind: - Type "Flags", for flags shared among Anabatic & Katana functions/methods (may also appear in some objects states). - Type "uint32_t" for flags belonging to an object internal state of from Hurricane functions flags (those should be grouped in a Flag subclass in a perfect world).
2017-05-16 07:53:33 -05:00
bool AutoSegment::moveUp ( Flags flags )
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
{
//if ( not canMoveUp(0.0,flags) ) return false;
changeDepth( Session::getRoutingGauge()->getLayerDepth(getLayer()) + 2, flags&Flags::Propagate );
return true;
}
Replace "unsigned int" by "Flags" in all AutoSegments collections. * Change: In Anabatic::AutoSegments collections, change the type of all the flags that where in "unsigned int" (32 bits) to Flags (uint64_t) as there is now more than 32 flags for functions. * New: In Ababatic::Constants, added new flag Flags::WithPerpands, which makes the number of flags tip over 32 bits, thus making mandatory to uses Flags and not unsigned int. * New: In Anabatic::AutoSegments_Perpandiculars, manage a new flag Flags::WithDoglegs to allow to propagate through global segments that are connecteds via doglegs on local segments. Meaning that there is a good chance that they could be aligned. Slighly change the way we propagate on aligned segments: no longer check for VTee or HTee, but only for same direction and layer as master. * New: In Anabatic & Katana, replace all the "int", "long" and their variants by the less implementation ambiguous "int32_t", "int64_t" (and variant). This should help to better detect bit trucation in flags. Use the type to give a hint about the flags kind: - Type "Flags", for flags shared among Anabatic & Katana functions/methods (may also appear in some objects states). - Type "uint32_t" for flags belonging to an object internal state of from Hurricane functions flags (those should be grouped in a Flag subclass in a perfect world).
2017-05-16 07:53:33 -05:00
bool AutoSegment::moveDown ( Flags flags )
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
{
//if ( not canPivotDown(0.0,flags) ) return false;
changeDepth( Session::getRoutingGauge()->getLayerDepth(getLayer()) - 2, flags&Flags::Propagate );
return true;
}
bool AutoSegment::reduceDoglegLayer ()
{
Manage minimum area for VIAs in the P&R. * New: In Hurricane::Layer, add support for a minimal area (given in microns). Exported in the Python interface. * New: In Anabatic::AutoSegment::bloatStackedStrap(), method to enlarge too small vertically stacked VIAs. For now just create a square ensuring the minimal area. Potentially cause DRC error in some cases so needs to be refined. * New: In Anabatic::_gutAnabatic(), check for too little stacked VIAs and display a report.
2020-11-15 17:55:49 -06:00
if (not isReduced()) return false;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
DebugSession::open( getNet(), 149, 160 );
cdebug_log(159,1) << "AutoSegment::reduceDoglegLayer(): " << this << endl;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
AutoContact* source = getAutoSource();
AutoContact* target = getAutoTarget();
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
unsigned int minSourceDepth = Session::getAllowedDepth();
unsigned int maxSourceDepth = 0;
unsigned int minTargetDepth = Session::getAllowedDepth();
unsigned int maxTargetDepth = 0;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
if (source->isTerminal()) {
unsigned int anchorDepth = Session::getLayerDepth( source->base()->getAnchor()->getLayer() );
minSourceDepth = std::min( minSourceDepth, anchorDepth );
maxSourceDepth = std::max( maxSourceDepth, anchorDepth );
} else {
for ( AutoSegment* perpandicular : source->getAutoSegments() ) {
if (perpandicular == this) continue;
minSourceDepth = std::min( minSourceDepth, perpandicular->getDepth() );
maxSourceDepth = std::max( maxSourceDepth, perpandicular->getDepth() );
}
}
if (target->isTerminal()) {
unsigned int anchorDepth = Session::getLayerDepth( target->base()->getAnchor()->getLayer() );
minTargetDepth = std::min( minTargetDepth, anchorDepth );
maxTargetDepth = std::max( maxTargetDepth, anchorDepth );
} else {
for ( AutoSegment* perpandicular : target->getAutoSegments() ) {
if (perpandicular == this) continue;
minTargetDepth = std::min( minTargetDepth, perpandicular->getDepth() );
maxTargetDepth = std::max( maxTargetDepth, perpandicular->getDepth() );
}
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
}
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
cdebug_log(159,0) << "Source span: [" << minSourceDepth << " " << maxSourceDepth << "]" << endl;
cdebug_log(159,0) << "Target span: [" << minTargetDepth << " " << maxTargetDepth << "]" << endl;
if ( (minSourceDepth == maxSourceDepth)
and (minTargetDepth == maxTargetDepth)
Bug in segment layer reduction causing METAL2 to go in METAL1 (RISC-V). * Bug: In Anabatic::AutoSegment::reduceDoglegLayer(), we did not check that not only source and target depth span was zero, but also both on the same one. This was seen in the RISC-V where a METAL2 small segment was demoted into METAL1 causing a gap in the routing (and a short in the underlying standart cell).
2019-08-18 06:20:50 -05:00
and (minSourceDepth == minTargetDepth) ) {
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
const Layer* layer = Session::getRoutingLayer(minSourceDepth);
DbU::Unit side = Session::getWireWidth (minSourceDepth);
cdebug_log(159,0) << "Reducing to " << minSourceDepth << " " << layer << endl;
In Anabatic/Katana, add support for VH gauges (real technos). * Change: In Anabatic::AutoContactTerminal::getNativeConstraintBox(), when the anchor is a RoutingPad (which must be always the case), perform the true computation of it's position based on the segment occurrence. It is a important change, previously the area was in fact the "center line" of the connector while now it is really an area (mandatory for "half-offgrid" terminals of real technologies). The change is not complete yet, the area should be shrinked by the half size of a VIA, because the area applies to the center coordinate of the VIA (to be done quickly). * Bug: In Anabatic::AutoContactTurn::updateTopology(), when a dogleg is created (restore connexity after a layer change) the layer of the VIA, based on the segments it connects to must be re-computed *after* the dogleg has been made. * Change: In all files of Anabatic, when comparing two layers, no longer use the Layer pointer itself, but the layer mask. This allow a transparent management of both real and symbolic layers (which do share the same mask). Real metal layers (not VIAs) will be BasicLayer and symbolic metal layers will be RegularLayer. * New: Anabatic::Configuration::selectRpComponent(), select the best RoutingPad component for metal1 terminals. Look for the metal1 component with the biggest accessibility on-grid. RoutingPad using other metals are left untoucheds. * New: New function Anabatic::Vertex::getNeighbor(Edge*) to get the neighbor Vertex through an Edge*. This method allows to write clearer code as we no longer need to access the neighbor through the underlying GCell. Also add proxies for GCell methods in Vertex. * Bug: In Anabatic::Dijkstra::_toSources(), in the ripup stage, when a component with multiples vertexes is reached *and* two of it's vertexes are reached *at the same time* (one from which we backtrack and one still in the queue) extraneous edges may be created by _materialize(). Case occurs on snx/c35b4, "abc_5360_n903_1". To solve this, Dijkstra::_toSource() is modificated, the "from" edges of the newly reacheds vertexes are reset to NULL, *except* for the one we will be backtracking from. That is, the one given in the source argument. * Change: In Anabatic::NetBuilder class, put the various Hooks and RoutingPad sorting functions as class ones. * Bug: In AutoSegment::setLayer(), raise the SegInvalidatedFayer flag. This unset flag was causing AutoContactTurn::updateTopology() to not work as expected and making gaps, this was the cause of the last remaining warnings about layer connexity.
2018-01-06 09:55:53 -06:00
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
source->setLayer( layer );
target->setLayer( layer );
setLayer( layer );
source->setSizes( side, side );
target->setSizes( side, side );
}
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
cdebug_tabw(159,-1);
DebugSession::close();
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
return true;
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
// if (not source->isTurn() or not target->isTurn()) return true;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
// unsigned int perpandicularDepth = getDepth();
// if (isSpinBottom()) --perpandicularDepth;
// if (isSpinTop ()) ++perpandicularDepth;
// if (perpandicularDepth == getDepth()) {
// cerr << Bug( "AutoSegment::reduceDoglegLayer(): Reduced segment spin is neither top (TT) nor bottom (BB).\n"
// " %s"
// , getString(this).c_str() ) << endl;
// return false;
// }
// const Layer* layer = Session::getRoutingLayer(perpandicularDepth);
// DbU::Unit side = Session::getWireWidth (perpandicularDepth);
// source->setLayer( layer );
// target->setLayer( layer );
// setLayer( layer );
// source->setSizes( side, side );
// target->setSizes( side, side );
// return true;
}
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
Manage minimum area for VIAs in the P&R. * New: In Hurricane::Layer, add support for a minimal area (given in microns). Exported in the Python interface. * New: In Anabatic::AutoSegment::bloatStackedStrap(), method to enlarge too small vertically stacked VIAs. For now just create a square ensuring the minimal area. Potentially cause DRC error in some cases so needs to be refined. * New: In Anabatic::_gutAnabatic(), check for too little stacked VIAs and display a report.
2020-11-15 17:55:49 -06:00
bool AutoSegment::bloatStackedStrap ()
{
DRC correct on Arlet6505 / TSMC C180. Integrate new features and bug fixes so the Arlet 6502 benchs successfully passes real DRC with reference industrial tools. Short summary: * Manage minimum area for VIAs in Katana::Tracks. * Allow different wire width for wires perpandicular to the prefered routing direction. * StackedVIAs used in the clock tree no longer assume an uniform routing grid (same offset & pitch all the way up). * Some hard-coded patches in PowerRails for FlexLib. * New: In CRL/symbolic/cmos/kite.py & cmos45/kite.py, update the RoutingLayerGauges by adding the new PWireWidth parameter. Always zero in case of symbolic layout (too fine tuning). * New: In CRL::RoutingGauge, add accessor to PWireWidth parameter. Modify the clone method. * New: In CRL::RoutingLayerGauge, add new parameter "PWireWidth" to give the width of a wire when it not drawn in the prefered routing direction. If it is set to zero, the normal width is used. * New: In CRL::PyRoutingGauge, export the updated constructor interface. It is *not* backward compatible, one must add the PWireWidth parameter in the various kite.py configuration files (in etc/). * Change: In AnabaticEngine::_gutAnabatic(), disable the minimum area detection mechanism, replaced by a more complete one in Katana::Track. Left commented out for now, but will be removed in the future. * Change: In Anabatic::AutoContact::updateLayer(), now systematically calls setLayerAndWidth() to potentially resize the VIAs. This is needed in real mode as VIAs are *not* macro-generated but have their real final size. * Change: In Anabatic::AutoContact::setLayerAndWidth(), select the width and height of the contact using the gauge wire width *and* perpandicular *wire width*. * Change: In Anabatic::AutoSegment::_initialize(), the "VIA to same cap" to PWireWidth/2, this will be the size of the VIA in the non-preferred direction at the end cap (non-square in real mode). * Change: In Anabatic::AutoSegment::getExtensionCap(), makes different cases for symbolic and real. Use raw length in real, add half the wire width in symbolic. Add a flag to get the extension cap *only*, not increased of half the minimal spacing. * Change: In Anabatic::AutoSegment::bloatStackedStrap(), enhanced, but finally unused... * New: In Anabatic::AutoSegment::create(), use the PWireWidth when the segment is not in the preferred routing direction (and of minimal width). * New: In Anabatic::Configuration, add new getPWirewidth(), DPHorizontalWidth() and DPVerticalWidth() accessors. * Change: In AnabaticEngine::setupPreRouteds(), skip components in in "cut" material. We are only interested in objects containing some metal (happens in real mode when VIAs cuts are really there). * New: In Katana::PowerRailsPlanes::Rail::doLayout(), add an hard-coded patch that artificially enlarge the *wide wire* so the spacing for wide wire is enforced. For now, two pitches on each side for "FlexLib" gauge. * New: In Katana::Track, add support to find and correct small wire chunks so they respect the minimum area rules. Two helper functions: * ::hasSameLayerTurn(), to find if a a TrackElement as non-zero length perpandicular is same layer connected to it. * ::toFoundryGrid(), to ensure that all coordinates will be on the foundry grid (may move in a more shared location). * ::expandToMinArea(), try to expand, *in the routing direction* the too small wire so it respect the minimal area. Check for the free space in the track. Track::minExpandArea() go through all the TrackElements in the track to look for too small ones and correct them. * Change: In Katana::RoutingPlane, add an accessor to get the tracks. * New: In KatanaEngine::finalizeLayout(), add a post-treatment to find for minimal area violations. * Change: In cumulus/plugins.block.configuration.GaugeConf, add a routingBb attribute that will serve as a common reference to all the functions calculation track positions. We must not have two different reference for the core and the corona. The reference is always the corona when we working on a complete chip. * New: In cumulus/plugins.block.configuration.GaugeConf.getTrack(), Simplified and more reliable way of getting tracks positions. Use the routingBb. * New: In cumulus/plugins.block.configuration.GaugeConf.rpAccess(), Make use of getTrack() to get every metal strap on the right X/Y position. * New: In cumulus/plugins.block.configuration.GaugeConf.expandMinArea(), As those wires are left alone by the router, it is our responsability to abide by the minimal area rule here. Hence the code duplication from the router (bad). Mainly wires made for the clock tree, I mean. * Bug: In cumulus/plugins.chip.configuration.ChipConf.setupICore(), the core instance must be placed on the GCell grid, defined by the slice height (X *and* Y). * Bug: In cumulus/plugins.chip.corona.Builder(), forgot to use bigvia for the corners of the inner ring. * Bug: In cumulus/plugins.chip.pads.corona._createCoreWire(), hard-coded patch for LibreSOCIO, the power/ground connectors toward the core are too wide and can create DRC errors when put side by side. Shrink them by the minimal distance.
2020-11-23 16:07:15 -06:00
DebugSession::open( getNet(), 145, 150 );
cdebug_log(149,1) << "AutoSegment::bloatStackedStrap() " << this << endl;
Manage minimum area for VIAs in the P&R. * New: In Hurricane::Layer, add support for a minimal area (given in microns). Exported in the Python interface. * New: In Anabatic::AutoSegment::bloatStackedStrap(), method to enlarge too small vertically stacked VIAs. For now just create a square ensuring the minimal area. Potentially cause DRC error in some cases so needs to be refined. * New: In Anabatic::_gutAnabatic(), check for too little stacked VIAs and display a report.
2020-11-15 17:55:49 -06:00
double minArea = getLayer()->getMinimalArea();
DRC correct on Arlet6505 / TSMC C180. Integrate new features and bug fixes so the Arlet 6502 benchs successfully passes real DRC with reference industrial tools. Short summary: * Manage minimum area for VIAs in Katana::Tracks. * Allow different wire width for wires perpandicular to the prefered routing direction. * StackedVIAs used in the clock tree no longer assume an uniform routing grid (same offset & pitch all the way up). * Some hard-coded patches in PowerRails for FlexLib. * New: In CRL/symbolic/cmos/kite.py & cmos45/kite.py, update the RoutingLayerGauges by adding the new PWireWidth parameter. Always zero in case of symbolic layout (too fine tuning). * New: In CRL::RoutingGauge, add accessor to PWireWidth parameter. Modify the clone method. * New: In CRL::RoutingLayerGauge, add new parameter "PWireWidth" to give the width of a wire when it not drawn in the prefered routing direction. If it is set to zero, the normal width is used. * New: In CRL::PyRoutingGauge, export the updated constructor interface. It is *not* backward compatible, one must add the PWireWidth parameter in the various kite.py configuration files (in etc/). * Change: In AnabaticEngine::_gutAnabatic(), disable the minimum area detection mechanism, replaced by a more complete one in Katana::Track. Left commented out for now, but will be removed in the future. * Change: In Anabatic::AutoContact::updateLayer(), now systematically calls setLayerAndWidth() to potentially resize the VIAs. This is needed in real mode as VIAs are *not* macro-generated but have their real final size. * Change: In Anabatic::AutoContact::setLayerAndWidth(), select the width and height of the contact using the gauge wire width *and* perpandicular *wire width*. * Change: In Anabatic::AutoSegment::_initialize(), the "VIA to same cap" to PWireWidth/2, this will be the size of the VIA in the non-preferred direction at the end cap (non-square in real mode). * Change: In Anabatic::AutoSegment::getExtensionCap(), makes different cases for symbolic and real. Use raw length in real, add half the wire width in symbolic. Add a flag to get the extension cap *only*, not increased of half the minimal spacing. * Change: In Anabatic::AutoSegment::bloatStackedStrap(), enhanced, but finally unused... * New: In Anabatic::AutoSegment::create(), use the PWireWidth when the segment is not in the preferred routing direction (and of minimal width). * New: In Anabatic::Configuration, add new getPWirewidth(), DPHorizontalWidth() and DPVerticalWidth() accessors. * Change: In AnabaticEngine::setupPreRouteds(), skip components in in "cut" material. We are only interested in objects containing some metal (happens in real mode when VIAs cuts are really there). * New: In Katana::PowerRailsPlanes::Rail::doLayout(), add an hard-coded patch that artificially enlarge the *wide wire* so the spacing for wide wire is enforced. For now, two pitches on each side for "FlexLib" gauge. * New: In Katana::Track, add support to find and correct small wire chunks so they respect the minimum area rules. Two helper functions: * ::hasSameLayerTurn(), to find if a a TrackElement as non-zero length perpandicular is same layer connected to it. * ::toFoundryGrid(), to ensure that all coordinates will be on the foundry grid (may move in a more shared location). * ::expandToMinArea(), try to expand, *in the routing direction* the too small wire so it respect the minimal area. Check for the free space in the track. Track::minExpandArea() go through all the TrackElements in the track to look for too small ones and correct them. * Change: In Katana::RoutingPlane, add an accessor to get the tracks. * New: In KatanaEngine::finalizeLayout(), add a post-treatment to find for minimal area violations. * Change: In cumulus/plugins.block.configuration.GaugeConf, add a routingBb attribute that will serve as a common reference to all the functions calculation track positions. We must not have two different reference for the core and the corona. The reference is always the corona when we working on a complete chip. * New: In cumulus/plugins.block.configuration.GaugeConf.getTrack(), Simplified and more reliable way of getting tracks positions. Use the routingBb. * New: In cumulus/plugins.block.configuration.GaugeConf.rpAccess(), Make use of getTrack() to get every metal strap on the right X/Y position. * New: In cumulus/plugins.block.configuration.GaugeConf.expandMinArea(), As those wires are left alone by the router, it is our responsability to abide by the minimal area rule here. Hence the code duplication from the router (bad). Mainly wires made for the clock tree, I mean. * Bug: In cumulus/plugins.chip.configuration.ChipConf.setupICore(), the core instance must be placed on the GCell grid, defined by the slice height (X *and* Y). * Bug: In cumulus/plugins.chip.corona.Builder(), forgot to use bigvia for the corners of the inner ring. * Bug: In cumulus/plugins.chip.pads.corona._createCoreWire(), hard-coded patch for LibreSOCIO, the power/ground connectors toward the core are too wide and can create DRC errors when put side by side. Shrink them by the minimal distance.
2020-11-23 16:07:15 -06:00
if (minArea == 0.0) {
cdebug_log(149,-1) << "False, NO minimal area." << endl;
DebugSession::close();
return false;
}
DbU::Unit minLength
= DbU::fromPhysical( minArea / DbU::toPhysical( getWidth(), DbU::UnitPower::Micro )
, DbU::UnitPower::Micro );
cdebug_log(149,0) << "Min length: " << DbU::getValueString(minLength) << " ." << endl;
if ((getSpanLength() >= minLength) or isReduced()) {
cdebug_log(149,-1) << "False, has length or is reduced." << endl;
DebugSession::close();
return false;
}
if (isDrag()) {
for ( AutoSegment* perpandicular : getPerpandiculars() ) {
if (perpandicular->getSpanLength() > minLength) {
cdebug_log(149,-1) << "False (drag), has length or PP has length." << endl;
DebugSession::close();
return false;
}
}
} else {
if ( ((_flags & (SegSourceBottom|SegTargetTop)) != (SegSourceBottom|SegTargetTop))
and ((_flags & (SegTargetBottom|SegSourceTop)) != (SegTargetBottom|SegSourceTop)) ) {
cdebug_log(149,-1) << "False, not part of a stacked VIA." << endl;
DebugSession::close();
return false;
}
}
Manage minimum area for VIAs in the P&R. * New: In Hurricane::Layer, add support for a minimal area (given in microns). Exported in the Python interface. * New: In Anabatic::AutoSegment::bloatStackedStrap(), method to enlarge too small vertically stacked VIAs. For now just create a square ensuring the minimal area. Potentially cause DRC error in some cases so needs to be refined. * New: In Anabatic::_gutAnabatic(), check for too little stacked VIAs and display a report.
2020-11-15 17:55:49 -06:00
DbU::Unit side = DbU::fromPhysical( std::sqrt(minArea) , DbU::UnitPower::Micro );
setWidth( side );
setDuSource( -side/2 );
setDuTarget( side/2 );
DRC correct on Arlet6505 / TSMC C180. Integrate new features and bug fixes so the Arlet 6502 benchs successfully passes real DRC with reference industrial tools. Short summary: * Manage minimum area for VIAs in Katana::Tracks. * Allow different wire width for wires perpandicular to the prefered routing direction. * StackedVIAs used in the clock tree no longer assume an uniform routing grid (same offset & pitch all the way up). * Some hard-coded patches in PowerRails for FlexLib. * New: In CRL/symbolic/cmos/kite.py & cmos45/kite.py, update the RoutingLayerGauges by adding the new PWireWidth parameter. Always zero in case of symbolic layout (too fine tuning). * New: In CRL::RoutingGauge, add accessor to PWireWidth parameter. Modify the clone method. * New: In CRL::RoutingLayerGauge, add new parameter "PWireWidth" to give the width of a wire when it not drawn in the prefered routing direction. If it is set to zero, the normal width is used. * New: In CRL::PyRoutingGauge, export the updated constructor interface. It is *not* backward compatible, one must add the PWireWidth parameter in the various kite.py configuration files (in etc/). * Change: In AnabaticEngine::_gutAnabatic(), disable the minimum area detection mechanism, replaced by a more complete one in Katana::Track. Left commented out for now, but will be removed in the future. * Change: In Anabatic::AutoContact::updateLayer(), now systematically calls setLayerAndWidth() to potentially resize the VIAs. This is needed in real mode as VIAs are *not* macro-generated but have their real final size. * Change: In Anabatic::AutoContact::setLayerAndWidth(), select the width and height of the contact using the gauge wire width *and* perpandicular *wire width*. * Change: In Anabatic::AutoSegment::_initialize(), the "VIA to same cap" to PWireWidth/2, this will be the size of the VIA in the non-preferred direction at the end cap (non-square in real mode). * Change: In Anabatic::AutoSegment::getExtensionCap(), makes different cases for symbolic and real. Use raw length in real, add half the wire width in symbolic. Add a flag to get the extension cap *only*, not increased of half the minimal spacing. * Change: In Anabatic::AutoSegment::bloatStackedStrap(), enhanced, but finally unused... * New: In Anabatic::AutoSegment::create(), use the PWireWidth when the segment is not in the preferred routing direction (and of minimal width). * New: In Anabatic::Configuration, add new getPWirewidth(), DPHorizontalWidth() and DPVerticalWidth() accessors. * Change: In AnabaticEngine::setupPreRouteds(), skip components in in "cut" material. We are only interested in objects containing some metal (happens in real mode when VIAs cuts are really there). * New: In Katana::PowerRailsPlanes::Rail::doLayout(), add an hard-coded patch that artificially enlarge the *wide wire* so the spacing for wide wire is enforced. For now, two pitches on each side for "FlexLib" gauge. * New: In Katana::Track, add support to find and correct small wire chunks so they respect the minimum area rules. Two helper functions: * ::hasSameLayerTurn(), to find if a a TrackElement as non-zero length perpandicular is same layer connected to it. * ::toFoundryGrid(), to ensure that all coordinates will be on the foundry grid (may move in a more shared location). * ::expandToMinArea(), try to expand, *in the routing direction* the too small wire so it respect the minimal area. Check for the free space in the track. Track::minExpandArea() go through all the TrackElements in the track to look for too small ones and correct them. * Change: In Katana::RoutingPlane, add an accessor to get the tracks. * New: In KatanaEngine::finalizeLayout(), add a post-treatment to find for minimal area violations. * Change: In cumulus/plugins.block.configuration.GaugeConf, add a routingBb attribute that will serve as a common reference to all the functions calculation track positions. We must not have two different reference for the core and the corona. The reference is always the corona when we working on a complete chip. * New: In cumulus/plugins.block.configuration.GaugeConf.getTrack(), Simplified and more reliable way of getting tracks positions. Use the routingBb. * New: In cumulus/plugins.block.configuration.GaugeConf.rpAccess(), Make use of getTrack() to get every metal strap on the right X/Y position. * New: In cumulus/plugins.block.configuration.GaugeConf.expandMinArea(), As those wires are left alone by the router, it is our responsability to abide by the minimal area rule here. Hence the code duplication from the router (bad). Mainly wires made for the clock tree, I mean. * Bug: In cumulus/plugins.chip.configuration.ChipConf.setupICore(), the core instance must be placed on the GCell grid, defined by the slice height (X *and* Y). * Bug: In cumulus/plugins.chip.corona.Builder(), forgot to use bigvia for the corners of the inner ring. * Bug: In cumulus/plugins.chip.pads.corona._createCoreWire(), hard-coded patch for LibreSOCIO, the power/ground connectors toward the core are too wide and can create DRC errors when put side by side. Shrink them by the minimal distance.
2020-11-23 16:07:15 -06:00
cdebug_log(149,-1) << "True, add area." << endl;
DebugSession::close();
Manage minimum area for VIAs in the P&R. * New: In Hurricane::Layer, add support for a minimal area (given in microns). Exported in the Python interface. * New: In Anabatic::AutoSegment::bloatStackedStrap(), method to enlarge too small vertically stacked VIAs. For now just create a square ensuring the minimal area. Potentially cause DRC error in some cases so needs to be refined. * New: In Anabatic::_gutAnabatic(), check for too little stacked VIAs and display a report.
2020-11-15 17:55:49 -06:00
return true;
}
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
#if THIS_IS_DISABLED
Replace "unsigned int" by "Flags" in all AutoSegments collections. * Change: In Anabatic::AutoSegments collections, change the type of all the flags that where in "unsigned int" (32 bits) to Flags (uint64_t) as there is now more than 32 flags for functions. * New: In Ababatic::Constants, added new flag Flags::WithPerpands, which makes the number of flags tip over 32 bits, thus making mandatory to uses Flags and not unsigned int. * New: In Anabatic::AutoSegments_Perpandiculars, manage a new flag Flags::WithDoglegs to allow to propagate through global segments that are connecteds via doglegs on local segments. Meaning that there is a good chance that they could be aligned. Slighly change the way we propagate on aligned segments: no longer check for VTee or HTee, but only for same direction and layer as master. * New: In Anabatic & Katana, replace all the "int", "long" and their variants by the less implementation ambiguous "int32_t", "int64_t" (and variant). This should help to better detect bit trucation in flags. Use the type to give a hint about the flags kind: - Type "Flags", for flags shared among Anabatic & Katana functions/methods (may also appear in some objects states). - Type "uint32_t" for flags belonging to an object internal state of from Hurricane functions flags (those should be grouped in a Flag subclass in a perfect world).
2017-05-16 07:53:33 -05:00
bool AutoSegment::shearUp ( GCell* upGCell, AutoSegment*& movedUp, float reserve, Flags flags )
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
{
cdebug_log(149,0) << "AutoSegment::shearUp() " << this << endl;
movedUp = NULL;
if ( isLayerChange() or isFixed() /*or isTerminal()*/ or isLocal() ) return false;
size_t upDepth = Session::getRoutingGauge()->getLayerDepth(getLayer()) + 2;
if ( upDepth > Session::getConfiguration()->getAllowedDepth() ) return false;
vector<GCell*> gcells;
getGCells ( gcells );
size_t iupGCell = 0;
for ( ; iupGCell<gcells.size() ; ++iupGCell ) {
if ( gcells[iupGCell] == upGCell ) break;
}
if ( iupGCell == gcells.size() ) {
cerr << Warning("Shear start %s not under %s."
,getString(upGCell).c_str()
,getString(this).c_str()
) << endl;
return false;
}
GCell* rightShear = NULL;
for ( size_t i=iupGCell ; i<gcells.size() ; i++ ) {
if ( not gcells[i]->hasFreeTrack(upDepth,reserve) ) {
cdebug_log(149,0) << "Right shearing @ " << gcells[i] << endl;
rightShear = gcells[i];
}
}
GCell* leftShear = NULL;
if ( iupGCell > 0 ) {
size_t i = iupGCell;
do {
--i;
if ( not gcells[i]->hasFreeTrack(upDepth,reserve) ) {
cdebug_log(149,0) << "Left shearing @ " << gcells[i] << endl;
leftShear = gcells[i];
}
} while (i > 0);
}
AutoSegment* before = this;
const vector<AutoSegment*>& doglegs = Session::getDoglegs();
if ( leftShear ) {
makeDogleg ( leftShear, true );
movedUp = doglegs[2];
} else {
before = NULL;
movedUp = this;
}
if ( rightShear ) makeDogleg(rightShear,true);
if ( movedUp->moveUp(flags) ) {
if ( rightShear or leftShear )
cinfo << "Shearing Up " << this << "." << endl;
return true;
}
movedUp = NULL;
return false;
}
#endif
Replace "unsigned int" by "Flags" in all AutoSegments collections. * Change: In Anabatic::AutoSegments collections, change the type of all the flags that where in "unsigned int" (32 bits) to Flags (uint64_t) as there is now more than 32 flags for functions. * New: In Ababatic::Constants, added new flag Flags::WithPerpands, which makes the number of flags tip over 32 bits, thus making mandatory to uses Flags and not unsigned int. * New: In Anabatic::AutoSegments_Perpandiculars, manage a new flag Flags::WithDoglegs to allow to propagate through global segments that are connecteds via doglegs on local segments. Meaning that there is a good chance that they could be aligned. Slighly change the way we propagate on aligned segments: no longer check for VTee or HTee, but only for same direction and layer as master. * New: In Anabatic & Katana, replace all the "int", "long" and their variants by the less implementation ambiguous "int32_t", "int64_t" (and variant). This should help to better detect bit trucation in flags. Use the type to give a hint about the flags kind: - Type "Flags", for flags shared among Anabatic & Katana functions/methods (may also appear in some objects states). - Type "uint32_t" for flags belonging to an object internal state of from Hurricane functions flags (those should be grouped in a Flag subclass in a perfect world).
2017-05-16 07:53:33 -05:00
Flags AutoSegment::canDogleg ( Interval interval )
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
{
cdebug_log(149,0) << "AutoSegment::canDogleg(Interval) " << interval << endl;
size_t leftDogleg = 0;
size_t rightDogleg = 0;
if (getSpanU().contains(interval.getVMin())) leftDogleg++;
if (getSpanU().contains(interval.getVMax())) rightDogleg++;
if (not isNotAligned()) {
Added support for 2-Metal block routing in Anabatic & Katana. * New: In AnabaticEngine::invalidateRoutingPads() this method is a temporary workaround for a Hurricane problems. When an instance is moved, the RoutingPads that use it must be moved accordingly, but they are not invalidated so they stay in the wrong QuadTree. New method ::_resizeMatrix() to be called when the associated Cell is resized. * Bug: In AutoHorizontal::getConstraints() and AutoVertical::getConstraints(), the *target* constraints where never merged. * Change: In AutoHorizontal::getCells() and AutoVertical::getGCells(), now return a boolean to tell if it was ok (must not encounter a NULL GCell while progessing from source to target). * New: In Anabatic::Configuration and Anabatic:Session, create new methods: - getDHorizontalLayer() - getDhorizontalDepth() - getDHorizontalWidth() - getDHorizontalPitch() And so on for Vertical and Contact. They supply depth-independant informations about the H/V layers to build the initial detailed routing. The AutoSegment::create() methods have been modificated accordingly. * New: In Anabatic::GCell, add two new types "StdCellRow" and "ChannelRow" for implementing 2-Metal blocks. Rename the GCell::setXY() method in GCell::setSouthWestCorner(), move the contents of GCell::updateContactsPosition() into it and suppress it. WARNING: In case of a GCell shrink this may cause problems. But for now we only expand... New method GCell::getNetCount() to count the number of Net going though the GCell. * Change: In Anabatic::Edge, add specific support for capacity of 2-Metal routing channels. * Change: In Anabatic::Dijsktra various methods, replace the "gcell->isMatrix()" calls by "not gcell->isAnalog()". Add more check so that the methods pertaining to the analog routing (GRData) are not called in digital mode. * New: In Anabatic::Dijkstra::materialize(), add support for 2-Metal specific cases. That is, always break in case of vertical pass-through or U-turn. The global routing must always be broken in H-Channel. * New: In Anabatic::GCell & Anabatic::Edge, make use of the Session mechanism to ensure the revalidation. The "::revalidate()" method is then moved as "::materialize()" (overload of Go) and "::_invalidate()" becomes "::invalidate()" * Change: In LoadGlobalRouting, cosmetic rename of SortHkByX in SortHookByX. * New: In GCellTopology, added support for building 2-Metal topologies. * ForkStack is now an object attribute as many methods do need it. * To push segments/hook on the stack, a new method "push()" is available. Perform NULL and fromHook checking. Can also setup _southWestContact or _northEastContact if it is the "from" edge. * N/S/E/W edges are now vector as in digital channel mode there can be more than one. * Added build topological build methods: - doRp_2m_Access() RoutingPad stem access. - _do_2m_1G_1M1() North or south access. - _do_2m_2G_1M1() North AND south access. - _do_2m_xG() H-Channel routing. * New: In Anabatic::Matrix, new ::resize() function, as Cell can be resizeds. * New: In Anabatic::Vertex, new static method ::getValueString() for a friendly text rendering. * New: In Katana::DigitalDistance, support for channel routing. * Change: In KatanaEngine::digitalSetup() and KatanaEngine::runGlobalrouter(), for channel routing, calls to setupPowerRails() and protectRoutingPads() must be called after the core block has been fully dimensionned. ::runGlobalrouter() contains the code tasked with the grid creation and channel sizing. * New: In KatanaEngine: Added support for core block, for 2-Metal routing. May be expanded for over-the-cell routing in the future. Added methods : - isDigitalMode() - isAnalogMode() - isMixedMode() - isChannelMode() - getBlock() / addBlock() - setupChannelMode() - createChannel() * New: In Katana, new class Block to manage core blocks and perform channel routing. * New: In Katana::Session, new convenience method "isOpen()".
2017-08-18 16:56:23 -05:00
for ( AutoSegment* segment : getAligneds() ) {
if (segment->getSpanU().contains(interval.getVMin())) {
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
if (segment->isFixed()) return Flags::NoFlags;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
leftDogleg++;
}
Added support for 2-Metal block routing in Anabatic & Katana. * New: In AnabaticEngine::invalidateRoutingPads() this method is a temporary workaround for a Hurricane problems. When an instance is moved, the RoutingPads that use it must be moved accordingly, but they are not invalidated so they stay in the wrong QuadTree. New method ::_resizeMatrix() to be called when the associated Cell is resized. * Bug: In AutoHorizontal::getConstraints() and AutoVertical::getConstraints(), the *target* constraints where never merged. * Change: In AutoHorizontal::getCells() and AutoVertical::getGCells(), now return a boolean to tell if it was ok (must not encounter a NULL GCell while progessing from source to target). * New: In Anabatic::Configuration and Anabatic:Session, create new methods: - getDHorizontalLayer() - getDhorizontalDepth() - getDHorizontalWidth() - getDHorizontalPitch() And so on for Vertical and Contact. They supply depth-independant informations about the H/V layers to build the initial detailed routing. The AutoSegment::create() methods have been modificated accordingly. * New: In Anabatic::GCell, add two new types "StdCellRow" and "ChannelRow" for implementing 2-Metal blocks. Rename the GCell::setXY() method in GCell::setSouthWestCorner(), move the contents of GCell::updateContactsPosition() into it and suppress it. WARNING: In case of a GCell shrink this may cause problems. But for now we only expand... New method GCell::getNetCount() to count the number of Net going though the GCell. * Change: In Anabatic::Edge, add specific support for capacity of 2-Metal routing channels. * Change: In Anabatic::Dijsktra various methods, replace the "gcell->isMatrix()" calls by "not gcell->isAnalog()". Add more check so that the methods pertaining to the analog routing (GRData) are not called in digital mode. * New: In Anabatic::Dijkstra::materialize(), add support for 2-Metal specific cases. That is, always break in case of vertical pass-through or U-turn. The global routing must always be broken in H-Channel. * New: In Anabatic::GCell & Anabatic::Edge, make use of the Session mechanism to ensure the revalidation. The "::revalidate()" method is then moved as "::materialize()" (overload of Go) and "::_invalidate()" becomes "::invalidate()" * Change: In LoadGlobalRouting, cosmetic rename of SortHkByX in SortHookByX. * New: In GCellTopology, added support for building 2-Metal topologies. * ForkStack is now an object attribute as many methods do need it. * To push segments/hook on the stack, a new method "push()" is available. Perform NULL and fromHook checking. Can also setup _southWestContact or _northEastContact if it is the "from" edge. * N/S/E/W edges are now vector as in digital channel mode there can be more than one. * Added build topological build methods: - doRp_2m_Access() RoutingPad stem access. - _do_2m_1G_1M1() North or south access. - _do_2m_2G_1M1() North AND south access. - _do_2m_xG() H-Channel routing. * New: In Anabatic::Matrix, new ::resize() function, as Cell can be resizeds. * New: In Anabatic::Vertex, new static method ::getValueString() for a friendly text rendering. * New: In Katana::DigitalDistance, support for channel routing. * Change: In KatanaEngine::digitalSetup() and KatanaEngine::runGlobalrouter(), for channel routing, calls to setupPowerRails() and protectRoutingPads() must be called after the core block has been fully dimensionned. ::runGlobalrouter() contains the code tasked with the grid creation and channel sizing. * New: In KatanaEngine: Added support for core block, for 2-Metal routing. May be expanded for over-the-cell routing in the future. Added methods : - isDigitalMode() - isAnalogMode() - isMixedMode() - isChannelMode() - getBlock() / addBlock() - setupChannelMode() - createChannel() * New: In Katana, new class Block to manage core blocks and perform channel routing. * New: In Katana::Session, new convenience method "isOpen()".
2017-08-18 16:56:23 -05:00
if (segment->getSpanU().contains(interval.getVMax())) {
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
if (segment->isFixed()) return Flags::NoFlags;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
rightDogleg++;
}
}
}
if ( (leftDogleg == 1) and (rightDogleg <= 1) ) return Flags::DoglegOnLeft;
if ( (leftDogleg <= 1) and (rightDogleg == 1) ) return Flags::DoglegOnRight;
cdebug_log(149,0) << "leftCount:" << leftDogleg << " rightCount:" << rightDogleg << endl;
Added support for 2-Metal block routing in Anabatic & Katana. * New: In AnabaticEngine::invalidateRoutingPads() this method is a temporary workaround for a Hurricane problems. When an instance is moved, the RoutingPads that use it must be moved accordingly, but they are not invalidated so they stay in the wrong QuadTree. New method ::_resizeMatrix() to be called when the associated Cell is resized. * Bug: In AutoHorizontal::getConstraints() and AutoVertical::getConstraints(), the *target* constraints where never merged. * Change: In AutoHorizontal::getCells() and AutoVertical::getGCells(), now return a boolean to tell if it was ok (must not encounter a NULL GCell while progessing from source to target). * New: In Anabatic::Configuration and Anabatic:Session, create new methods: - getDHorizontalLayer() - getDhorizontalDepth() - getDHorizontalWidth() - getDHorizontalPitch() And so on for Vertical and Contact. They supply depth-independant informations about the H/V layers to build the initial detailed routing. The AutoSegment::create() methods have been modificated accordingly. * New: In Anabatic::GCell, add two new types "StdCellRow" and "ChannelRow" for implementing 2-Metal blocks. Rename the GCell::setXY() method in GCell::setSouthWestCorner(), move the contents of GCell::updateContactsPosition() into it and suppress it. WARNING: In case of a GCell shrink this may cause problems. But for now we only expand... New method GCell::getNetCount() to count the number of Net going though the GCell. * Change: In Anabatic::Edge, add specific support for capacity of 2-Metal routing channels. * Change: In Anabatic::Dijsktra various methods, replace the "gcell->isMatrix()" calls by "not gcell->isAnalog()". Add more check so that the methods pertaining to the analog routing (GRData) are not called in digital mode. * New: In Anabatic::Dijkstra::materialize(), add support for 2-Metal specific cases. That is, always break in case of vertical pass-through or U-turn. The global routing must always be broken in H-Channel. * New: In Anabatic::GCell & Anabatic::Edge, make use of the Session mechanism to ensure the revalidation. The "::revalidate()" method is then moved as "::materialize()" (overload of Go) and "::_invalidate()" becomes "::invalidate()" * Change: In LoadGlobalRouting, cosmetic rename of SortHkByX in SortHookByX. * New: In GCellTopology, added support for building 2-Metal topologies. * ForkStack is now an object attribute as many methods do need it. * To push segments/hook on the stack, a new method "push()" is available. Perform NULL and fromHook checking. Can also setup _southWestContact or _northEastContact if it is the "from" edge. * N/S/E/W edges are now vector as in digital channel mode there can be more than one. * Added build topological build methods: - doRp_2m_Access() RoutingPad stem access. - _do_2m_1G_1M1() North or south access. - _do_2m_2G_1M1() North AND south access. - _do_2m_xG() H-Channel routing. * New: In Anabatic::Matrix, new ::resize() function, as Cell can be resizeds. * New: In Anabatic::Vertex, new static method ::getValueString() for a friendly text rendering. * New: In Katana::DigitalDistance, support for channel routing. * Change: In KatanaEngine::digitalSetup() and KatanaEngine::runGlobalrouter(), for channel routing, calls to setupPowerRails() and protectRoutingPads() must be called after the core block has been fully dimensionned. ::runGlobalrouter() contains the code tasked with the grid creation and channel sizing. * New: In KatanaEngine: Added support for core block, for 2-Metal routing. May be expanded for over-the-cell routing in the future. Added methods : - isDigitalMode() - isAnalogMode() - isMixedMode() - isChannelMode() - getBlock() / addBlock() - setupChannelMode() - createChannel() * New: In Katana, new class Block to manage core blocks and perform channel routing. * New: In Katana::Session, new convenience method "isOpen()".
2017-08-18 16:56:23 -05:00
return Flags::NoFlags;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
}
AutoSegment* AutoSegment::makeDogleg ( AutoContact* from )
{
cdebug_log(149,1) << "AutoSegment::makeDogleg(AutoContact*) " << from << endl;
cdebug_log(149,0) << this << endl;
RoutingGauge* rg = Session::getRoutingGauge();
size_t segmentDepth = rg->getLayerDepth( getLayer() );
const vector<AutoSegment*>& doglegs = Session::getDoglegs();
size_t index = doglegs.size();
bool isSource = (getAutoSource() == from);
cdebug_log(149,0) << "isSource:" << isSource << endl;
makeDogleg( from->getGCell(), Flags::NoCheckLayer );
if (doglegs.size() == index) {
cdebug_tabw(149,-1);
return NULL;
}
doglegs[ index+1 ]->setAxis( isHorizontal() ? from->getX() : from->getY() );
if (not from->getLayer()->contains(getLayer())) {
cdebug_log(149,0) << "Contact layer do not contains Segment layer, adjust layers" << endl;
if (getLayer()->above(from->getLayer())) {
cdebug_log(149,0) << "Go Down from depth " << segmentDepth << endl;
In Anabatic, correctly set the wire width according to the routing gauge. * New: Anabatic::AutoContact::setLayerAndWidth() to set both layer and VIA width/side according to the RoutingGauge. If the delta in zero, use the metal gauge, and the VIA gauge otherwise. * Bug: In Anabatic::AutoContactTerminal, Anabatic::AutoContactTurn, Anabatic::AutoContactHTee & Anabatic::AutoContactVTee, in the updateTopology() method, set both the layer and the VIA with when there is a change of layer. Note that this default size may be overriden later by updateSeize() in the case of non-default width segments. * New: In Anabatic::AutoSegment, new overload setLayer(size_t) to set both layer and segment with according to the routing gauge. * Bug: In Anabatic::AutoHorizontal and Anabatic::AutoVertical, in method _makeDogleg(), make use of the new setLayer() to correctly set up the wire width. Idem for Anabatic::AutoSegment::changeDepth() and ::makeDogleg(), and in Anabatic::LayerAssign.
2019-02-12 05:43:09 -06:00
doglegs[ index + 1 ]->setLayer( segmentDepth-1 );
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
cdebug_log(149,0) << "doglegs[i+1]: " << doglegs[index+1] << endl;
if (isSource) {
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
doglegs[ index + 0 ]->setLayer( std::max((size_t)1,segmentDepth-2) );
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
cdebug_log(149,0) << "doglegs[i+0]: " << doglegs[index+0] << endl;
cdebug_log(149,0) << "doglegs[i+1]: " << doglegs[index+1]->getAutoSource() << endl;
cdebug_log(149,0) << "doglegs[i+1]: " << doglegs[index+1]->getAutoTarget() << endl;
} else {
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
doglegs[ index + 2 ]->setLayer( std::max((size_t)1,segmentDepth-2) );
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
cdebug_log(149,0) << "doglegs[i+2]: " << doglegs[index+2] << endl;
cdebug_log(149,0) << "doglegs[i+1]: " << doglegs[index+1]->getAutoTarget() << endl;
cdebug_log(149,0) << "doglegs[i+1]: " << doglegs[index+1]->getAutoSource() << endl;
}
} else {
cdebug_log(149,0) << "Go Up from depth " << segmentDepth << endl;
In Anabatic, correctly set the wire width according to the routing gauge. * New: Anabatic::AutoContact::setLayerAndWidth() to set both layer and VIA width/side according to the RoutingGauge. If the delta in zero, use the metal gauge, and the VIA gauge otherwise. * Bug: In Anabatic::AutoContactTerminal, Anabatic::AutoContactTurn, Anabatic::AutoContactHTee & Anabatic::AutoContactVTee, in the updateTopology() method, set both the layer and the VIA with when there is a change of layer. Note that this default size may be overriden later by updateSeize() in the case of non-default width segments. * New: In Anabatic::AutoSegment, new overload setLayer(size_t) to set both layer and segment with according to the routing gauge. * Bug: In Anabatic::AutoHorizontal and Anabatic::AutoVertical, in method _makeDogleg(), make use of the new setLayer() to correctly set up the wire width. Idem for Anabatic::AutoSegment::changeDepth() and ::makeDogleg(), and in Anabatic::LayerAssign.
2019-02-12 05:43:09 -06:00
doglegs[ index + 1 ]->setLayer( segmentDepth+1 );
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
cdebug_log(149,0) << "doglegs[i+1]: " << doglegs[index+1] << endl;
if (isSource) {
In Anabatic, correctly set the wire width according to the routing gauge. * New: Anabatic::AutoContact::setLayerAndWidth() to set both layer and VIA width/side according to the RoutingGauge. If the delta in zero, use the metal gauge, and the VIA gauge otherwise. * Bug: In Anabatic::AutoContactTerminal, Anabatic::AutoContactTurn, Anabatic::AutoContactHTee & Anabatic::AutoContactVTee, in the updateTopology() method, set both the layer and the VIA with when there is a change of layer. Note that this default size may be overriden later by updateSeize() in the case of non-default width segments. * New: In Anabatic::AutoSegment, new overload setLayer(size_t) to set both layer and segment with according to the routing gauge. * Bug: In Anabatic::AutoHorizontal and Anabatic::AutoVertical, in method _makeDogleg(), make use of the new setLayer() to correctly set up the wire width. Idem for Anabatic::AutoSegment::changeDepth() and ::makeDogleg(), and in Anabatic::LayerAssign.
2019-02-12 05:43:09 -06:00
doglegs[ index + 0 ]->setLayer( segmentDepth+2 );
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
cdebug_log(149,0) << "doglegs[i+0]: " << doglegs[index+0] << endl;
cdebug_log(149,0) << "doglegs[i+1]: " << doglegs[index+1]->getAutoSource() << endl;
cdebug_log(149,0) << "doglegs[i+1]: " << doglegs[index+1]->getAutoTarget() << endl;
} else {
In Anabatic, correctly set the wire width according to the routing gauge. * New: Anabatic::AutoContact::setLayerAndWidth() to set both layer and VIA width/side according to the RoutingGauge. If the delta in zero, use the metal gauge, and the VIA gauge otherwise. * Bug: In Anabatic::AutoContactTerminal, Anabatic::AutoContactTurn, Anabatic::AutoContactHTee & Anabatic::AutoContactVTee, in the updateTopology() method, set both the layer and the VIA with when there is a change of layer. Note that this default size may be overriden later by updateSeize() in the case of non-default width segments. * New: In Anabatic::AutoSegment, new overload setLayer(size_t) to set both layer and segment with according to the routing gauge. * Bug: In Anabatic::AutoHorizontal and Anabatic::AutoVertical, in method _makeDogleg(), make use of the new setLayer() to correctly set up the wire width. Idem for Anabatic::AutoSegment::changeDepth() and ::makeDogleg(), and in Anabatic::LayerAssign.
2019-02-12 05:43:09 -06:00
doglegs[ index + 2 ]->setLayer( segmentDepth+2 );
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
cdebug_log(149,0) << "doglegs[i+2]: " << doglegs[index+2] << endl;
cdebug_log(149,0) << "doglegs[i+1]: " << doglegs[index+1]->getAutoTarget() << endl;
cdebug_log(149,0) << "doglegs[i+1]: " << doglegs[index+1]->getAutoSource() << endl;
}
}
Migrating the initialisation system to be completely Python-like. * New: In bootstrap/coriolisEnv.py, add the "etc" directory to the PYTHONPATH as initialization are now Python modules. * New: In Hurricane/analogic, first groundwork for the integration of PIP/MIM/MOM multi-capacitors. Add C++ and Python interface for the allocation matrix and the list of capacities values. * Change: In Hurricane::RegularLayer, add a layer parameter to the constructor so the association between the RegularLayer and it's BasicLayer can readily be done. * Change: In Hurricane::Layer, add a new getCut() accessor to get the cut layer in ViaLayer. * Change: In Hurricane::DataBase::get(), the Python wrapper should no longer consider an error if the data-base has not been created yet. Just return None. * Bug: In Isobar::PyLayer::getEnclosure() wrapper, if the overall enclosure is requested, pass the right parameter to the C++ function. * Change: In AllianceFramework, make public _bindLibraries() and export it to the Python interface. * Change: In AllianceFramework::create(), do not longer call bindLibraries(). This now must be done explicitely and afterwards. * Change: In AllianceFramework::createLibrary() and Environement::addSYSTEM_LIBRARY(), minor bug corrections that I don't recall. * Change: In SearchPath::prepend(), set the selected index to zero and return it. * Change: In CRL::System CTOR, add "etc" to the PYTHONPATH as the configuration files are now organized as Python modules. * New: In PyCRL, export the CRL::System singleton, it's creation is no longer triggered by the one of AllianceFramework. * New: In CRL/etc/, convert most of the configuration files into the Python module format. For now, keep the old ".conf", but that are no longer used. For the real technologies, we cannot keep the directory name as "180" or "45" as it not allowed by Python syntax, so we create "node180" or "node45" instead. Most of the helpers and coriolisInit.py are no longer used now. To be removed in future commits after being sure that everything works... * Bug: In AutoSegment::makeDogleg(AutoContact*), the layer of the contacts where badly computed when one end of the original segment was attached to a non-preferred direction segment (mostly on terminal contacts). Now use the new AutoContact::updateLayer() method. * Bug: In Dijkstra::load(), limit symetric search area only if the net is a symmetric one ! * Change: In Katana/python/katanaInit.py, comply with the new initialisation scheme. * Change: In Unicorn/cgt.py, comply to the new inititalization scheme. * Change: In cumulus various Python scripts remove the call to helpers.staticInitialization() as they are not needed now (we run in only *one* interpreter, so we correctly share all init). In plugins/__init__.py, read the new NDA directory variable. * Bug: In cumulus/plugins/Chip.doCoronafloorplan(), self.railsNb was not correctly managed when there was no clock. * Change: In cumulus/plugins/Configuration.coronaContactArray(), compute the viaPitch from the technology instead of the hard-coded 4.0 lambdas. In Configuration.loadConfiguration(), read the "ioring.py" from the new user's settings module. * Bug: In stratus.dpgen_ADSB2F, gives coordinates translated into DbU to the XY functions. In st_model.Save(), use the VstUseConcat flag to get correct VST files. In st_net.hur_net(), when a net is POWER/GROUND or CLOCK also make it global. * Change: In Oroshi/python/WIP_Transistor.py, encapsulate the generator inside a try/except block to get prettier error (and stop at the first).
2019-10-28 12:09:14 -05:00
doglegs[ index + 1 ]->getAutoSource()->updateLayer();
doglegs[ index + 1 ]->getAutoTarget()->updateLayer();
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
}
cdebug_tabw(149,-1);
return doglegs[ index + (isSource?0:2) ];
}
Replace "unsigned int" by "Flags" in all AutoSegments collections. * Change: In Anabatic::AutoSegments collections, change the type of all the flags that where in "unsigned int" (32 bits) to Flags (uint64_t) as there is now more than 32 flags for functions. * New: In Ababatic::Constants, added new flag Flags::WithPerpands, which makes the number of flags tip over 32 bits, thus making mandatory to uses Flags and not unsigned int. * New: In Anabatic::AutoSegments_Perpandiculars, manage a new flag Flags::WithDoglegs to allow to propagate through global segments that are connecteds via doglegs on local segments. Meaning that there is a good chance that they could be aligned. Slighly change the way we propagate on aligned segments: no longer check for VTee or HTee, but only for same direction and layer as master. * New: In Anabatic & Katana, replace all the "int", "long" and their variants by the less implementation ambiguous "int32_t", "int64_t" (and variant). This should help to better detect bit trucation in flags. Use the type to give a hint about the flags kind: - Type "Flags", for flags shared among Anabatic & Katana functions/methods (may also appear in some objects states). - Type "uint32_t" for flags belonging to an object internal state of from Hurricane functions flags (those should be grouped in a Flag subclass in a perfect world).
2017-05-16 07:53:33 -05:00
Flags AutoSegment::makeDogleg ( Interval interval, Flags flags )
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
{
cdebug_log(149,1) << "AutoSegment::makeDogleg(Interval) - " << interval << endl;
bool leftDogleg = true;
Replace "unsigned int" by "Flags" in all AutoSegments collections. * Change: In Anabatic::AutoSegments collections, change the type of all the flags that where in "unsigned int" (32 bits) to Flags (uint64_t) as there is now more than 32 flags for functions. * New: In Ababatic::Constants, added new flag Flags::WithPerpands, which makes the number of flags tip over 32 bits, thus making mandatory to uses Flags and not unsigned int. * New: In Anabatic::AutoSegments_Perpandiculars, manage a new flag Flags::WithDoglegs to allow to propagate through global segments that are connecteds via doglegs on local segments. Meaning that there is a good chance that they could be aligned. Slighly change the way we propagate on aligned segments: no longer check for VTee or HTee, but only for same direction and layer as master. * New: In Anabatic & Katana, replace all the "int", "long" and their variants by the less implementation ambiguous "int32_t", "int64_t" (and variant). This should help to better detect bit trucation in flags. Use the type to give a hint about the flags kind: - Type "Flags", for flags shared among Anabatic & Katana functions/methods (may also appear in some objects states). - Type "uint32_t" for flags belonging to an object internal state of from Hurricane functions flags (those should be grouped in a Flag subclass in a perfect world).
2017-05-16 07:53:33 -05:00
Flags rflags = Flags::NoFlags;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
size_t leftDoglegCount = 0;
size_t rightDoglegCount = 0;
AutoSegment* leftCandidate = NULL;
AutoSegment* rightCandidate = NULL;
if (getSpanU().contains(interval.getVMin())) { leftCandidate = this; leftDoglegCount++; }
if (getSpanU().contains(interval.getVMax())) { rightCandidate = this; rightDoglegCount++; }
if (not isNotAligned()) {
forEach ( AutoSegment*, isegment, getAligneds(flags) ) {
if (isegment->getSpanU().contains(interval.getVMin())) { leftCandidate = *isegment; leftDoglegCount++; }
if (isegment->getSpanU().contains(interval.getVMax())) { rightCandidate = *isegment; rightDoglegCount++; }
}
}
if ( (leftDoglegCount != 1) and (rightDoglegCount != 1) ) { cdebug_tabw(149,-1); return 0; }
if (not leftDoglegCount) {
leftDogleg = false;
leftCandidate = rightCandidate;
rightCandidate = NULL;
}
if (leftCandidate and rightCandidate) {
cdebug_log(149,0) << "Left Constraint: " << leftCandidate->getSourceConstraints(Flags::NativeConstraints) << endl;
cdebug_log(149,0) << "Right Constraint: " << rightCandidate->getTargetConstraints(Flags::NativeConstraints) << endl;
if ( leftCandidate ->getTargetConstraints(Flags::NativeConstraints).getSize()
< rightCandidate->getSourceConstraints(Flags::NativeConstraints).getSize() ) {
leftCandidate = rightCandidate;
leftDogleg = false;
}
} else {
if (not leftCandidate) {
leftCandidate = rightCandidate;
leftDogleg = false;
}
}
if (leftCandidate) {
DbU::Unit axis;
// Ugly: Hard-wired track spacing.
if (leftDogleg) axis = interval.getVMin() - getPitch();
else axis = interval.getVMax() + getPitch();
cdebug_log(149,0) << "Break @" << DbU::getValueString(axis) << " " << leftCandidate << endl;
Replace "unsigned int" by "Flags" in all AutoSegments collections. * Change: In Anabatic::AutoSegments collections, change the type of all the flags that where in "unsigned int" (32 bits) to Flags (uint64_t) as there is now more than 32 flags for functions. * New: In Ababatic::Constants, added new flag Flags::WithPerpands, which makes the number of flags tip over 32 bits, thus making mandatory to uses Flags and not unsigned int. * New: In Anabatic::AutoSegments_Perpandiculars, manage a new flag Flags::WithDoglegs to allow to propagate through global segments that are connecteds via doglegs on local segments. Meaning that there is a good chance that they could be aligned. Slighly change the way we propagate on aligned segments: no longer check for VTee or HTee, but only for same direction and layer as master. * New: In Anabatic & Katana, replace all the "int", "long" and their variants by the less implementation ambiguous "int32_t", "int64_t" (and variant). This should help to better detect bit trucation in flags. Use the type to give a hint about the flags kind: - Type "Flags", for flags shared among Anabatic & Katana functions/methods (may also appear in some objects states). - Type "uint32_t" for flags belonging to an object internal state of from Hurricane functions flags (those should be grouped in a Flag subclass in a perfect world).
2017-05-16 07:53:33 -05:00
Flags direction = getDirection();
GCell* gcell = leftCandidate->getAutoSource()->getGCell();
GCell* end = leftCandidate->getAutoTarget()->getGCell();
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
while ( gcell != end ) {
if (gcell->getSide(direction).contains(axis)) break;
gcell = (direction == Flags::Horizontal) ? gcell->getEast (getNativeMin())
: gcell->getNorth(getNativeMin());
}
cdebug_log(149,0) << "In " << gcell << endl;
rflags = leftCandidate->_makeDogleg( gcell, flags );
const vector<AutoSegment*>& doglegs = Session::getDoglegs();
if (doglegs.size() >= 2) {
cdebug_log(149,0) << "AutoSegment::makeDogleg(): @" << DbU::getValueString(axis) << endl;
doglegs[1]->setAxis( axis );
}
}
cdebug_tabw(149,-1);
return rflags | (leftDogleg ? Flags::DoglegOnLeft : Flags::DoglegOnRight);
}
Replace "unsigned int" by "Flags" in all AutoSegments collections. * Change: In Anabatic::AutoSegments collections, change the type of all the flags that where in "unsigned int" (32 bits) to Flags (uint64_t) as there is now more than 32 flags for functions. * New: In Ababatic::Constants, added new flag Flags::WithPerpands, which makes the number of flags tip over 32 bits, thus making mandatory to uses Flags and not unsigned int. * New: In Anabatic::AutoSegments_Perpandiculars, manage a new flag Flags::WithDoglegs to allow to propagate through global segments that are connecteds via doglegs on local segments. Meaning that there is a good chance that they could be aligned. Slighly change the way we propagate on aligned segments: no longer check for VTee or HTee, but only for same direction and layer as master. * New: In Anabatic & Katana, replace all the "int", "long" and their variants by the less implementation ambiguous "int32_t", "int64_t" (and variant). This should help to better detect bit trucation in flags. Use the type to give a hint about the flags kind: - Type "Flags", for flags shared among Anabatic & Katana functions/methods (may also appear in some objects states). - Type "uint32_t" for flags belonging to an object internal state of from Hurricane functions flags (those should be grouped in a Flag subclass in a perfect world).
2017-05-16 07:53:33 -05:00
Flags AutoSegment::makeDogleg ( GCell* doglegGCell, Flags flags )
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
{
cdebug_log(9000,0) << "Deter| AutoSegment::makeDogleg(GCell*) " << doglegGCell << endl;
cdebug_log(9000,0) << "Deter| in " << this << endl;
cdebug_tabw(149,1);
Replace "unsigned int" by "Flags" in all AutoSegments collections. * Change: In Anabatic::AutoSegments collections, change the type of all the flags that where in "unsigned int" (32 bits) to Flags (uint64_t) as there is now more than 32 flags for functions. * New: In Ababatic::Constants, added new flag Flags::WithPerpands, which makes the number of flags tip over 32 bits, thus making mandatory to uses Flags and not unsigned int. * New: In Anabatic::AutoSegments_Perpandiculars, manage a new flag Flags::WithDoglegs to allow to propagate through global segments that are connecteds via doglegs on local segments. Meaning that there is a good chance that they could be aligned. Slighly change the way we propagate on aligned segments: no longer check for VTee or HTee, but only for same direction and layer as master. * New: In Anabatic & Katana, replace all the "int", "long" and their variants by the less implementation ambiguous "int32_t", "int64_t" (and variant). This should help to better detect bit trucation in flags. Use the type to give a hint about the flags kind: - Type "Flags", for flags shared among Anabatic & Katana functions/methods (may also appear in some objects states). - Type "uint32_t" for flags belonging to an object internal state of from Hurricane functions flags (those should be grouped in a Flag subclass in a perfect world).
2017-05-16 07:53:33 -05:00
Flags rflags = Flags::NoFlags;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
if ( doglegGCell->isIoPad()
and (Session::getAnabatic()->getState() != EngineGlobalLoaded) ) {
cerr << Bug( "Attempt to make a dogleg in a GCell under a Pad\n"
" %s\n"
" %s"
, getString(this).c_str()
, getString(doglegGCell).c_str() ) << endl;
}
if (isFixed()) {
cerr << Error( "AutoSegment::makeDogleg(): Cannot make a dog leg on a fixed segment.\n"
" (on: %s)", _getString().c_str() ) << endl;
return 0;
}
if (doglegGCell->getSide(getDirection()).intersect(getSpanU())) {
Solved more problems related to GCell not owning north/east boundary. * Bug: In Anabatic, in AutoHorizontal & AutoVertical, native constraints must use GCell::getConstraintsUMax(). * Bug: In Anabatic::GCell, correct terribly bad implementation of both getConstraintUmax(). In getSide(), use getConstraintUMax(). * Bug: In Anabatic::AutoContact, setCBUMax() must be set using the GCell getContraintUMax().
2016-08-30 09:05:15 -05:00
cdebug_log(149,0) << "Dogleg in " << this << " spanU:" << getSpanU() << endl;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
rflags = _makeDogleg( doglegGCell, flags );
} else {
cdebug_log(149,0) << "Looking in aligneds." << endl;
if (not isNotAligned()) {
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
for ( AutoSegment* aligned : getAligneds(flags) ) {
cdebug_log(149,0) << "| Try in " << aligned << endl;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
if (doglegGCell->getSide(getDirection()).intersect(aligned->getSpanU())) {
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
cdebug_log(149,0) << "Dogleg in " << aligned << endl;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
rflags = aligned->_makeDogleg( doglegGCell, flags );
cdebug_tabw(149,-1);
return 0;
}
}
}
cerr << Bug("Cannot make a dogleg in %s at %s"
,_getString().c_str(), getString(doglegGCell).c_str()) << endl;
}
cdebug_tabw(149,-1);
return rflags;
}
bool AutoSegment::_check () const
{
bool coherency = true;
coherency = checkNotInvalidated() and coherency;
coherency = checkPositions() and coherency;
coherency = checkConstraints() and coherency;
coherency = checkDepthSpin() and coherency;
Improved management of AutoContactTerminal for VH gauges (real ones). * New: In Anabatic & Katana, add the new "drag" feature. With VH gauges used by real technologies (M1-H, M2-V, M3-H) a new routing configuration that was not efficiently handled did appear. While the preferred routing direction for metal1 is officially horizontal, due to the way the standard cell must be designed, their metal1 terminals are still verticals (or punctuals). Thus, when connecting to them, we face the case where the metal1 terminal (RoutingPad) is vertical *and* the metal2 wire is also vertical. With that setup, the position of the AutoContactTerminal via12 cannot be deduced, it may range all the way over the metal1 RoutingPad. What may define it's position is the metal3 the metal2 finally connects to. That, is, when we have one horizontal (the metal3) and one vertical (the metal1 RoutingPad). The intermediate wire of metal2 can be kept to a minimum size by "dragging" the via12 close to the via23 when the metal3 wire is moved. * New: In Anabatic & Katana, problem of closely vertically aligneds RoutingPads in metal1 is managed first in PreProcess by restricting the span of the connecteds metal3 and in _makeDogleg also by restricting the span even more tightly (to the RoutingPad itself). * New: In Anabatic::AutoContactTerminal, add the "drag" support. Automatically check if the connecting segment is in the same direction as the RoutingPad, if so, sets the "SegDrag" flag. The dragging state can be known with the "::canDrag()" predicate. * New: In Anabatic::AutoHorizontal, add the "drag" support. The drag state can be known with the "::isDrag()" predicate. In "::_makeDogleg()", when making a dogleg on a dragable segment pass the drag state correctly and restrict the perpandicular span of the perpandicular to the RoutingPad (though segment user constraints). If we make a dogleg on the metal2 is it likely than we cannot go straigth out vertically from the RoutingPad, so the new perpandicular *is* restricted to the RoutingPad span. Idem for AutoVertical. * New: In Katana::Manipulator, add method "::dragMinimize()" which find a hole where to minimize a draggable segment. We finally did not use it, but keep it for potential further use. * New: In Katana::PreProcess, adds a "protectAlignedaccesses()" local function to check for vertically aligned metal1 RoutingPads, in that case setup user constraints on the metal3 segments so they cannot completly cover the other RoutingPad with metal2. We also keep a "metal2protect()" function that create a fixed segment to lock/protect a RoutingPad. Not used for now. * New: In Katana::Session, add a RoutingPad locking event mechanism. This allows us to request the creation of a locking (fixed segment) over a draggable segment. Not used for now. Lock events are processeds before all others as they create new TrackElements. * New: In Katana::Track, "::getNextFree()" and "::getPreviousFree()" method to find the nearest free interval in a Track after/before a position. * Bug: In Anabatic::AutoHorizontal::getConstraints(), merge with user constraints *only* if it's not an empty interval (as we use min/max functions). Idem for AutoVertical. * Bug: In AutoSegments_OnContacts::Locator::isValid(), the boolean test must be inverted. Seems it never worked, but we never used it until now...
2018-01-25 04:58:04 -06:00
if (isDrag() xor (getAutoSource()->canDrag() or getAutoTarget()->canDrag())) {
cerr << Error( "%s\n"
" Discrepency between segment \"drag\" state and it's contacts.\n"
" source:%s\n"
" target:%s"
, getString(this).c_str()
, getString(getAutoSource()).c_str()
, getString(getAutoTarget()).c_str()
) << endl;
coherency = false;
}
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
return coherency;
}
string AutoSegment::_getStringFlags () const
{
string state;
In Anabatic/Katana, add support for VH gauges (real technos). * Change: In Anabatic::AutoContactTerminal::getNativeConstraintBox(), when the anchor is a RoutingPad (which must be always the case), perform the true computation of it's position based on the segment occurrence. It is a important change, previously the area was in fact the "center line" of the connector while now it is really an area (mandatory for "half-offgrid" terminals of real technologies). The change is not complete yet, the area should be shrinked by the half size of a VIA, because the area applies to the center coordinate of the VIA (to be done quickly). * Bug: In Anabatic::AutoContactTurn::updateTopology(), when a dogleg is created (restore connexity after a layer change) the layer of the VIA, based on the segments it connects to must be re-computed *after* the dogleg has been made. * Change: In all files of Anabatic, when comparing two layers, no longer use the Layer pointer itself, but the layer mask. This allow a transparent management of both real and symbolic layers (which do share the same mask). Real metal layers (not VIAs) will be BasicLayer and symbolic metal layers will be RegularLayer. * New: Anabatic::Configuration::selectRpComponent(), select the best RoutingPad component for metal1 terminals. Look for the metal1 component with the biggest accessibility on-grid. RoutingPad using other metals are left untoucheds. * New: New function Anabatic::Vertex::getNeighbor(Edge*) to get the neighbor Vertex through an Edge*. This method allows to write clearer code as we no longer need to access the neighbor through the underlying GCell. Also add proxies for GCell methods in Vertex. * Bug: In Anabatic::Dijkstra::_toSources(), in the ripup stage, when a component with multiples vertexes is reached *and* two of it's vertexes are reached *at the same time* (one from which we backtrack and one still in the queue) extraneous edges may be created by _materialize(). Case occurs on snx/c35b4, "abc_5360_n903_1". To solve this, Dijkstra::_toSource() is modificated, the "from" edges of the newly reacheds vertexes are reset to NULL, *except* for the one we will be backtracking from. That is, the one given in the source argument. * Change: In Anabatic::NetBuilder class, put the various Hooks and RoutingPad sorting functions as class ones. * Bug: In AutoSegment::setLayer(), raise the SegInvalidatedFayer flag. This unset flag was causing AutoContactTurn::updateTopology() to not work as expected and making gaps, this was the cause of the last remaining warnings about layer connexity.
2018-01-06 09:55:53 -06:00
state += isFixed () ?" F":" -";
Basic support for FreePDK 45 completed. * New: In Commons, inspector support for std::pair<T,U>. * New: In Hurricane::Layer, ContactLayer & ViaLayer, support for non square VIAs. The hole (cut) remains square, but the various metal extensions can now be different in X and Y. The ::getEnclosure() method now takes a flag EnclosureH / EnclosureV. * New: In Hurricane::DbU, inspector support for: std::pair<DbU::Unit,DbU::Unit> std::array<DbU::Unit,3> Must be defined here as DbU do not exists yet in Commons.h * Bug: In Hurricane::Interval::getSize(), when the interval is "full span", do not return the difference between min and max, but directly DbU::Max. (the previous result was -1 !) * New: In CRL Core Python/Technology.py, support for non square VIAs in the configuration files. Applied to FreePDK 45. * New: In CRL::RoutingGauge, added a "symbolic" flag to tell if a gauge is for symbolic layout or not. Exported to Python. * New: In Anabatic::AutoHorizontal::updatePosition(), differentiated computation for soure or target taking account of the VIA extension in the right segment metal (due to non-square VIAs). * Change: In Anabatic::AutoHorizontal::_makeDogleg(), the dogleg is UP for HV gauges and DOWN for VH. * New: In Anabatic::AutoSegment::_initialize(), create a cache of the various extension length for each layer (viaToTop, viaToBottom, viaToSame). New implementation of getExtensionCap() using the previous cached extension table. See updatePositions(). New static functions to access the extension cache in the header: getViaTotopCap() ... * Change: In Anabatic::AutoSegment, in various update methods, updateOrient() must always be called *before* updatePositions() as extensions are dependant on source/target. * New: In Anabatic::AutoSegment::getEndAxes() compute the position of the first source and last target position (center/axes) on an *aligned* set of segments. * New: In Anabatic::AutoSegment, add a new state flag SegAxisFixed to signal segments that can be put on only one track. Specific case to VH gauge for a M1 vertical terminal with a M2 vertical segment. The M2 is effectively bound to the M1 axis position. * Bug: In Anabatic::NetBuilderVH::_do_xG_xM1_xM3(), in case of E/W global and only one RoutingPad the connexion to the RoutingPad was duplicated. It was valid, but totally stupid. * Bug: In Anabatic::Session::_canonize(), for an aligned segment set, intersect the user constraints from all segments instead of only considering the canonical one. Issue a warning about too tight constraints only for symbolic gauges. It may be correct for the real ones. * New: In Katata::DataNegociate::update(), more accurate computation of the perpandicular free interval. Use segment extension cap calculation. Create a special case for fixed axis segments allowing them to find alternative free interval, try under source and under target as they are likely to be draggable segments. * Change: In Katana::Manipulator::relax(), use the extension cap value to compute the axis of the perpandicular segemnts. * Change: In Katana::Manipulator::moveUp(), now move up the whole set of aligned segments instead of just the canonical one. * Change: In Katana::NegociateWindow::loadRoutingPads(), more accurate TrackMarkers insertions for fixed terminals. * New: In Katana::RoutingEvent::Key::Compare::operator(), segments with fixed axis are processed prior to any others. * New: In Katana::RoutingEventLoop, store segment pointers instead of ids to generate more accurate error messages. * Change: In Katana::RoutingPlane::create(), perform local track assignment only for HV gauges. * Change: In Katana::SegmentFsm::_slackenLocal(), add a "dragMinimize" step in the automaton. Mutliple states transitions can occurs in a row if an action fails. * New: In Katana::Session::_toIntervalAxis(), normalize interval bounds so they are on track positions (by shrinking the interval). * Bug: In Katana::TrackMarker CTOR, the weigh computation was wrong.
2018-02-17 13:27:38 -06:00
state += isFixedAxis () ? "X": "-";
In Anabatic/Katana, add support for VH gauges (real technos). * Change: In Anabatic::AutoContactTerminal::getNativeConstraintBox(), when the anchor is a RoutingPad (which must be always the case), perform the true computation of it's position based on the segment occurrence. It is a important change, previously the area was in fact the "center line" of the connector while now it is really an area (mandatory for "half-offgrid" terminals of real technologies). The change is not complete yet, the area should be shrinked by the half size of a VIA, because the area applies to the center coordinate of the VIA (to be done quickly). * Bug: In Anabatic::AutoContactTurn::updateTopology(), when a dogleg is created (restore connexity after a layer change) the layer of the VIA, based on the segments it connects to must be re-computed *after* the dogleg has been made. * Change: In all files of Anabatic, when comparing two layers, no longer use the Layer pointer itself, but the layer mask. This allow a transparent management of both real and symbolic layers (which do share the same mask). Real metal layers (not VIAs) will be BasicLayer and symbolic metal layers will be RegularLayer. * New: Anabatic::Configuration::selectRpComponent(), select the best RoutingPad component for metal1 terminals. Look for the metal1 component with the biggest accessibility on-grid. RoutingPad using other metals are left untoucheds. * New: New function Anabatic::Vertex::getNeighbor(Edge*) to get the neighbor Vertex through an Edge*. This method allows to write clearer code as we no longer need to access the neighbor through the underlying GCell. Also add proxies for GCell methods in Vertex. * Bug: In Anabatic::Dijkstra::_toSources(), in the ripup stage, when a component with multiples vertexes is reached *and* two of it's vertexes are reached *at the same time* (one from which we backtrack and one still in the queue) extraneous edges may be created by _materialize(). Case occurs on snx/c35b4, "abc_5360_n903_1". To solve this, Dijkstra::_toSource() is modificated, the "from" edges of the newly reacheds vertexes are reset to NULL, *except* for the one we will be backtracking from. That is, the one given in the source argument. * Change: In Anabatic::NetBuilder class, put the various Hooks and RoutingPad sorting functions as class ones. * Bug: In AutoSegment::setLayer(), raise the SegInvalidatedFayer flag. This unset flag was causing AutoContactTurn::updateTopology() to not work as expected and making gaps, this was the cause of the last remaining warnings about layer connexity.
2018-01-06 09:55:53 -06:00
state += isUnsetAxis () ? "u": "-";
state += isStrap () ? "S": "-";
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
state += isUnbreakable () ? "U": "-";
In Anabatic/Katana, add support for VH gauges (real technos). * Change: In Anabatic::AutoContactTerminal::getNativeConstraintBox(), when the anchor is a RoutingPad (which must be always the case), perform the true computation of it's position based on the segment occurrence. It is a important change, previously the area was in fact the "center line" of the connector while now it is really an area (mandatory for "half-offgrid" terminals of real technologies). The change is not complete yet, the area should be shrinked by the half size of a VIA, because the area applies to the center coordinate of the VIA (to be done quickly). * Bug: In Anabatic::AutoContactTurn::updateTopology(), when a dogleg is created (restore connexity after a layer change) the layer of the VIA, based on the segments it connects to must be re-computed *after* the dogleg has been made. * Change: In all files of Anabatic, when comparing two layers, no longer use the Layer pointer itself, but the layer mask. This allow a transparent management of both real and symbolic layers (which do share the same mask). Real metal layers (not VIAs) will be BasicLayer and symbolic metal layers will be RegularLayer. * New: Anabatic::Configuration::selectRpComponent(), select the best RoutingPad component for metal1 terminals. Look for the metal1 component with the biggest accessibility on-grid. RoutingPad using other metals are left untoucheds. * New: New function Anabatic::Vertex::getNeighbor(Edge*) to get the neighbor Vertex through an Edge*. This method allows to write clearer code as we no longer need to access the neighbor through the underlying GCell. Also add proxies for GCell methods in Vertex. * Bug: In Anabatic::Dijkstra::_toSources(), in the ripup stage, when a component with multiples vertexes is reached *and* two of it's vertexes are reached *at the same time* (one from which we backtrack and one still in the queue) extraneous edges may be created by _materialize(). Case occurs on snx/c35b4, "abc_5360_n903_1". To solve this, Dijkstra::_toSource() is modificated, the "from" edges of the newly reacheds vertexes are reset to NULL, *except* for the one we will be backtracking from. That is, the one given in the source argument. * Change: In Anabatic::NetBuilder class, put the various Hooks and RoutingPad sorting functions as class ones. * Bug: In AutoSegment::setLayer(), raise the SegInvalidatedFayer flag. This unset flag was causing AutoContactTurn::updateTopology() to not work as expected and making gaps, this was the cause of the last remaining warnings about layer connexity.
2018-01-06 09:55:53 -06:00
state += isCanonical () ? "C": "-";
state += isGlobal () ? "G": "-";
state += isWeakGlobal () ? "g": "-";
state += isLongLocal () ? "L": "-";
state += isStrongTerminal () ? "T": "-";
Improved management of AutoContactTerminal for VH gauges (real ones). * New: In Anabatic & Katana, add the new "drag" feature. With VH gauges used by real technologies (M1-H, M2-V, M3-H) a new routing configuration that was not efficiently handled did appear. While the preferred routing direction for metal1 is officially horizontal, due to the way the standard cell must be designed, their metal1 terminals are still verticals (or punctuals). Thus, when connecting to them, we face the case where the metal1 terminal (RoutingPad) is vertical *and* the metal2 wire is also vertical. With that setup, the position of the AutoContactTerminal via12 cannot be deduced, it may range all the way over the metal1 RoutingPad. What may define it's position is the metal3 the metal2 finally connects to. That, is, when we have one horizontal (the metal3) and one vertical (the metal1 RoutingPad). The intermediate wire of metal2 can be kept to a minimum size by "dragging" the via12 close to the via23 when the metal3 wire is moved. * New: In Anabatic & Katana, problem of closely vertically aligneds RoutingPads in metal1 is managed first in PreProcess by restricting the span of the connecteds metal3 and in _makeDogleg also by restricting the span even more tightly (to the RoutingPad itself). * New: In Anabatic::AutoContactTerminal, add the "drag" support. Automatically check if the connecting segment is in the same direction as the RoutingPad, if so, sets the "SegDrag" flag. The dragging state can be known with the "::canDrag()" predicate. * New: In Anabatic::AutoHorizontal, add the "drag" support. The drag state can be known with the "::isDrag()" predicate. In "::_makeDogleg()", when making a dogleg on a dragable segment pass the drag state correctly and restrict the perpandicular span of the perpandicular to the RoutingPad (though segment user constraints). If we make a dogleg on the metal2 is it likely than we cannot go straigth out vertically from the RoutingPad, so the new perpandicular *is* restricted to the RoutingPad span. Idem for AutoVertical. * New: In Katana::Manipulator, add method "::dragMinimize()" which find a hole where to minimize a draggable segment. We finally did not use it, but keep it for potential further use. * New: In Katana::PreProcess, adds a "protectAlignedaccesses()" local function to check for vertically aligned metal1 RoutingPads, in that case setup user constraints on the metal3 segments so they cannot completly cover the other RoutingPad with metal2. We also keep a "metal2protect()" function that create a fixed segment to lock/protect a RoutingPad. Not used for now. * New: In Katana::Session, add a RoutingPad locking event mechanism. This allows us to request the creation of a locking (fixed segment) over a draggable segment. Not used for now. Lock events are processeds before all others as they create new TrackElements. * New: In Katana::Track, "::getNextFree()" and "::getPreviousFree()" method to find the nearest free interval in a Track after/before a position. * Bug: In Anabatic::AutoHorizontal::getConstraints(), merge with user constraints *only* if it's not an empty interval (as we use min/max functions). Idem for AutoVertical. * Bug: In AutoSegments_OnContacts::Locator::isValid(), the boolean test must be inverted. Seems it never worked, but we never used it until now...
2018-01-25 04:58:04 -06:00
state += isDrag () ? "D": "-";
In Anabatic/Katana, add support for VH gauges (real technos). * Change: In Anabatic::AutoContactTerminal::getNativeConstraintBox(), when the anchor is a RoutingPad (which must be always the case), perform the true computation of it's position based on the segment occurrence. It is a important change, previously the area was in fact the "center line" of the connector while now it is really an area (mandatory for "half-offgrid" terminals of real technologies). The change is not complete yet, the area should be shrinked by the half size of a VIA, because the area applies to the center coordinate of the VIA (to be done quickly). * Bug: In Anabatic::AutoContactTurn::updateTopology(), when a dogleg is created (restore connexity after a layer change) the layer of the VIA, based on the segments it connects to must be re-computed *after* the dogleg has been made. * Change: In all files of Anabatic, when comparing two layers, no longer use the Layer pointer itself, but the layer mask. This allow a transparent management of both real and symbolic layers (which do share the same mask). Real metal layers (not VIAs) will be BasicLayer and symbolic metal layers will be RegularLayer. * New: Anabatic::Configuration::selectRpComponent(), select the best RoutingPad component for metal1 terminals. Look for the metal1 component with the biggest accessibility on-grid. RoutingPad using other metals are left untoucheds. * New: New function Anabatic::Vertex::getNeighbor(Edge*) to get the neighbor Vertex through an Edge*. This method allows to write clearer code as we no longer need to access the neighbor through the underlying GCell. Also add proxies for GCell methods in Vertex. * Bug: In Anabatic::Dijkstra::_toSources(), in the ripup stage, when a component with multiples vertexes is reached *and* two of it's vertexes are reached *at the same time* (one from which we backtrack and one still in the queue) extraneous edges may be created by _materialize(). Case occurs on snx/c35b4, "abc_5360_n903_1". To solve this, Dijkstra::_toSource() is modificated, the "from" edges of the newly reacheds vertexes are reset to NULL, *except* for the one we will be backtracking from. That is, the one given in the source argument. * Change: In Anabatic::NetBuilder class, put the various Hooks and RoutingPad sorting functions as class ones. * Bug: In AutoSegment::setLayer(), raise the SegInvalidatedFayer flag. This unset flag was causing AutoContactTurn::updateTopology() to not work as expected and making gaps, this was the cause of the last remaining warnings about layer connexity.
2018-01-06 09:55:53 -06:00
state += isWeakTerminal1 () ? "W": "-";
state += isWeakTerminal2 () ? "w": "-";
state += isNotAligned () ? "A": "-";
state += isSlackened () ? "S": "-";
state += isReduced () ? "r": "-";
state += isInvalidated () ? "i": "-";
state += isInvalidatedLayer() ? "l": "-";
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
Redesigning access to analog terminals in devices. * New: In Anabatic::AutoContact, added forceOnGrid() method to allow a AutoContactTerminal to be set (and fixed) outside it's constraint Box. Be aware that this may generate gaps in the routing. Stored using the UserNativeConstraints mechanism. * New: In Anabatic::Session() new method getNearestgridpoint() to easily compute the nearest grid point from a position. * Change: In Anabatic::AutoSegment::_revalidate(), use layer mask to compute the spin of the segment instead of the layers (more robust for analogic). * New: In ::GCellTopology::doRp_Access() new method to create terminal contacts in _doDevice(). Issue more warnings and force the terminal to on grid (even if it create gaps). This a temporary hack.
2016-10-05 16:43:07 -05:00
if (_flags & SegSourceTop) state += 't';
else if (_flags & SegSourceBottom) state += 'b';
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
else state += '-';
Redesigning access to analog terminals in devices. * New: In Anabatic::AutoContact, added forceOnGrid() method to allow a AutoContactTerminal to be set (and fixed) outside it's constraint Box. Be aware that this may generate gaps in the routing. Stored using the UserNativeConstraints mechanism. * New: In Anabatic::Session() new method getNearestgridpoint() to easily compute the nearest grid point from a position. * Change: In Anabatic::AutoSegment::_revalidate(), use layer mask to compute the spin of the segment instead of the layers (more robust for analogic). * New: In ::GCellTopology::doRp_Access() new method to create terminal contacts in _doDevice(). Issue more warnings and force the terminal to on grid (even if it create gaps). This a temporary hack.
2016-10-05 16:43:07 -05:00
if (_flags & SegTargetTop) state += 't';
else if (_flags & SegTargetBottom) state += 'b';
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
else state += '-';
Improved handling of short nets (fully included in one GCell). The short net mode degrade the routing in some cases. This will be fixed in a next batch of commits. * New: In Hurricane::NetRoutingProperty, added "ShortNet" flag for Nets that are completly inside *one* GCell. * Bug: In CRL::BlifParser::Model::staticInit(), when looking for the output of zero and one cell, also skip the blockage net (as well as automatic and supplies). * New: In Anabatic::AutoSegment, added "ShortNet" flag to know if the segment is part of a short net (fully included in *one* GCell). Also add accessor/mutators for the _analogMode flag (was it ever used before?). * New: In Anabatic::NetBuilder::singleGCell(), if a RoutingPad is vertically small, add a vertical segment to give it some slack. * New: In Anabatic::Dijkstra::_materialize(), detect "short net" as they have only one GCell in their source list... * Bug: In AnabaticEngine::_loadGrbyNet(), reset the AutoSegment "short net" and "analog mode" creation flags between two different nets. * New: In Katana::Configuration, added dedicated ripup for short net segmnts. * New: In Katana: partially implemented support for "short dogleg", that is dogleg that are always kept in same metal because they connect neighboring perpandicular tracks. Not finished neither activated yet. * New: In Katana::TreckElement and derived, export the the *short net* support from AutoSegment. * Bug: In Katana::RoutingEvent::_processRepair(), when a segment is successfully inserted, re-process any perpandicular that is in repair state, as it may have a new chance to be placed. * New: In Katana::SegmentFsm::slackenTopology(), always reject short nets. * Bug: In Katana::Track::check(), correctly handle wide segments instead of issuing false check messages.
2018-07-16 04:16:51 -05:00
state += isShortNet () ? "s": "-";
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
return state;
}
string AutoSegment::_getString () const
{
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
string sdistance = " rpD:" + getString(_rpDistance);
string s = base()->_getString();
s.insert ( s.size()-1, sdistance );
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
s.insert ( s.size()-1, _getStringFlags() );
return s;
}
Record* AutoSegment::_getRecord () const
{
Record* record = base()->_getRecord ();
Basic support for FreePDK 45 completed. * New: In Commons, inspector support for std::pair<T,U>. * New: In Hurricane::Layer, ContactLayer & ViaLayer, support for non square VIAs. The hole (cut) remains square, but the various metal extensions can now be different in X and Y. The ::getEnclosure() method now takes a flag EnclosureH / EnclosureV. * New: In Hurricane::DbU, inspector support for: std::pair<DbU::Unit,DbU::Unit> std::array<DbU::Unit,3> Must be defined here as DbU do not exists yet in Commons.h * Bug: In Hurricane::Interval::getSize(), when the interval is "full span", do not return the difference between min and max, but directly DbU::Max. (the previous result was -1 !) * New: In CRL Core Python/Technology.py, support for non square VIAs in the configuration files. Applied to FreePDK 45. * New: In CRL::RoutingGauge, added a "symbolic" flag to tell if a gauge is for symbolic layout or not. Exported to Python. * New: In Anabatic::AutoHorizontal::updatePosition(), differentiated computation for soure or target taking account of the VIA extension in the right segment metal (due to non-square VIAs). * Change: In Anabatic::AutoHorizontal::_makeDogleg(), the dogleg is UP for HV gauges and DOWN for VH. * New: In Anabatic::AutoSegment::_initialize(), create a cache of the various extension length for each layer (viaToTop, viaToBottom, viaToSame). New implementation of getExtensionCap() using the previous cached extension table. See updatePositions(). New static functions to access the extension cache in the header: getViaTotopCap() ... * Change: In Anabatic::AutoSegment, in various update methods, updateOrient() must always be called *before* updatePositions() as extensions are dependant on source/target. * New: In Anabatic::AutoSegment::getEndAxes() compute the position of the first source and last target position (center/axes) on an *aligned* set of segments. * New: In Anabatic::AutoSegment, add a new state flag SegAxisFixed to signal segments that can be put on only one track. Specific case to VH gauge for a M1 vertical terminal with a M2 vertical segment. The M2 is effectively bound to the M1 axis position. * Bug: In Anabatic::NetBuilderVH::_do_xG_xM1_xM3(), in case of E/W global and only one RoutingPad the connexion to the RoutingPad was duplicated. It was valid, but totally stupid. * Bug: In Anabatic::Session::_canonize(), for an aligned segment set, intersect the user constraints from all segments instead of only considering the canonical one. Issue a warning about too tight constraints only for symbolic gauges. It may be correct for the real ones. * New: In Katata::DataNegociate::update(), more accurate computation of the perpandicular free interval. Use segment extension cap calculation. Create a special case for fixed axis segments allowing them to find alternative free interval, try under source and under target as they are likely to be draggable segments. * Change: In Katana::Manipulator::relax(), use the extension cap value to compute the axis of the perpandicular segemnts. * Change: In Katana::Manipulator::moveUp(), now move up the whole set of aligned segments instead of just the canonical one. * Change: In Katana::NegociateWindow::loadRoutingPads(), more accurate TrackMarkers insertions for fixed terminals. * New: In Katana::RoutingEvent::Key::Compare::operator(), segments with fixed axis are processed prior to any others. * New: In Katana::RoutingEventLoop, store segment pointers instead of ids to generate more accurate error messages. * Change: In Katana::RoutingPlane::create(), perform local track assignment only for HV gauges. * Change: In Katana::SegmentFsm::_slackenLocal(), add a "dragMinimize" step in the automaton. Mutliple states transitions can occurs in a row if an action fails. * New: In Katana::Session::_toIntervalAxis(), normalize interval bounds so they are on track positions (by shrinking the interval). * Bug: In Katana::TrackMarker CTOR, the weigh computation was wrong.
2018-02-17 13:27:38 -06:00
record->add ( getSlot ( "_extensionCaps" , &_extensionCaps ) );
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
record->add ( getSlot ( "_gcell" , _gcell ) );
record->add ( getSlot ( "_id" , &_id ) );
record->add ( getSlot ( "_flags" , &_flags ) );
record->add ( getSlot ( "_userContraints", &_userConstraints ) );
record->add ( getSlot ( "_sourcePosition", &_sourcePosition ) );
record->add ( getSlot ( "_targetPosition", &_targetPosition ) );
record->add ( getSlot ( "_parent" , _parent ) );
return record;
}
AutoSegment* AutoSegment::create ( AutoContact* source
, AutoContact* target
, Segment* hurricaneSegment
)
{
Added support for 2-Metal block routing in Anabatic & Katana. * New: In AnabaticEngine::invalidateRoutingPads() this method is a temporary workaround for a Hurricane problems. When an instance is moved, the RoutingPads that use it must be moved accordingly, but they are not invalidated so they stay in the wrong QuadTree. New method ::_resizeMatrix() to be called when the associated Cell is resized. * Bug: In AutoHorizontal::getConstraints() and AutoVertical::getConstraints(), the *target* constraints where never merged. * Change: In AutoHorizontal::getCells() and AutoVertical::getGCells(), now return a boolean to tell if it was ok (must not encounter a NULL GCell while progessing from source to target). * New: In Anabatic::Configuration and Anabatic:Session, create new methods: - getDHorizontalLayer() - getDhorizontalDepth() - getDHorizontalWidth() - getDHorizontalPitch() And so on for Vertical and Contact. They supply depth-independant informations about the H/V layers to build the initial detailed routing. The AutoSegment::create() methods have been modificated accordingly. * New: In Anabatic::GCell, add two new types "StdCellRow" and "ChannelRow" for implementing 2-Metal blocks. Rename the GCell::setXY() method in GCell::setSouthWestCorner(), move the contents of GCell::updateContactsPosition() into it and suppress it. WARNING: In case of a GCell shrink this may cause problems. But for now we only expand... New method GCell::getNetCount() to count the number of Net going though the GCell. * Change: In Anabatic::Edge, add specific support for capacity of 2-Metal routing channels. * Change: In Anabatic::Dijsktra various methods, replace the "gcell->isMatrix()" calls by "not gcell->isAnalog()". Add more check so that the methods pertaining to the analog routing (GRData) are not called in digital mode. * New: In Anabatic::Dijkstra::materialize(), add support for 2-Metal specific cases. That is, always break in case of vertical pass-through or U-turn. The global routing must always be broken in H-Channel. * New: In Anabatic::GCell & Anabatic::Edge, make use of the Session mechanism to ensure the revalidation. The "::revalidate()" method is then moved as "::materialize()" (overload of Go) and "::_invalidate()" becomes "::invalidate()" * Change: In LoadGlobalRouting, cosmetic rename of SortHkByX in SortHookByX. * New: In GCellTopology, added support for building 2-Metal topologies. * ForkStack is now an object attribute as many methods do need it. * To push segments/hook on the stack, a new method "push()" is available. Perform NULL and fromHook checking. Can also setup _southWestContact or _northEastContact if it is the "from" edge. * N/S/E/W edges are now vector as in digital channel mode there can be more than one. * Added build topological build methods: - doRp_2m_Access() RoutingPad stem access. - _do_2m_1G_1M1() North or south access. - _do_2m_2G_1M1() North AND south access. - _do_2m_xG() H-Channel routing. * New: In Anabatic::Matrix, new ::resize() function, as Cell can be resizeds. * New: In Anabatic::Vertex, new static method ::getValueString() for a friendly text rendering. * New: In Katana::DigitalDistance, support for channel routing. * Change: In KatanaEngine::digitalSetup() and KatanaEngine::runGlobalrouter(), for channel routing, calls to setupPowerRails() and protectRoutingPads() must be called after the core block has been fully dimensionned. ::runGlobalrouter() contains the code tasked with the grid creation and channel sizing. * New: In KatanaEngine: Added support for core block, for 2-Metal routing. May be expanded for over-the-cell routing in the future. Added methods : - isDigitalMode() - isAnalogMode() - isMixedMode() - isChannelMode() - getBlock() / addBlock() - setupChannelMode() - createChannel() * New: In Katana, new class Block to manage core blocks and perform channel routing. * New: In Katana::Session, new convenience method "isOpen()".
2017-08-18 16:56:23 -05:00
const Layer* horizontalLayer = Session::getDHorizontalLayer();
DbU::Unit horizontalWidth = Session::getDHorizontalWidth();
const Layer* verticalLayer = Session::getDVerticalLayer();
DbU::Unit verticalWidth = Session::getDVerticalWidth();
Katana manage wide wires, and they can also be symmetric. * New: In Anabatic::AutoContact and the derived classes, manages wide wires. The contact self dimension itself according to the segments it is connected to. Special case for the AutoContactTerminal which also read the size of the component it is anchored upon. New refresh method "updateSize()" and flag CntInvalidatedWidth. to compute the size. In AutoContactTerminal, compute the constraint box according to the width of the segment. * New: In Anabatic::AutoSegment, flags are now implemented as "static const" attributes of the class. The flags are stored into a uint64_t as they are more than 32. Added new flag "SegWide" and associated predicates. * Change: In GCellTopology::_doHChannel() and GCellTopology::_doVChannel(), uses the simpler overload of AutoSegment::create() in order to detect the wire width automatically. * New: In Katana::Manipulator, split insertToTrack() and forceToTrack() into a one-track method and a segment level method that iterate over the track span of the segment. * New: In Katana::SegmentFsm, for each cost in the table, now allow access to a specific track. So the base functions have now two parameters: "icost" and "itrack" (has a cost can have multiple tracks in the case of wide segments). * Change: In Katana::TrackElement, remove the index of the element inside it's track, as for a wide segment it will not be meaningful for the non-base track. This means that we have to use the Track::find() method each time instead. Remove the wide flag, as it is a duplicate of the one in AutoSegment. Added a getTrackCount() method to tell the number of track the segment is inserted into. Needed in the Track destroy step to delete a segment only when the last track that refers it is destroyed. Added getSymmetricAxis() to correct the computation of the symmetric base track in case of wide segment as the base track is not centered but the the leftmost one. * Change: In Track::insert() insert wide segments in their whole track span. * Change: In TrackCost, create an array of costs according to the segment track span. * Change: In TrackSegment::create(), now activate the factory and create wide segments. * Bug: In Katana::AutoSegments_Perpandicular, correct the debug indentation problem (ever shifting to the right).
2017-07-28 08:30:22 -05:00
uint32_t wPitch = NetRoutingExtension::getWPitch( source->getNet() );
if (wPitch > 1) {
Added support for 2-Metal block routing in Anabatic & Katana. * New: In AnabaticEngine::invalidateRoutingPads() this method is a temporary workaround for a Hurricane problems. When an instance is moved, the RoutingPads that use it must be moved accordingly, but they are not invalidated so they stay in the wrong QuadTree. New method ::_resizeMatrix() to be called when the associated Cell is resized. * Bug: In AutoHorizontal::getConstraints() and AutoVertical::getConstraints(), the *target* constraints where never merged. * Change: In AutoHorizontal::getCells() and AutoVertical::getGCells(), now return a boolean to tell if it was ok (must not encounter a NULL GCell while progessing from source to target). * New: In Anabatic::Configuration and Anabatic:Session, create new methods: - getDHorizontalLayer() - getDhorizontalDepth() - getDHorizontalWidth() - getDHorizontalPitch() And so on for Vertical and Contact. They supply depth-independant informations about the H/V layers to build the initial detailed routing. The AutoSegment::create() methods have been modificated accordingly. * New: In Anabatic::GCell, add two new types "StdCellRow" and "ChannelRow" for implementing 2-Metal blocks. Rename the GCell::setXY() method in GCell::setSouthWestCorner(), move the contents of GCell::updateContactsPosition() into it and suppress it. WARNING: In case of a GCell shrink this may cause problems. But for now we only expand... New method GCell::getNetCount() to count the number of Net going though the GCell. * Change: In Anabatic::Edge, add specific support for capacity of 2-Metal routing channels. * Change: In Anabatic::Dijsktra various methods, replace the "gcell->isMatrix()" calls by "not gcell->isAnalog()". Add more check so that the methods pertaining to the analog routing (GRData) are not called in digital mode. * New: In Anabatic::Dijkstra::materialize(), add support for 2-Metal specific cases. That is, always break in case of vertical pass-through or U-turn. The global routing must always be broken in H-Channel. * New: In Anabatic::GCell & Anabatic::Edge, make use of the Session mechanism to ensure the revalidation. The "::revalidate()" method is then moved as "::materialize()" (overload of Go) and "::_invalidate()" becomes "::invalidate()" * Change: In LoadGlobalRouting, cosmetic rename of SortHkByX in SortHookByX. * New: In GCellTopology, added support for building 2-Metal topologies. * ForkStack is now an object attribute as many methods do need it. * To push segments/hook on the stack, a new method "push()" is available. Perform NULL and fromHook checking. Can also setup _southWestContact or _northEastContact if it is the "from" edge. * N/S/E/W edges are now vector as in digital channel mode there can be more than one. * Added build topological build methods: - doRp_2m_Access() RoutingPad stem access. - _do_2m_1G_1M1() North or south access. - _do_2m_2G_1M1() North AND south access. - _do_2m_xG() H-Channel routing. * New: In Anabatic::Matrix, new ::resize() function, as Cell can be resizeds. * New: In Anabatic::Vertex, new static method ::getValueString() for a friendly text rendering. * New: In Katana::DigitalDistance, support for channel routing. * Change: In KatanaEngine::digitalSetup() and KatanaEngine::runGlobalrouter(), for channel routing, calls to setupPowerRails() and protectRoutingPads() must be called after the core block has been fully dimensionned. ::runGlobalrouter() contains the code tasked with the grid creation and channel sizing. * New: In KatanaEngine: Added support for core block, for 2-Metal routing. May be expanded for over-the-cell routing in the future. Added methods : - isDigitalMode() - isAnalogMode() - isMixedMode() - isChannelMode() - getBlock() / addBlock() - setupChannelMode() - createChannel() * New: In Katana, new class Block to manage core blocks and perform channel routing. * New: In Katana::Session, new convenience method "isOpen()".
2017-08-18 16:56:23 -05:00
horizontalWidth += (wPitch-1) * Session::getDHorizontalPitch();
verticalWidth += (wPitch-1) * Session::getDVerticalPitch ();
Katana manage wide wires, and they can also be symmetric. * New: In Anabatic::AutoContact and the derived classes, manages wide wires. The contact self dimension itself according to the segments it is connected to. Special case for the AutoContactTerminal which also read the size of the component it is anchored upon. New refresh method "updateSize()" and flag CntInvalidatedWidth. to compute the size. In AutoContactTerminal, compute the constraint box according to the width of the segment. * New: In Anabatic::AutoSegment, flags are now implemented as "static const" attributes of the class. The flags are stored into a uint64_t as they are more than 32. Added new flag "SegWide" and associated predicates. * Change: In GCellTopology::_doHChannel() and GCellTopology::_doVChannel(), uses the simpler overload of AutoSegment::create() in order to detect the wire width automatically. * New: In Katana::Manipulator, split insertToTrack() and forceToTrack() into a one-track method and a segment level method that iterate over the track span of the segment. * New: In Katana::SegmentFsm, for each cost in the table, now allow access to a specific track. So the base functions have now two parameters: "icost" and "itrack" (has a cost can have multiple tracks in the case of wide segments). * Change: In Katana::TrackElement, remove the index of the element inside it's track, as for a wide segment it will not be meaningful for the non-base track. This means that we have to use the Track::find() method each time instead. Remove the wide flag, as it is a duplicate of the one in AutoSegment. Added a getTrackCount() method to tell the number of track the segment is inserted into. Needed in the Track destroy step to delete a segment only when the last track that refers it is destroyed. Added getSymmetricAxis() to correct the computation of the symmetric base track in case of wide segment as the base track is not centered but the the leftmost one. * Change: In Track::insert() insert wide segments in their whole track span. * Change: In TrackCost, create an array of costs according to the segment track span. * Change: In TrackSegment::create(), now activate the factory and create wide segments. * Bug: In Katana::AutoSegments_Perpandicular, correct the debug indentation problem (ever shifting to the right).
2017-07-28 08:30:22 -05:00
}
cdebug_log(149,0) << "wPitch:" << wPitch << " hW:" << DbU::getValueString(horizontalWidth) << endl;
if (wPitch > 2) {
throw Error( "wPitch %d for \"%s\"", wPitch, getString(source->getNet()->getName()).c_str() );
}
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
bool reattachSource = false;
bool reattachTarget = false;
AutoSegment* segment;
Horizontal* horizontal = dynamic_cast<Horizontal*>( hurricaneSegment );
Vertical* vertical = dynamic_cast<Vertical* >( hurricaneSegment );
Validating channel routing mode (two metals) on SNX. * New: In Hurricane::Entity, add an id counter limit and a memory size limit. The two limits are checked only when a new Entity object is created. This should help avoiding massive memory links. * New: In CRL Core, add a "crlcore.groundName" and a "crlcore.powerName" parameter to specify the name of the ground/power signals to be created if they are missing in a Cell. For Alliance libraries it would be "vss" & "vdd" (default values), but for real technologies, it is often "gnd!" & "vdd!". The Blif parser is modificated to make use of it. * Bug: In AnabaticEngine::unify(), set the resulting unified segment in the center of the GCells common side. Gcells under a segment are found by using the edge that cover the segment axis. When we have a "bend" GCell stack and the axis is wrong, they could be ommited. This was causing deleted segments to be not removed from some Edges, then core dump. * Change: In Anabatic::AutoSegment::create(), smarter choosing of the reference contact, select the fixed or terminal one instead of always the source one. * New: In Anabatic::Edge::isEnding(), new function to check if a segment going through an Edge is starting/ending in either source or target GCell of the edge (active only when running in channel mode). * New: In Anabatic::Edge::add(), a segment takes part in the occupancy only if it is not ending in either source or target (channel mode only). The occupancy due to terminal is pre-computed in Katana. * New: In Anabatic::Edge::ripup(), in channel mode, never ripup a segment which is ending in either source or target (we *have* to access this edge to connect to the terminal). * Bug: In Anabatic::GCell::hcut() and vcut(), force the update of the Edge which is on the side that will get splitted by the cut. It's capacity will be reduced to it must be updated. * Change: In Anabatic::GCell::updateGContacts() add a flag to conditionnally update horizontals or verticals only. We may require only a partial update when resizing the GCell in only one direction. This, again, related to the fact that we compute the GCells under a segment thanks to it's axis position, so we need to be very careful when modificating axis. * Change: In Katana::Block::resizeChannels(), only update GContact vertical position. Do not disturb X positions of segments. * Bug: In Katana::GlobalRoute::DigitalDistance, in channel mode, some Edges can have a zero capacity, but still be reachable if the net has a terminal in either source or target. Look for this case and return a distance of zero instead of "unreachable". This was causing the global routing not to complete in channel mode. For computing the edge distance, makes the vertical edges much more long (10 times) than the horizontal ones as the vertical capacity is very limited. Hard coded for now, should make it a parameter in the future. * Change: In KatanaEngine::annotateGlobalGraph(), decrease the capacity of edges with reserveCapacity for each terminal inside a GCell. Both north and south edges are decreased as we a terminal will block both north and south edges. As a counterpart, the Edge capacity is not decreased when the global router connect to a terminal. * Change: In Katana::RoutingEvent::revalidate(), when in repair stage, do not expand the slack for horizontal segments in channel mode. So they may not overlap the standard cell row. * Bug: In Stratus documentation, do not use the french option in babel, the documentation is in english! * New: In Documentation, added Hurricane/Python tutorial, part for drawing layout.
2018-03-16 10:20:04 -05:00
AutoContact* reference = NULL;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
cdebug_log(149,0) << "Source:" << source << endl;
cdebug_log(149,0) << "Target:" << target << endl;
Validating channel routing mode (two metals) on SNX. * New: In Hurricane::Entity, add an id counter limit and a memory size limit. The two limits are checked only when a new Entity object is created. This should help avoiding massive memory links. * New: In CRL Core, add a "crlcore.groundName" and a "crlcore.powerName" parameter to specify the name of the ground/power signals to be created if they are missing in a Cell. For Alliance libraries it would be "vss" & "vdd" (default values), but for real technologies, it is often "gnd!" & "vdd!". The Blif parser is modificated to make use of it. * Bug: In AnabaticEngine::unify(), set the resulting unified segment in the center of the GCells common side. Gcells under a segment are found by using the edge that cover the segment axis. When we have a "bend" GCell stack and the axis is wrong, they could be ommited. This was causing deleted segments to be not removed from some Edges, then core dump. * Change: In Anabatic::AutoSegment::create(), smarter choosing of the reference contact, select the fixed or terminal one instead of always the source one. * New: In Anabatic::Edge::isEnding(), new function to check if a segment going through an Edge is starting/ending in either source or target GCell of the edge (active only when running in channel mode). * New: In Anabatic::Edge::add(), a segment takes part in the occupancy only if it is not ending in either source or target (channel mode only). The occupancy due to terminal is pre-computed in Katana. * New: In Anabatic::Edge::ripup(), in channel mode, never ripup a segment which is ending in either source or target (we *have* to access this edge to connect to the terminal). * Bug: In Anabatic::GCell::hcut() and vcut(), force the update of the Edge which is on the side that will get splitted by the cut. It's capacity will be reduced to it must be updated. * Change: In Anabatic::GCell::updateGContacts() add a flag to conditionnally update horizontals or verticals only. We may require only a partial update when resizing the GCell in only one direction. This, again, related to the fact that we compute the GCells under a segment thanks to it's axis position, so we need to be very careful when modificating axis. * Change: In Katana::Block::resizeChannels(), only update GContact vertical position. Do not disturb X positions of segments. * Bug: In Katana::GlobalRoute::DigitalDistance, in channel mode, some Edges can have a zero capacity, but still be reachable if the net has a terminal in either source or target. Look for this case and return a distance of zero instead of "unreachable". This was causing the global routing not to complete in channel mode. For computing the edge distance, makes the vertical edges much more long (10 times) than the horizontal ones as the vertical capacity is very limited. Hard coded for now, should make it a parameter in the future. * Change: In KatanaEngine::annotateGlobalGraph(), decrease the capacity of edges with reserveCapacity for each terminal inside a GCell. Both north and south edges are decreased as we a terminal will block both north and south edges. As a counterpart, the Edge capacity is not decreased when the global router connect to a terminal. * Change: In Katana::RoutingEvent::revalidate(), when in repair stage, do not expand the slack for horizontal segments in channel mode. So they may not overlap the standard cell row. * Bug: In Stratus documentation, do not use the french option in babel, the documentation is in english! * New: In Documentation, added Hurricane/Python tutorial, part for drawing layout.
2018-03-16 10:20:04 -05:00
if (source->isFixed() and target->isFixed()) {
if ( (horizontal) and (source->getY() != target->getY()))
cerr << Warning( "Straight AutoHorizontal connecting misaligned contacts:\n"
" %s\n"
" %s"
, getString(source).c_str()
, getString(target).c_str()
) << endl;
if ( (vertical) and (source->getX() != target->getX()))
cerr << Warning( "Straight AutoVertical connecting misaligned contacts:\n"
" %s\n"
" %s"
, getString(source).c_str()
, getString(target).c_str()
) << endl;
}
if (target->isFixed() or target->isTerminal()) reference = target;
if (source->isFixed() or source->isTerminal()) reference = source;
Added support for 2-Metal block routing in Anabatic & Katana. * New: In AnabaticEngine::invalidateRoutingPads() this method is a temporary workaround for a Hurricane problems. When an instance is moved, the RoutingPads that use it must be moved accordingly, but they are not invalidated so they stay in the wrong QuadTree. New method ::_resizeMatrix() to be called when the associated Cell is resized. * Bug: In AutoHorizontal::getConstraints() and AutoVertical::getConstraints(), the *target* constraints where never merged. * Change: In AutoHorizontal::getCells() and AutoVertical::getGCells(), now return a boolean to tell if it was ok (must not encounter a NULL GCell while progessing from source to target). * New: In Anabatic::Configuration and Anabatic:Session, create new methods: - getDHorizontalLayer() - getDhorizontalDepth() - getDHorizontalWidth() - getDHorizontalPitch() And so on for Vertical and Contact. They supply depth-independant informations about the H/V layers to build the initial detailed routing. The AutoSegment::create() methods have been modificated accordingly. * New: In Anabatic::GCell, add two new types "StdCellRow" and "ChannelRow" for implementing 2-Metal blocks. Rename the GCell::setXY() method in GCell::setSouthWestCorner(), move the contents of GCell::updateContactsPosition() into it and suppress it. WARNING: In case of a GCell shrink this may cause problems. But for now we only expand... New method GCell::getNetCount() to count the number of Net going though the GCell. * Change: In Anabatic::Edge, add specific support for capacity of 2-Metal routing channels. * Change: In Anabatic::Dijsktra various methods, replace the "gcell->isMatrix()" calls by "not gcell->isAnalog()". Add more check so that the methods pertaining to the analog routing (GRData) are not called in digital mode. * New: In Anabatic::Dijkstra::materialize(), add support for 2-Metal specific cases. That is, always break in case of vertical pass-through or U-turn. The global routing must always be broken in H-Channel. * New: In Anabatic::GCell & Anabatic::Edge, make use of the Session mechanism to ensure the revalidation. The "::revalidate()" method is then moved as "::materialize()" (overload of Go) and "::_invalidate()" becomes "::invalidate()" * Change: In LoadGlobalRouting, cosmetic rename of SortHkByX in SortHookByX. * New: In GCellTopology, added support for building 2-Metal topologies. * ForkStack is now an object attribute as many methods do need it. * To push segments/hook on the stack, a new method "push()" is available. Perform NULL and fromHook checking. Can also setup _southWestContact or _northEastContact if it is the "from" edge. * N/S/E/W edges are now vector as in digital channel mode there can be more than one. * Added build topological build methods: - doRp_2m_Access() RoutingPad stem access. - _do_2m_1G_1M1() North or south access. - _do_2m_2G_1M1() North AND south access. - _do_2m_xG() H-Channel routing. * New: In Anabatic::Matrix, new ::resize() function, as Cell can be resizeds. * New: In Anabatic::Vertex, new static method ::getValueString() for a friendly text rendering. * New: In Katana::DigitalDistance, support for channel routing. * Change: In KatanaEngine::digitalSetup() and KatanaEngine::runGlobalrouter(), for channel routing, calls to setupPowerRails() and protectRoutingPads() must be called after the core block has been fully dimensionned. ::runGlobalrouter() contains the code tasked with the grid creation and channel sizing. * New: In KatanaEngine: Added support for core block, for 2-Metal routing. May be expanded for over-the-cell routing in the future. Added methods : - isDigitalMode() - isAnalogMode() - isMixedMode() - isChannelMode() - getBlock() / addBlock() - setupChannelMode() - createChannel() * New: In Katana, new class Block to manage core blocks and perform channel routing. * New: In Katana::Session, new convenience method "isOpen()".
2017-08-18 16:56:23 -05:00
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
Contact* contact = dynamic_cast<Contact*>( hurricaneSegment->getSource() );
AutoContact* autoContact = Session::lookup( contact );
if (contact == NULL) {
throw Error( badSegmentSource, getString(hurricaneSegment).c_str() );
if ( autoContact and (autoContact != source) )
throw Error( mismatchSegmentSource
, getString(hurricaneSegment).c_str()
, getString(contact).c_str() );
} else {
if (autoContact != source) reattachSource = true;
}
contact = dynamic_cast<Contact*>( hurricaneSegment->getTarget() );
autoContact = Session::lookup( contact );
if (contact == NULL) {
throw Error( badSegmentTarget, getString(hurricaneSegment).c_str() );
if ( autoContact and (autoContact != target) )
throw Error ( mismatchSegmentTarget
, getString(hurricaneSegment).c_str()
, getString(contact).c_str() );
} else {
if (autoContact != source) reattachTarget = true;
}
if (reattachSource) {
Hook* hook = hurricaneSegment->getSourceHook();
hook->detach ();
hook->attach ( source->getBodyHook() );
}
if (reattachTarget) {
Hook* hook = hurricaneSegment->getTargetHook();
hook->detach ();
hook->attach ( target->getBodyHook() );
}
if (horizontal) {
if (horizontal->getLayer() != horizontalLayer) {
if (Session::getAnabatic()->getConfiguration()->isGMetal(horizontal->getLayer())) {
horizontal->setLayer( horizontalLayer );
horizontal->setWidth( horizontalWidth );
} else {
if (horizontal->getWidth() != horizontalWidth) {
cerr << Warning("Segment %s has non-default width %s."
,getString(horizontal).c_str()
,DbU::getValueString(horizontal->getWidth()).c_str()) << endl;
}
}
}
Validating channel routing mode (two metals) on SNX. * New: In Hurricane::Entity, add an id counter limit and a memory size limit. The two limits are checked only when a new Entity object is created. This should help avoiding massive memory links. * New: In CRL Core, add a "crlcore.groundName" and a "crlcore.powerName" parameter to specify the name of the ground/power signals to be created if they are missing in a Cell. For Alliance libraries it would be "vss" & "vdd" (default values), but for real technologies, it is often "gnd!" & "vdd!". The Blif parser is modificated to make use of it. * Bug: In AnabaticEngine::unify(), set the resulting unified segment in the center of the GCells common side. Gcells under a segment are found by using the edge that cover the segment axis. When we have a "bend" GCell stack and the axis is wrong, they could be ommited. This was causing deleted segments to be not removed from some Edges, then core dump. * Change: In Anabatic::AutoSegment::create(), smarter choosing of the reference contact, select the fixed or terminal one instead of always the source one. * New: In Anabatic::Edge::isEnding(), new function to check if a segment going through an Edge is starting/ending in either source or target GCell of the edge (active only when running in channel mode). * New: In Anabatic::Edge::add(), a segment takes part in the occupancy only if it is not ending in either source or target (channel mode only). The occupancy due to terminal is pre-computed in Katana. * New: In Anabatic::Edge::ripup(), in channel mode, never ripup a segment which is ending in either source or target (we *have* to access this edge to connect to the terminal). * Bug: In Anabatic::GCell::hcut() and vcut(), force the update of the Edge which is on the side that will get splitted by the cut. It's capacity will be reduced to it must be updated. * Change: In Anabatic::GCell::updateGContacts() add a flag to conditionnally update horizontals or verticals only. We may require only a partial update when resizing the GCell in only one direction. This, again, related to the fact that we compute the GCells under a segment thanks to it's axis position, so we need to be very careful when modificating axis. * Change: In Katana::Block::resizeChannels(), only update GContact vertical position. Do not disturb X positions of segments. * Bug: In Katana::GlobalRoute::DigitalDistance, in channel mode, some Edges can have a zero capacity, but still be reachable if the net has a terminal in either source or target. Look for this case and return a distance of zero instead of "unreachable". This was causing the global routing not to complete in channel mode. For computing the edge distance, makes the vertical edges much more long (10 times) than the horizontal ones as the vertical capacity is very limited. Hard coded for now, should make it a parameter in the future. * Change: In KatanaEngine::annotateGlobalGraph(), decrease the capacity of edges with reserveCapacity for each terminal inside a GCell. Both north and south edges are decreased as we a terminal will block both north and south edges. As a counterpart, the Edge capacity is not decreased when the global router connect to a terminal. * Change: In Katana::RoutingEvent::revalidate(), when in repair stage, do not expand the slack for horizontal segments in channel mode. So they may not overlap the standard cell row. * Bug: In Stratus documentation, do not use the french option in babel, the documentation is in english! * New: In Documentation, added Hurricane/Python tutorial, part for drawing layout.
2018-03-16 10:20:04 -05:00
if (reference) horizontal->setY( reference->getY() );
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
segment = new AutoHorizontal ( horizontal );
segment->_postCreate();
} else if (vertical) {
if (vertical->getLayer() != verticalLayer) {
DRC correct on Arlet6505 / TSMC C180. Integrate new features and bug fixes so the Arlet 6502 benchs successfully passes real DRC with reference industrial tools. Short summary: * Manage minimum area for VIAs in Katana::Tracks. * Allow different wire width for wires perpandicular to the prefered routing direction. * StackedVIAs used in the clock tree no longer assume an uniform routing grid (same offset & pitch all the way up). * Some hard-coded patches in PowerRails for FlexLib. * New: In CRL/symbolic/cmos/kite.py & cmos45/kite.py, update the RoutingLayerGauges by adding the new PWireWidth parameter. Always zero in case of symbolic layout (too fine tuning). * New: In CRL::RoutingGauge, add accessor to PWireWidth parameter. Modify the clone method. * New: In CRL::RoutingLayerGauge, add new parameter "PWireWidth" to give the width of a wire when it not drawn in the prefered routing direction. If it is set to zero, the normal width is used. * New: In CRL::PyRoutingGauge, export the updated constructor interface. It is *not* backward compatible, one must add the PWireWidth parameter in the various kite.py configuration files (in etc/). * Change: In AnabaticEngine::_gutAnabatic(), disable the minimum area detection mechanism, replaced by a more complete one in Katana::Track. Left commented out for now, but will be removed in the future. * Change: In Anabatic::AutoContact::updateLayer(), now systematically calls setLayerAndWidth() to potentially resize the VIAs. This is needed in real mode as VIAs are *not* macro-generated but have their real final size. * Change: In Anabatic::AutoContact::setLayerAndWidth(), select the width and height of the contact using the gauge wire width *and* perpandicular *wire width*. * Change: In Anabatic::AutoSegment::_initialize(), the "VIA to same cap" to PWireWidth/2, this will be the size of the VIA in the non-preferred direction at the end cap (non-square in real mode). * Change: In Anabatic::AutoSegment::getExtensionCap(), makes different cases for symbolic and real. Use raw length in real, add half the wire width in symbolic. Add a flag to get the extension cap *only*, not increased of half the minimal spacing. * Change: In Anabatic::AutoSegment::bloatStackedStrap(), enhanced, but finally unused... * New: In Anabatic::AutoSegment::create(), use the PWireWidth when the segment is not in the preferred routing direction (and of minimal width). * New: In Anabatic::Configuration, add new getPWirewidth(), DPHorizontalWidth() and DPVerticalWidth() accessors. * Change: In AnabaticEngine::setupPreRouteds(), skip components in in "cut" material. We are only interested in objects containing some metal (happens in real mode when VIAs cuts are really there). * New: In Katana::PowerRailsPlanes::Rail::doLayout(), add an hard-coded patch that artificially enlarge the *wide wire* so the spacing for wide wire is enforced. For now, two pitches on each side for "FlexLib" gauge. * New: In Katana::Track, add support to find and correct small wire chunks so they respect the minimum area rules. Two helper functions: * ::hasSameLayerTurn(), to find if a a TrackElement as non-zero length perpandicular is same layer connected to it. * ::toFoundryGrid(), to ensure that all coordinates will be on the foundry grid (may move in a more shared location). * ::expandToMinArea(), try to expand, *in the routing direction* the too small wire so it respect the minimal area. Check for the free space in the track. Track::minExpandArea() go through all the TrackElements in the track to look for too small ones and correct them. * Change: In Katana::RoutingPlane, add an accessor to get the tracks. * New: In KatanaEngine::finalizeLayout(), add a post-treatment to find for minimal area violations. * Change: In cumulus/plugins.block.configuration.GaugeConf, add a routingBb attribute that will serve as a common reference to all the functions calculation track positions. We must not have two different reference for the core and the corona. The reference is always the corona when we working on a complete chip. * New: In cumulus/plugins.block.configuration.GaugeConf.getTrack(), Simplified and more reliable way of getting tracks positions. Use the routingBb. * New: In cumulus/plugins.block.configuration.GaugeConf.rpAccess(), Make use of getTrack() to get every metal strap on the right X/Y position. * New: In cumulus/plugins.block.configuration.GaugeConf.expandMinArea(), As those wires are left alone by the router, it is our responsability to abide by the minimal area rule here. Hence the code duplication from the router (bad). Mainly wires made for the clock tree, I mean. * Bug: In cumulus/plugins.chip.configuration.ChipConf.setupICore(), the core instance must be placed on the GCell grid, defined by the slice height (X *and* Y). * Bug: In cumulus/plugins.chip.corona.Builder(), forgot to use bigvia for the corners of the inner ring. * Bug: In cumulus/plugins.chip.pads.corona._createCoreWire(), hard-coded patch for LibreSOCIO, the power/ground connectors toward the core are too wide and can create DRC errors when put side by side. Shrink them by the minimal distance.
2020-11-23 16:07:15 -06:00
if (Session::getAnabatic()->getConfiguration()->isGMetal(vertical->getLayer()) ) {
vertical->setLayer( verticalLayer );
vertical->setWidth( verticalWidth );
}
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
} else {
if (vertical->getWidth() != verticalWidth) {
cerr << Warning("Segment %s has non-default width %s."
,getString(vertical).c_str()
,DbU::getValueString(vertical->getWidth()).c_str()) << endl;
}
}
Validating channel routing mode (two metals) on SNX. * New: In Hurricane::Entity, add an id counter limit and a memory size limit. The two limits are checked only when a new Entity object is created. This should help avoiding massive memory links. * New: In CRL Core, add a "crlcore.groundName" and a "crlcore.powerName" parameter to specify the name of the ground/power signals to be created if they are missing in a Cell. For Alliance libraries it would be "vss" & "vdd" (default values), but for real technologies, it is often "gnd!" & "vdd!". The Blif parser is modificated to make use of it. * Bug: In AnabaticEngine::unify(), set the resulting unified segment in the center of the GCells common side. Gcells under a segment are found by using the edge that cover the segment axis. When we have a "bend" GCell stack and the axis is wrong, they could be ommited. This was causing deleted segments to be not removed from some Edges, then core dump. * Change: In Anabatic::AutoSegment::create(), smarter choosing of the reference contact, select the fixed or terminal one instead of always the source one. * New: In Anabatic::Edge::isEnding(), new function to check if a segment going through an Edge is starting/ending in either source or target GCell of the edge (active only when running in channel mode). * New: In Anabatic::Edge::add(), a segment takes part in the occupancy only if it is not ending in either source or target (channel mode only). The occupancy due to terminal is pre-computed in Katana. * New: In Anabatic::Edge::ripup(), in channel mode, never ripup a segment which is ending in either source or target (we *have* to access this edge to connect to the terminal). * Bug: In Anabatic::GCell::hcut() and vcut(), force the update of the Edge which is on the side that will get splitted by the cut. It's capacity will be reduced to it must be updated. * Change: In Anabatic::GCell::updateGContacts() add a flag to conditionnally update horizontals or verticals only. We may require only a partial update when resizing the GCell in only one direction. This, again, related to the fact that we compute the GCells under a segment thanks to it's axis position, so we need to be very careful when modificating axis. * Change: In Katana::Block::resizeChannels(), only update GContact vertical position. Do not disturb X positions of segments. * Bug: In Katana::GlobalRoute::DigitalDistance, in channel mode, some Edges can have a zero capacity, but still be reachable if the net has a terminal in either source or target. Look for this case and return a distance of zero instead of "unreachable". This was causing the global routing not to complete in channel mode. For computing the edge distance, makes the vertical edges much more long (10 times) than the horizontal ones as the vertical capacity is very limited. Hard coded for now, should make it a parameter in the future. * Change: In KatanaEngine::annotateGlobalGraph(), decrease the capacity of edges with reserveCapacity for each terminal inside a GCell. Both north and south edges are decreased as we a terminal will block both north and south edges. As a counterpart, the Edge capacity is not decreased when the global router connect to a terminal. * Change: In Katana::RoutingEvent::revalidate(), when in repair stage, do not expand the slack for horizontal segments in channel mode. So they may not overlap the standard cell row. * Bug: In Stratus documentation, do not use the french option in babel, the documentation is in english! * New: In Documentation, added Hurricane/Python tutorial, part for drawing layout.
2018-03-16 10:20:04 -05:00
if (reference) vertical->setX( reference->getX() );
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
segment = new AutoVertical ( vertical );
segment->_postCreate();
} else {
throw Error( badSegment, getString(source).c_str(), getString(target).c_str() );
}
Katana manage wide wires, and they can also be symmetric. * New: In Anabatic::AutoContact and the derived classes, manages wide wires. The contact self dimension itself according to the segments it is connected to. Special case for the AutoContactTerminal which also read the size of the component it is anchored upon. New refresh method "updateSize()" and flag CntInvalidatedWidth. to compute the size. In AutoContactTerminal, compute the constraint box according to the width of the segment. * New: In Anabatic::AutoSegment, flags are now implemented as "static const" attributes of the class. The flags are stored into a uint64_t as they are more than 32. Added new flag "SegWide" and associated predicates. * Change: In GCellTopology::_doHChannel() and GCellTopology::_doVChannel(), uses the simpler overload of AutoSegment::create() in order to detect the wire width automatically. * New: In Katana::Manipulator, split insertToTrack() and forceToTrack() into a one-track method and a segment level method that iterate over the track span of the segment. * New: In Katana::SegmentFsm, for each cost in the table, now allow access to a specific track. So the base functions have now two parameters: "icost" and "itrack" (has a cost can have multiple tracks in the case of wide segments). * Change: In Katana::TrackElement, remove the index of the element inside it's track, as for a wide segment it will not be meaningful for the non-base track. This means that we have to use the Track::find() method each time instead. Remove the wide flag, as it is a duplicate of the one in AutoSegment. Added a getTrackCount() method to tell the number of track the segment is inserted into. Needed in the Track destroy step to delete a segment only when the last track that refers it is destroyed. Added getSymmetricAxis() to correct the computation of the symmetric base track in case of wide segment as the base track is not centered but the the leftmost one. * Change: In Track::insert() insert wide segments in their whole track span. * Change: In TrackCost, create an array of costs according to the segment track span. * Change: In TrackSegment::create(), now activate the factory and create wide segments. * Bug: In Katana::AutoSegments_Perpandicular, correct the debug indentation problem (ever shifting to the right).
2017-07-28 08:30:22 -05:00
if (wPitch > 1) segment->setFlags( SegWide );
Improved management of AutoContactTerminal for VH gauges (real ones). * New: In Anabatic & Katana, add the new "drag" feature. With VH gauges used by real technologies (M1-H, M2-V, M3-H) a new routing configuration that was not efficiently handled did appear. While the preferred routing direction for metal1 is officially horizontal, due to the way the standard cell must be designed, their metal1 terminals are still verticals (or punctuals). Thus, when connecting to them, we face the case where the metal1 terminal (RoutingPad) is vertical *and* the metal2 wire is also vertical. With that setup, the position of the AutoContactTerminal via12 cannot be deduced, it may range all the way over the metal1 RoutingPad. What may define it's position is the metal3 the metal2 finally connects to. That, is, when we have one horizontal (the metal3) and one vertical (the metal1 RoutingPad). The intermediate wire of metal2 can be kept to a minimum size by "dragging" the via12 close to the via23 when the metal3 wire is moved. * New: In Anabatic & Katana, problem of closely vertically aligneds RoutingPads in metal1 is managed first in PreProcess by restricting the span of the connecteds metal3 and in _makeDogleg also by restricting the span even more tightly (to the RoutingPad itself). * New: In Anabatic::AutoContactTerminal, add the "drag" support. Automatically check if the connecting segment is in the same direction as the RoutingPad, if so, sets the "SegDrag" flag. The dragging state can be known with the "::canDrag()" predicate. * New: In Anabatic::AutoHorizontal, add the "drag" support. The drag state can be known with the "::isDrag()" predicate. In "::_makeDogleg()", when making a dogleg on a dragable segment pass the drag state correctly and restrict the perpandicular span of the perpandicular to the RoutingPad (though segment user constraints). If we make a dogleg on the metal2 is it likely than we cannot go straigth out vertically from the RoutingPad, so the new perpandicular *is* restricted to the RoutingPad span. Idem for AutoVertical. * New: In Katana::Manipulator, add method "::dragMinimize()" which find a hole where to minimize a draggable segment. We finally did not use it, but keep it for potential further use. * New: In Katana::PreProcess, adds a "protectAlignedaccesses()" local function to check for vertically aligned metal1 RoutingPads, in that case setup user constraints on the metal3 segments so they cannot completly cover the other RoutingPad with metal2. We also keep a "metal2protect()" function that create a fixed segment to lock/protect a RoutingPad. Not used for now. * New: In Katana::Session, add a RoutingPad locking event mechanism. This allows us to request the creation of a locking (fixed segment) over a draggable segment. Not used for now. Lock events are processeds before all others as they create new TrackElements. * New: In Katana::Track, "::getNextFree()" and "::getPreviousFree()" method to find the nearest free interval in a Track after/before a position. * Bug: In Anabatic::AutoHorizontal::getConstraints(), merge with user constraints *only* if it's not an empty interval (as we use min/max functions). Idem for AutoVertical. * Bug: In AutoSegments_OnContacts::Locator::isValid(), the boolean test must be inverted. Seems it never worked, but we never used it until now...
2018-01-25 04:58:04 -06:00
if (source->canDrag() or target->canDrag()) segment->setFlags( SegDrag );
Katana manage wide wires, and they can also be symmetric. * New: In Anabatic::AutoContact and the derived classes, manages wide wires. The contact self dimension itself according to the segments it is connected to. Special case for the AutoContactTerminal which also read the size of the component it is anchored upon. New refresh method "updateSize()" and flag CntInvalidatedWidth. to compute the size. In AutoContactTerminal, compute the constraint box according to the width of the segment. * New: In Anabatic::AutoSegment, flags are now implemented as "static const" attributes of the class. The flags are stored into a uint64_t as they are more than 32. Added new flag "SegWide" and associated predicates. * Change: In GCellTopology::_doHChannel() and GCellTopology::_doVChannel(), uses the simpler overload of AutoSegment::create() in order to detect the wire width automatically. * New: In Katana::Manipulator, split insertToTrack() and forceToTrack() into a one-track method and a segment level method that iterate over the track span of the segment. * New: In Katana::SegmentFsm, for each cost in the table, now allow access to a specific track. So the base functions have now two parameters: "icost" and "itrack" (has a cost can have multiple tracks in the case of wide segments). * Change: In Katana::TrackElement, remove the index of the element inside it's track, as for a wide segment it will not be meaningful for the non-base track. This means that we have to use the Track::find() method each time instead. Remove the wide flag, as it is a duplicate of the one in AutoSegment. Added a getTrackCount() method to tell the number of track the segment is inserted into. Needed in the Track destroy step to delete a segment only when the last track that refers it is destroyed. Added getSymmetricAxis() to correct the computation of the symmetric base track in case of wide segment as the base track is not centered but the the leftmost one. * Change: In Track::insert() insert wide segments in their whole track span. * Change: In TrackCost, create an array of costs according to the segment track span. * Change: In TrackSegment::create(), now activate the factory and create wide segments. * Bug: In Katana::AutoSegments_Perpandicular, correct the debug indentation problem (ever shifting to the right).
2017-07-28 08:30:22 -05:00
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
return segment;
}
Replace "unsigned int" by "Flags" in all AutoSegments collections. * Change: In Anabatic::AutoSegments collections, change the type of all the flags that where in "unsigned int" (32 bits) to Flags (uint64_t) as there is now more than 32 flags for functions. * New: In Ababatic::Constants, added new flag Flags::WithPerpands, which makes the number of flags tip over 32 bits, thus making mandatory to uses Flags and not unsigned int. * New: In Anabatic::AutoSegments_Perpandiculars, manage a new flag Flags::WithDoglegs to allow to propagate through global segments that are connecteds via doglegs on local segments. Meaning that there is a good chance that they could be aligned. Slighly change the way we propagate on aligned segments: no longer check for VTee or HTee, but only for same direction and layer as master. * New: In Anabatic & Katana, replace all the "int", "long" and their variants by the less implementation ambiguous "int32_t", "int64_t" (and variant). This should help to better detect bit trucation in flags. Use the type to give a hint about the flags kind: - Type "Flags", for flags shared among Anabatic & Katana functions/methods (may also appear in some objects states). - Type "uint32_t" for flags belonging to an object internal state of from Hurricane functions flags (those should be grouped in a Flag subclass in a perfect world).
2017-05-16 07:53:33 -05:00
AutoSegment* AutoSegment::create ( AutoContact* source
, AutoContact* target
, Flags dir
, size_t depth
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
)
{
// Hardcoded: make the assumption that,
// depth=0 is terminal reserved | METAL1
// depth=1 is horizontal | METAL2
// depth=2 is vertical | METAL3
// Should be based on gauge informations.
Added support for 2-Metal block routing in Anabatic & Katana. * New: In AnabaticEngine::invalidateRoutingPads() this method is a temporary workaround for a Hurricane problems. When an instance is moved, the RoutingPads that use it must be moved accordingly, but they are not invalidated so they stay in the wrong QuadTree. New method ::_resizeMatrix() to be called when the associated Cell is resized. * Bug: In AutoHorizontal::getConstraints() and AutoVertical::getConstraints(), the *target* constraints where never merged. * Change: In AutoHorizontal::getCells() and AutoVertical::getGCells(), now return a boolean to tell if it was ok (must not encounter a NULL GCell while progessing from source to target). * New: In Anabatic::Configuration and Anabatic:Session, create new methods: - getDHorizontalLayer() - getDhorizontalDepth() - getDHorizontalWidth() - getDHorizontalPitch() And so on for Vertical and Contact. They supply depth-independant informations about the H/V layers to build the initial detailed routing. The AutoSegment::create() methods have been modificated accordingly. * New: In Anabatic::GCell, add two new types "StdCellRow" and "ChannelRow" for implementing 2-Metal blocks. Rename the GCell::setXY() method in GCell::setSouthWestCorner(), move the contents of GCell::updateContactsPosition() into it and suppress it. WARNING: In case of a GCell shrink this may cause problems. But for now we only expand... New method GCell::getNetCount() to count the number of Net going though the GCell. * Change: In Anabatic::Edge, add specific support for capacity of 2-Metal routing channels. * Change: In Anabatic::Dijsktra various methods, replace the "gcell->isMatrix()" calls by "not gcell->isAnalog()". Add more check so that the methods pertaining to the analog routing (GRData) are not called in digital mode. * New: In Anabatic::Dijkstra::materialize(), add support for 2-Metal specific cases. That is, always break in case of vertical pass-through or U-turn. The global routing must always be broken in H-Channel. * New: In Anabatic::GCell & Anabatic::Edge, make use of the Session mechanism to ensure the revalidation. The "::revalidate()" method is then moved as "::materialize()" (overload of Go) and "::_invalidate()" becomes "::invalidate()" * Change: In LoadGlobalRouting, cosmetic rename of SortHkByX in SortHookByX. * New: In GCellTopology, added support for building 2-Metal topologies. * ForkStack is now an object attribute as many methods do need it. * To push segments/hook on the stack, a new method "push()" is available. Perform NULL and fromHook checking. Can also setup _southWestContact or _northEastContact if it is the "from" edge. * N/S/E/W edges are now vector as in digital channel mode there can be more than one. * Added build topological build methods: - doRp_2m_Access() RoutingPad stem access. - _do_2m_1G_1M1() North or south access. - _do_2m_2G_1M1() North AND south access. - _do_2m_xG() H-Channel routing. * New: In Anabatic::Matrix, new ::resize() function, as Cell can be resizeds. * New: In Anabatic::Vertex, new static method ::getValueString() for a friendly text rendering. * New: In Katana::DigitalDistance, support for channel routing. * Change: In KatanaEngine::digitalSetup() and KatanaEngine::runGlobalrouter(), for channel routing, calls to setupPowerRails() and protectRoutingPads() must be called after the core block has been fully dimensionned. ::runGlobalrouter() contains the code tasked with the grid creation and channel sizing. * New: In KatanaEngine: Added support for core block, for 2-Metal routing. May be expanded for over-the-cell routing in the future. Added methods : - isDigitalMode() - isAnalogMode() - isMixedMode() - isChannelMode() - getBlock() / addBlock() - setupChannelMode() - createChannel() * New: In Katana, new class Block to manage core blocks and perform channel routing. * New: In Katana::Session, new convenience method "isOpen()".
2017-08-18 16:56:23 -05:00
const Layer* hLayer = Session::getDHorizontalLayer();
DbU::Unit hWidth = Session::getDHorizontalWidth();
const Layer* vLayer = Session::getDVerticalLayer();
DbU::Unit vWidth = Session::getDVerticalWidth();
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
if (dir & Flags::UseNonPref) {
if (dir & Flags::Vertical) {
vLayer = hLayer;
DRC correct on Arlet6505 / TSMC C180. Integrate new features and bug fixes so the Arlet 6502 benchs successfully passes real DRC with reference industrial tools. Short summary: * Manage minimum area for VIAs in Katana::Tracks. * Allow different wire width for wires perpandicular to the prefered routing direction. * StackedVIAs used in the clock tree no longer assume an uniform routing grid (same offset & pitch all the way up). * Some hard-coded patches in PowerRails for FlexLib. * New: In CRL/symbolic/cmos/kite.py & cmos45/kite.py, update the RoutingLayerGauges by adding the new PWireWidth parameter. Always zero in case of symbolic layout (too fine tuning). * New: In CRL::RoutingGauge, add accessor to PWireWidth parameter. Modify the clone method. * New: In CRL::RoutingLayerGauge, add new parameter "PWireWidth" to give the width of a wire when it not drawn in the prefered routing direction. If it is set to zero, the normal width is used. * New: In CRL::PyRoutingGauge, export the updated constructor interface. It is *not* backward compatible, one must add the PWireWidth parameter in the various kite.py configuration files (in etc/). * Change: In AnabaticEngine::_gutAnabatic(), disable the minimum area detection mechanism, replaced by a more complete one in Katana::Track. Left commented out for now, but will be removed in the future. * Change: In Anabatic::AutoContact::updateLayer(), now systematically calls setLayerAndWidth() to potentially resize the VIAs. This is needed in real mode as VIAs are *not* macro-generated but have their real final size. * Change: In Anabatic::AutoContact::setLayerAndWidth(), select the width and height of the contact using the gauge wire width *and* perpandicular *wire width*. * Change: In Anabatic::AutoSegment::_initialize(), the "VIA to same cap" to PWireWidth/2, this will be the size of the VIA in the non-preferred direction at the end cap (non-square in real mode). * Change: In Anabatic::AutoSegment::getExtensionCap(), makes different cases for symbolic and real. Use raw length in real, add half the wire width in symbolic. Add a flag to get the extension cap *only*, not increased of half the minimal spacing. * Change: In Anabatic::AutoSegment::bloatStackedStrap(), enhanced, but finally unused... * New: In Anabatic::AutoSegment::create(), use the PWireWidth when the segment is not in the preferred routing direction (and of minimal width). * New: In Anabatic::Configuration, add new getPWirewidth(), DPHorizontalWidth() and DPVerticalWidth() accessors. * Change: In AnabaticEngine::setupPreRouteds(), skip components in in "cut" material. We are only interested in objects containing some metal (happens in real mode when VIAs cuts are really there). * New: In Katana::PowerRailsPlanes::Rail::doLayout(), add an hard-coded patch that artificially enlarge the *wide wire* so the spacing for wide wire is enforced. For now, two pitches on each side for "FlexLib" gauge. * New: In Katana::Track, add support to find and correct small wire chunks so they respect the minimum area rules. Two helper functions: * ::hasSameLayerTurn(), to find if a a TrackElement as non-zero length perpandicular is same layer connected to it. * ::toFoundryGrid(), to ensure that all coordinates will be on the foundry grid (may move in a more shared location). * ::expandToMinArea(), try to expand, *in the routing direction* the too small wire so it respect the minimal area. Check for the free space in the track. Track::minExpandArea() go through all the TrackElements in the track to look for too small ones and correct them. * Change: In Katana::RoutingPlane, add an accessor to get the tracks. * New: In KatanaEngine::finalizeLayout(), add a post-treatment to find for minimal area violations. * Change: In cumulus/plugins.block.configuration.GaugeConf, add a routingBb attribute that will serve as a common reference to all the functions calculation track positions. We must not have two different reference for the core and the corona. The reference is always the corona when we working on a complete chip. * New: In cumulus/plugins.block.configuration.GaugeConf.getTrack(), Simplified and more reliable way of getting tracks positions. Use the routingBb. * New: In cumulus/plugins.block.configuration.GaugeConf.rpAccess(), Make use of getTrack() to get every metal strap on the right X/Y position. * New: In cumulus/plugins.block.configuration.GaugeConf.expandMinArea(), As those wires are left alone by the router, it is our responsability to abide by the minimal area rule here. Hence the code duplication from the router (bad). Mainly wires made for the clock tree, I mean. * Bug: In cumulus/plugins.chip.configuration.ChipConf.setupICore(), the core instance must be placed on the GCell grid, defined by the slice height (X *and* Y). * Bug: In cumulus/plugins.chip.corona.Builder(), forgot to use bigvia for the corners of the inner ring. * Bug: In cumulus/plugins.chip.pads.corona._createCoreWire(), hard-coded patch for LibreSOCIO, the power/ground connectors toward the core are too wide and can create DRC errors when put side by side. Shrink them by the minimal distance.
2020-11-23 16:07:15 -06:00
vWidth = Session::getDPHorizontalWidth();
cdebug_log(149,0) << "Make vertical in non-preferred direction (ppW:"
<< DbU::getValueString(vWidth).c_str() << ")." << endl;
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
}
if (dir & Flags::Horizontal) {
hLayer = vLayer;
DRC correct on Arlet6505 / TSMC C180. Integrate new features and bug fixes so the Arlet 6502 benchs successfully passes real DRC with reference industrial tools. Short summary: * Manage minimum area for VIAs in Katana::Tracks. * Allow different wire width for wires perpandicular to the prefered routing direction. * StackedVIAs used in the clock tree no longer assume an uniform routing grid (same offset & pitch all the way up). * Some hard-coded patches in PowerRails for FlexLib. * New: In CRL/symbolic/cmos/kite.py & cmos45/kite.py, update the RoutingLayerGauges by adding the new PWireWidth parameter. Always zero in case of symbolic layout (too fine tuning). * New: In CRL::RoutingGauge, add accessor to PWireWidth parameter. Modify the clone method. * New: In CRL::RoutingLayerGauge, add new parameter "PWireWidth" to give the width of a wire when it not drawn in the prefered routing direction. If it is set to zero, the normal width is used. * New: In CRL::PyRoutingGauge, export the updated constructor interface. It is *not* backward compatible, one must add the PWireWidth parameter in the various kite.py configuration files (in etc/). * Change: In AnabaticEngine::_gutAnabatic(), disable the minimum area detection mechanism, replaced by a more complete one in Katana::Track. Left commented out for now, but will be removed in the future. * Change: In Anabatic::AutoContact::updateLayer(), now systematically calls setLayerAndWidth() to potentially resize the VIAs. This is needed in real mode as VIAs are *not* macro-generated but have their real final size. * Change: In Anabatic::AutoContact::setLayerAndWidth(), select the width and height of the contact using the gauge wire width *and* perpandicular *wire width*. * Change: In Anabatic::AutoSegment::_initialize(), the "VIA to same cap" to PWireWidth/2, this will be the size of the VIA in the non-preferred direction at the end cap (non-square in real mode). * Change: In Anabatic::AutoSegment::getExtensionCap(), makes different cases for symbolic and real. Use raw length in real, add half the wire width in symbolic. Add a flag to get the extension cap *only*, not increased of half the minimal spacing. * Change: In Anabatic::AutoSegment::bloatStackedStrap(), enhanced, but finally unused... * New: In Anabatic::AutoSegment::create(), use the PWireWidth when the segment is not in the preferred routing direction (and of minimal width). * New: In Anabatic::Configuration, add new getPWirewidth(), DPHorizontalWidth() and DPVerticalWidth() accessors. * Change: In AnabaticEngine::setupPreRouteds(), skip components in in "cut" material. We are only interested in objects containing some metal (happens in real mode when VIAs cuts are really there). * New: In Katana::PowerRailsPlanes::Rail::doLayout(), add an hard-coded patch that artificially enlarge the *wide wire* so the spacing for wide wire is enforced. For now, two pitches on each side for "FlexLib" gauge. * New: In Katana::Track, add support to find and correct small wire chunks so they respect the minimum area rules. Two helper functions: * ::hasSameLayerTurn(), to find if a a TrackElement as non-zero length perpandicular is same layer connected to it. * ::toFoundryGrid(), to ensure that all coordinates will be on the foundry grid (may move in a more shared location). * ::expandToMinArea(), try to expand, *in the routing direction* the too small wire so it respect the minimal area. Check for the free space in the track. Track::minExpandArea() go through all the TrackElements in the track to look for too small ones and correct them. * Change: In Katana::RoutingPlane, add an accessor to get the tracks. * New: In KatanaEngine::finalizeLayout(), add a post-treatment to find for minimal area violations. * Change: In cumulus/plugins.block.configuration.GaugeConf, add a routingBb attribute that will serve as a common reference to all the functions calculation track positions. We must not have two different reference for the core and the corona. The reference is always the corona when we working on a complete chip. * New: In cumulus/plugins.block.configuration.GaugeConf.getTrack(), Simplified and more reliable way of getting tracks positions. Use the routingBb. * New: In cumulus/plugins.block.configuration.GaugeConf.rpAccess(), Make use of getTrack() to get every metal strap on the right X/Y position. * New: In cumulus/plugins.block.configuration.GaugeConf.expandMinArea(), As those wires are left alone by the router, it is our responsability to abide by the minimal area rule here. Hence the code duplication from the router (bad). Mainly wires made for the clock tree, I mean. * Bug: In cumulus/plugins.chip.configuration.ChipConf.setupICore(), the core instance must be placed on the GCell grid, defined by the slice height (X *and* Y). * Bug: In cumulus/plugins.chip.corona.Builder(), forgot to use bigvia for the corners of the inner ring. * Bug: In cumulus/plugins.chip.pads.corona._createCoreWire(), hard-coded patch for LibreSOCIO, the power/ground connectors toward the core are too wide and can create DRC errors when put side by side. Shrink them by the minimal distance.
2020-11-23 16:07:15 -06:00
hWidth = Session::getDPVerticalWidth();
cdebug_log(149,0) << "Make horizontal in non-preferred direction (ppW:"
<< DbU::getValueString(hWidth).c_str() << ")." << endl;
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
}
}
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
const Layer* horizontalLayer = hLayer;
DbU::Unit horizontalWidth = hWidth;
const Layer* verticalLayer = vLayer;
DbU::Unit verticalWidth = vWidth;
DRC correct on Arlet6505 / TSMC C180. Integrate new features and bug fixes so the Arlet 6502 benchs successfully passes real DRC with reference industrial tools. Short summary: * Manage minimum area for VIAs in Katana::Tracks. * Allow different wire width for wires perpandicular to the prefered routing direction. * StackedVIAs used in the clock tree no longer assume an uniform routing grid (same offset & pitch all the way up). * Some hard-coded patches in PowerRails for FlexLib. * New: In CRL/symbolic/cmos/kite.py & cmos45/kite.py, update the RoutingLayerGauges by adding the new PWireWidth parameter. Always zero in case of symbolic layout (too fine tuning). * New: In CRL::RoutingGauge, add accessor to PWireWidth parameter. Modify the clone method. * New: In CRL::RoutingLayerGauge, add new parameter "PWireWidth" to give the width of a wire when it not drawn in the prefered routing direction. If it is set to zero, the normal width is used. * New: In CRL::PyRoutingGauge, export the updated constructor interface. It is *not* backward compatible, one must add the PWireWidth parameter in the various kite.py configuration files (in etc/). * Change: In AnabaticEngine::_gutAnabatic(), disable the minimum area detection mechanism, replaced by a more complete one in Katana::Track. Left commented out for now, but will be removed in the future. * Change: In Anabatic::AutoContact::updateLayer(), now systematically calls setLayerAndWidth() to potentially resize the VIAs. This is needed in real mode as VIAs are *not* macro-generated but have their real final size. * Change: In Anabatic::AutoContact::setLayerAndWidth(), select the width and height of the contact using the gauge wire width *and* perpandicular *wire width*. * Change: In Anabatic::AutoSegment::_initialize(), the "VIA to same cap" to PWireWidth/2, this will be the size of the VIA in the non-preferred direction at the end cap (non-square in real mode). * Change: In Anabatic::AutoSegment::getExtensionCap(), makes different cases for symbolic and real. Use raw length in real, add half the wire width in symbolic. Add a flag to get the extension cap *only*, not increased of half the minimal spacing. * Change: In Anabatic::AutoSegment::bloatStackedStrap(), enhanced, but finally unused... * New: In Anabatic::AutoSegment::create(), use the PWireWidth when the segment is not in the preferred routing direction (and of minimal width). * New: In Anabatic::Configuration, add new getPWirewidth(), DPHorizontalWidth() and DPVerticalWidth() accessors. * Change: In AnabaticEngine::setupPreRouteds(), skip components in in "cut" material. We are only interested in objects containing some metal (happens in real mode when VIAs cuts are really there). * New: In Katana::PowerRailsPlanes::Rail::doLayout(), add an hard-coded patch that artificially enlarge the *wide wire* so the spacing for wide wire is enforced. For now, two pitches on each side for "FlexLib" gauge. * New: In Katana::Track, add support to find and correct small wire chunks so they respect the minimum area rules. Two helper functions: * ::hasSameLayerTurn(), to find if a a TrackElement as non-zero length perpandicular is same layer connected to it. * ::toFoundryGrid(), to ensure that all coordinates will be on the foundry grid (may move in a more shared location). * ::expandToMinArea(), try to expand, *in the routing direction* the too small wire so it respect the minimal area. Check for the free space in the track. Track::minExpandArea() go through all the TrackElements in the track to look for too small ones and correct them. * Change: In Katana::RoutingPlane, add an accessor to get the tracks. * New: In KatanaEngine::finalizeLayout(), add a post-treatment to find for minimal area violations. * Change: In cumulus/plugins.block.configuration.GaugeConf, add a routingBb attribute that will serve as a common reference to all the functions calculation track positions. We must not have two different reference for the core and the corona. The reference is always the corona when we working on a complete chip. * New: In cumulus/plugins.block.configuration.GaugeConf.getTrack(), Simplified and more reliable way of getting tracks positions. Use the routingBb. * New: In cumulus/plugins.block.configuration.GaugeConf.rpAccess(), Make use of getTrack() to get every metal strap on the right X/Y position. * New: In cumulus/plugins.block.configuration.GaugeConf.expandMinArea(), As those wires are left alone by the router, it is our responsability to abide by the minimal area rule here. Hence the code duplication from the router (bad). Mainly wires made for the clock tree, I mean. * Bug: In cumulus/plugins.chip.configuration.ChipConf.setupICore(), the core instance must be placed on the GCell grid, defined by the slice height (X *and* Y). * Bug: In cumulus/plugins.chip.corona.Builder(), forgot to use bigvia for the corners of the inner ring. * Bug: In cumulus/plugins.chip.pads.corona._createCoreWire(), hard-coded patch for LibreSOCIO, the power/ground connectors toward the core are too wide and can create DRC errors when put side by side. Shrink them by the minimal distance.
2020-11-23 16:07:15 -06:00
cdebug_log(149,0) << "verticalWidth:" << DbU::getValueString(verticalWidth).c_str() << endl;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
Katana manage wide wires, and they can also be symmetric. * New: In Anabatic::AutoContact and the derived classes, manages wide wires. The contact self dimension itself according to the segments it is connected to. Special case for the AutoContactTerminal which also read the size of the component it is anchored upon. New refresh method "updateSize()" and flag CntInvalidatedWidth. to compute the size. In AutoContactTerminal, compute the constraint box according to the width of the segment. * New: In Anabatic::AutoSegment, flags are now implemented as "static const" attributes of the class. The flags are stored into a uint64_t as they are more than 32. Added new flag "SegWide" and associated predicates. * Change: In GCellTopology::_doHChannel() and GCellTopology::_doVChannel(), uses the simpler overload of AutoSegment::create() in order to detect the wire width automatically. * New: In Katana::Manipulator, split insertToTrack() and forceToTrack() into a one-track method and a segment level method that iterate over the track span of the segment. * New: In Katana::SegmentFsm, for each cost in the table, now allow access to a specific track. So the base functions have now two parameters: "icost" and "itrack" (has a cost can have multiple tracks in the case of wide segments). * Change: In Katana::TrackElement, remove the index of the element inside it's track, as for a wide segment it will not be meaningful for the non-base track. This means that we have to use the Track::find() method each time instead. Remove the wide flag, as it is a duplicate of the one in AutoSegment. Added a getTrackCount() method to tell the number of track the segment is inserted into. Needed in the Track destroy step to delete a segment only when the last track that refers it is destroyed. Added getSymmetricAxis() to correct the computation of the symmetric base track in case of wide segment as the base track is not centered but the the leftmost one. * Change: In Track::insert() insert wide segments in their whole track span. * Change: In TrackCost, create an array of costs according to the segment track span. * Change: In TrackSegment::create(), now activate the factory and create wide segments. * Bug: In Katana::AutoSegments_Perpandicular, correct the debug indentation problem (ever shifting to the right).
2017-07-28 08:30:22 -05:00
uint32_t wPitch = NetRoutingExtension::getWPitch( source->getNet() );
if (wPitch > 1) {
Added support for 2-Metal block routing in Anabatic & Katana. * New: In AnabaticEngine::invalidateRoutingPads() this method is a temporary workaround for a Hurricane problems. When an instance is moved, the RoutingPads that use it must be moved accordingly, but they are not invalidated so they stay in the wrong QuadTree. New method ::_resizeMatrix() to be called when the associated Cell is resized. * Bug: In AutoHorizontal::getConstraints() and AutoVertical::getConstraints(), the *target* constraints where never merged. * Change: In AutoHorizontal::getCells() and AutoVertical::getGCells(), now return a boolean to tell if it was ok (must not encounter a NULL GCell while progessing from source to target). * New: In Anabatic::Configuration and Anabatic:Session, create new methods: - getDHorizontalLayer() - getDhorizontalDepth() - getDHorizontalWidth() - getDHorizontalPitch() And so on for Vertical and Contact. They supply depth-independant informations about the H/V layers to build the initial detailed routing. The AutoSegment::create() methods have been modificated accordingly. * New: In Anabatic::GCell, add two new types "StdCellRow" and "ChannelRow" for implementing 2-Metal blocks. Rename the GCell::setXY() method in GCell::setSouthWestCorner(), move the contents of GCell::updateContactsPosition() into it and suppress it. WARNING: In case of a GCell shrink this may cause problems. But for now we only expand... New method GCell::getNetCount() to count the number of Net going though the GCell. * Change: In Anabatic::Edge, add specific support for capacity of 2-Metal routing channels. * Change: In Anabatic::Dijsktra various methods, replace the "gcell->isMatrix()" calls by "not gcell->isAnalog()". Add more check so that the methods pertaining to the analog routing (GRData) are not called in digital mode. * New: In Anabatic::Dijkstra::materialize(), add support for 2-Metal specific cases. That is, always break in case of vertical pass-through or U-turn. The global routing must always be broken in H-Channel. * New: In Anabatic::GCell & Anabatic::Edge, make use of the Session mechanism to ensure the revalidation. The "::revalidate()" method is then moved as "::materialize()" (overload of Go) and "::_invalidate()" becomes "::invalidate()" * Change: In LoadGlobalRouting, cosmetic rename of SortHkByX in SortHookByX. * New: In GCellTopology, added support for building 2-Metal topologies. * ForkStack is now an object attribute as many methods do need it. * To push segments/hook on the stack, a new method "push()" is available. Perform NULL and fromHook checking. Can also setup _southWestContact or _northEastContact if it is the "from" edge. * N/S/E/W edges are now vector as in digital channel mode there can be more than one. * Added build topological build methods: - doRp_2m_Access() RoutingPad stem access. - _do_2m_1G_1M1() North or south access. - _do_2m_2G_1M1() North AND south access. - _do_2m_xG() H-Channel routing. * New: In Anabatic::Matrix, new ::resize() function, as Cell can be resizeds. * New: In Anabatic::Vertex, new static method ::getValueString() for a friendly text rendering. * New: In Katana::DigitalDistance, support for channel routing. * Change: In KatanaEngine::digitalSetup() and KatanaEngine::runGlobalrouter(), for channel routing, calls to setupPowerRails() and protectRoutingPads() must be called after the core block has been fully dimensionned. ::runGlobalrouter() contains the code tasked with the grid creation and channel sizing. * New: In KatanaEngine: Added support for core block, for 2-Metal routing. May be expanded for over-the-cell routing in the future. Added methods : - isDigitalMode() - isAnalogMode() - isMixedMode() - isChannelMode() - getBlock() / addBlock() - setupChannelMode() - createChannel() * New: In Katana, new class Block to manage core blocks and perform channel routing. * New: In Katana::Session, new convenience method "isOpen()".
2017-08-18 16:56:23 -05:00
horizontalWidth = (wPitch-1) * Session::getDHorizontalPitch() + hWidth;
verticalWidth = (wPitch-1) * Session::getDVerticalPitch () + vWidth;
Katana manage wide wires, and they can also be symmetric. * New: In Anabatic::AutoContact and the derived classes, manages wide wires. The contact self dimension itself according to the segments it is connected to. Special case for the AutoContactTerminal which also read the size of the component it is anchored upon. New refresh method "updateSize()" and flag CntInvalidatedWidth. to compute the size. In AutoContactTerminal, compute the constraint box according to the width of the segment. * New: In Anabatic::AutoSegment, flags are now implemented as "static const" attributes of the class. The flags are stored into a uint64_t as they are more than 32. Added new flag "SegWide" and associated predicates. * Change: In GCellTopology::_doHChannel() and GCellTopology::_doVChannel(), uses the simpler overload of AutoSegment::create() in order to detect the wire width automatically. * New: In Katana::Manipulator, split insertToTrack() and forceToTrack() into a one-track method and a segment level method that iterate over the track span of the segment. * New: In Katana::SegmentFsm, for each cost in the table, now allow access to a specific track. So the base functions have now two parameters: "icost" and "itrack" (has a cost can have multiple tracks in the case of wide segments). * Change: In Katana::TrackElement, remove the index of the element inside it's track, as for a wide segment it will not be meaningful for the non-base track. This means that we have to use the Track::find() method each time instead. Remove the wide flag, as it is a duplicate of the one in AutoSegment. Added a getTrackCount() method to tell the number of track the segment is inserted into. Needed in the Track destroy step to delete a segment only when the last track that refers it is destroyed. Added getSymmetricAxis() to correct the computation of the symmetric base track in case of wide segment as the base track is not centered but the the leftmost one. * Change: In Track::insert() insert wide segments in their whole track span. * Change: In TrackCost, create an array of costs according to the segment track span. * Change: In TrackSegment::create(), now activate the factory and create wide segments. * Bug: In Katana::AutoSegments_Perpandicular, correct the debug indentation problem (ever shifting to the right).
2017-07-28 08:30:22 -05:00
}
DRC correct on Arlet6505 / TSMC C180. Integrate new features and bug fixes so the Arlet 6502 benchs successfully passes real DRC with reference industrial tools. Short summary: * Manage minimum area for VIAs in Katana::Tracks. * Allow different wire width for wires perpandicular to the prefered routing direction. * StackedVIAs used in the clock tree no longer assume an uniform routing grid (same offset & pitch all the way up). * Some hard-coded patches in PowerRails for FlexLib. * New: In CRL/symbolic/cmos/kite.py & cmos45/kite.py, update the RoutingLayerGauges by adding the new PWireWidth parameter. Always zero in case of symbolic layout (too fine tuning). * New: In CRL::RoutingGauge, add accessor to PWireWidth parameter. Modify the clone method. * New: In CRL::RoutingLayerGauge, add new parameter "PWireWidth" to give the width of a wire when it not drawn in the prefered routing direction. If it is set to zero, the normal width is used. * New: In CRL::PyRoutingGauge, export the updated constructor interface. It is *not* backward compatible, one must add the PWireWidth parameter in the various kite.py configuration files (in etc/). * Change: In AnabaticEngine::_gutAnabatic(), disable the minimum area detection mechanism, replaced by a more complete one in Katana::Track. Left commented out for now, but will be removed in the future. * Change: In Anabatic::AutoContact::updateLayer(), now systematically calls setLayerAndWidth() to potentially resize the VIAs. This is needed in real mode as VIAs are *not* macro-generated but have their real final size. * Change: In Anabatic::AutoContact::setLayerAndWidth(), select the width and height of the contact using the gauge wire width *and* perpandicular *wire width*. * Change: In Anabatic::AutoSegment::_initialize(), the "VIA to same cap" to PWireWidth/2, this will be the size of the VIA in the non-preferred direction at the end cap (non-square in real mode). * Change: In Anabatic::AutoSegment::getExtensionCap(), makes different cases for symbolic and real. Use raw length in real, add half the wire width in symbolic. Add a flag to get the extension cap *only*, not increased of half the minimal spacing. * Change: In Anabatic::AutoSegment::bloatStackedStrap(), enhanced, but finally unused... * New: In Anabatic::AutoSegment::create(), use the PWireWidth when the segment is not in the preferred routing direction (and of minimal width). * New: In Anabatic::Configuration, add new getPWirewidth(), DPHorizontalWidth() and DPVerticalWidth() accessors. * Change: In AnabaticEngine::setupPreRouteds(), skip components in in "cut" material. We are only interested in objects containing some metal (happens in real mode when VIAs cuts are really there). * New: In Katana::PowerRailsPlanes::Rail::doLayout(), add an hard-coded patch that artificially enlarge the *wide wire* so the spacing for wide wire is enforced. For now, two pitches on each side for "FlexLib" gauge. * New: In Katana::Track, add support to find and correct small wire chunks so they respect the minimum area rules. Two helper functions: * ::hasSameLayerTurn(), to find if a a TrackElement as non-zero length perpandicular is same layer connected to it. * ::toFoundryGrid(), to ensure that all coordinates will be on the foundry grid (may move in a more shared location). * ::expandToMinArea(), try to expand, *in the routing direction* the too small wire so it respect the minimal area. Check for the free space in the track. Track::minExpandArea() go through all the TrackElements in the track to look for too small ones and correct them. * Change: In Katana::RoutingPlane, add an accessor to get the tracks. * New: In KatanaEngine::finalizeLayout(), add a post-treatment to find for minimal area violations. * Change: In cumulus/plugins.block.configuration.GaugeConf, add a routingBb attribute that will serve as a common reference to all the functions calculation track positions. We must not have two different reference for the core and the corona. The reference is always the corona when we working on a complete chip. * New: In cumulus/plugins.block.configuration.GaugeConf.getTrack(), Simplified and more reliable way of getting tracks positions. Use the routingBb. * New: In cumulus/plugins.block.configuration.GaugeConf.rpAccess(), Make use of getTrack() to get every metal strap on the right X/Y position. * New: In cumulus/plugins.block.configuration.GaugeConf.expandMinArea(), As those wires are left alone by the router, it is our responsability to abide by the minimal area rule here. Hence the code duplication from the router (bad). Mainly wires made for the clock tree, I mean. * Bug: In cumulus/plugins.chip.configuration.ChipConf.setupICore(), the core instance must be placed on the GCell grid, defined by the slice height (X *and* Y). * Bug: In cumulus/plugins.chip.corona.Builder(), forgot to use bigvia for the corners of the inner ring. * Bug: In cumulus/plugins.chip.pads.corona._createCoreWire(), hard-coded patch for LibreSOCIO, the power/ground connectors toward the core are too wide and can create DRC errors when put side by side. Shrink them by the minimal distance.
2020-11-23 16:07:15 -06:00
cdebug_log(149,0) << "verticalWidth:" << DbU::getValueString(verticalWidth).c_str() << endl;
Katana manage wide wires, and they can also be symmetric. * New: In Anabatic::AutoContact and the derived classes, manages wide wires. The contact self dimension itself according to the segments it is connected to. Special case for the AutoContactTerminal which also read the size of the component it is anchored upon. New refresh method "updateSize()" and flag CntInvalidatedWidth. to compute the size. In AutoContactTerminal, compute the constraint box according to the width of the segment. * New: In Anabatic::AutoSegment, flags are now implemented as "static const" attributes of the class. The flags are stored into a uint64_t as they are more than 32. Added new flag "SegWide" and associated predicates. * Change: In GCellTopology::_doHChannel() and GCellTopology::_doVChannel(), uses the simpler overload of AutoSegment::create() in order to detect the wire width automatically. * New: In Katana::Manipulator, split insertToTrack() and forceToTrack() into a one-track method and a segment level method that iterate over the track span of the segment. * New: In Katana::SegmentFsm, for each cost in the table, now allow access to a specific track. So the base functions have now two parameters: "icost" and "itrack" (has a cost can have multiple tracks in the case of wide segments). * Change: In Katana::TrackElement, remove the index of the element inside it's track, as for a wide segment it will not be meaningful for the non-base track. This means that we have to use the Track::find() method each time instead. Remove the wide flag, as it is a duplicate of the one in AutoSegment. Added a getTrackCount() method to tell the number of track the segment is inserted into. Needed in the Track destroy step to delete a segment only when the last track that refers it is destroyed. Added getSymmetricAxis() to correct the computation of the symmetric base track in case of wide segment as the base track is not centered but the the leftmost one. * Change: In Track::insert() insert wide segments in their whole track span. * Change: In TrackCost, create an array of costs according to the segment track span. * Change: In TrackSegment::create(), now activate the factory and create wide segments. * Bug: In Katana::AutoSegments_Perpandicular, correct the debug indentation problem (ever shifting to the right).
2017-07-28 08:30:22 -05:00
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
if (depth != RoutingGauge::nlayerdepth) {
horizontalLayer = verticalLayer = Session::getRoutingLayer( depth );
Katana manage wide wires, and they can also be symmetric. * New: In Anabatic::AutoContact and the derived classes, manages wide wires. The contact self dimension itself according to the segments it is connected to. Special case for the AutoContactTerminal which also read the size of the component it is anchored upon. New refresh method "updateSize()" and flag CntInvalidatedWidth. to compute the size. In AutoContactTerminal, compute the constraint box according to the width of the segment. * New: In Anabatic::AutoSegment, flags are now implemented as "static const" attributes of the class. The flags are stored into a uint64_t as they are more than 32. Added new flag "SegWide" and associated predicates. * Change: In GCellTopology::_doHChannel() and GCellTopology::_doVChannel(), uses the simpler overload of AutoSegment::create() in order to detect the wire width automatically. * New: In Katana::Manipulator, split insertToTrack() and forceToTrack() into a one-track method and a segment level method that iterate over the track span of the segment. * New: In Katana::SegmentFsm, for each cost in the table, now allow access to a specific track. So the base functions have now two parameters: "icost" and "itrack" (has a cost can have multiple tracks in the case of wide segments). * Change: In Katana::TrackElement, remove the index of the element inside it's track, as for a wide segment it will not be meaningful for the non-base track. This means that we have to use the Track::find() method each time instead. Remove the wide flag, as it is a duplicate of the one in AutoSegment. Added a getTrackCount() method to tell the number of track the segment is inserted into. Needed in the Track destroy step to delete a segment only when the last track that refers it is destroyed. Added getSymmetricAxis() to correct the computation of the symmetric base track in case of wide segment as the base track is not centered but the the leftmost one. * Change: In Track::insert() insert wide segments in their whole track span. * Change: In TrackCost, create an array of costs according to the segment track span. * Change: In TrackSegment::create(), now activate the factory and create wide segments. * Bug: In Katana::AutoSegments_Perpandicular, correct the debug indentation problem (ever shifting to the right).
2017-07-28 08:30:22 -05:00
if (wPitch > 1) {
horizontalWidth = verticalWidth = (wPitch-1) * Session::getPitch (depth)
+ Session::getWireWidth(depth);
} else {
DRC correct on Arlet6505 / TSMC C180. Integrate new features and bug fixes so the Arlet 6502 benchs successfully passes real DRC with reference industrial tools. Short summary: * Manage minimum area for VIAs in Katana::Tracks. * Allow different wire width for wires perpandicular to the prefered routing direction. * StackedVIAs used in the clock tree no longer assume an uniform routing grid (same offset & pitch all the way up). * Some hard-coded patches in PowerRails for FlexLib. * New: In CRL/symbolic/cmos/kite.py & cmos45/kite.py, update the RoutingLayerGauges by adding the new PWireWidth parameter. Always zero in case of symbolic layout (too fine tuning). * New: In CRL::RoutingGauge, add accessor to PWireWidth parameter. Modify the clone method. * New: In CRL::RoutingLayerGauge, add new parameter "PWireWidth" to give the width of a wire when it not drawn in the prefered routing direction. If it is set to zero, the normal width is used. * New: In CRL::PyRoutingGauge, export the updated constructor interface. It is *not* backward compatible, one must add the PWireWidth parameter in the various kite.py configuration files (in etc/). * Change: In AnabaticEngine::_gutAnabatic(), disable the minimum area detection mechanism, replaced by a more complete one in Katana::Track. Left commented out for now, but will be removed in the future. * Change: In Anabatic::AutoContact::updateLayer(), now systematically calls setLayerAndWidth() to potentially resize the VIAs. This is needed in real mode as VIAs are *not* macro-generated but have their real final size. * Change: In Anabatic::AutoContact::setLayerAndWidth(), select the width and height of the contact using the gauge wire width *and* perpandicular *wire width*. * Change: In Anabatic::AutoSegment::_initialize(), the "VIA to same cap" to PWireWidth/2, this will be the size of the VIA in the non-preferred direction at the end cap (non-square in real mode). * Change: In Anabatic::AutoSegment::getExtensionCap(), makes different cases for symbolic and real. Use raw length in real, add half the wire width in symbolic. Add a flag to get the extension cap *only*, not increased of half the minimal spacing. * Change: In Anabatic::AutoSegment::bloatStackedStrap(), enhanced, but finally unused... * New: In Anabatic::AutoSegment::create(), use the PWireWidth when the segment is not in the preferred routing direction (and of minimal width). * New: In Anabatic::Configuration, add new getPWirewidth(), DPHorizontalWidth() and DPVerticalWidth() accessors. * Change: In AnabaticEngine::setupPreRouteds(), skip components in in "cut" material. We are only interested in objects containing some metal (happens in real mode when VIAs cuts are really there). * New: In Katana::PowerRailsPlanes::Rail::doLayout(), add an hard-coded patch that artificially enlarge the *wide wire* so the spacing for wide wire is enforced. For now, two pitches on each side for "FlexLib" gauge. * New: In Katana::Track, add support to find and correct small wire chunks so they respect the minimum area rules. Two helper functions: * ::hasSameLayerTurn(), to find if a a TrackElement as non-zero length perpandicular is same layer connected to it. * ::toFoundryGrid(), to ensure that all coordinates will be on the foundry grid (may move in a more shared location). * ::expandToMinArea(), try to expand, *in the routing direction* the too small wire so it respect the minimal area. Check for the free space in the track. Track::minExpandArea() go through all the TrackElements in the track to look for too small ones and correct them. * Change: In Katana::RoutingPlane, add an accessor to get the tracks. * New: In KatanaEngine::finalizeLayout(), add a post-treatment to find for minimal area violations. * Change: In cumulus/plugins.block.configuration.GaugeConf, add a routingBb attribute that will serve as a common reference to all the functions calculation track positions. We must not have two different reference for the core and the corona. The reference is always the corona when we working on a complete chip. * New: In cumulus/plugins.block.configuration.GaugeConf.getTrack(), Simplified and more reliable way of getting tracks positions. Use the routingBb. * New: In cumulus/plugins.block.configuration.GaugeConf.rpAccess(), Make use of getTrack() to get every metal strap on the right X/Y position. * New: In cumulus/plugins.block.configuration.GaugeConf.expandMinArea(), As those wires are left alone by the router, it is our responsability to abide by the minimal area rule here. Hence the code duplication from the router (bad). Mainly wires made for the clock tree, I mean. * Bug: In cumulus/plugins.chip.configuration.ChipConf.setupICore(), the core instance must be placed on the GCell grid, defined by the slice height (X *and* Y). * Bug: In cumulus/plugins.chip.corona.Builder(), forgot to use bigvia for the corners of the inner ring. * Bug: In cumulus/plugins.chip.pads.corona._createCoreWire(), hard-coded patch for LibreSOCIO, the power/ground connectors toward the core are too wide and can create DRC errors when put side by side. Shrink them by the minimal distance.
2020-11-23 16:07:15 -06:00
if (dir & Flags::Horizontal) {
horizontalWidth = Session::getWireWidth ( depth );
verticalWidth = Session::getPWireWidth( depth );
} else {
verticalWidth = Session::getWireWidth ( depth );
horizontalWidth = Session::getPWireWidth( depth );
}
cdebug_log(149,0) << "hW:" << DbU::getValueString(horizontalWidth).c_str()
<< "vW:" << DbU::getValueString( verticalWidth).c_str()
<< endl;
if (dir & Flags::UseNonPref) {
cdebug_log(149,0) << "swap H/W width." << endl;
std::swap( horizontalWidth, verticalWidth );
}
Katana manage wide wires, and they can also be symmetric. * New: In Anabatic::AutoContact and the derived classes, manages wide wires. The contact self dimension itself according to the segments it is connected to. Special case for the AutoContactTerminal which also read the size of the component it is anchored upon. New refresh method "updateSize()" and flag CntInvalidatedWidth. to compute the size. In AutoContactTerminal, compute the constraint box according to the width of the segment. * New: In Anabatic::AutoSegment, flags are now implemented as "static const" attributes of the class. The flags are stored into a uint64_t as they are more than 32. Added new flag "SegWide" and associated predicates. * Change: In GCellTopology::_doHChannel() and GCellTopology::_doVChannel(), uses the simpler overload of AutoSegment::create() in order to detect the wire width automatically. * New: In Katana::Manipulator, split insertToTrack() and forceToTrack() into a one-track method and a segment level method that iterate over the track span of the segment. * New: In Katana::SegmentFsm, for each cost in the table, now allow access to a specific track. So the base functions have now two parameters: "icost" and "itrack" (has a cost can have multiple tracks in the case of wide segments). * Change: In Katana::TrackElement, remove the index of the element inside it's track, as for a wide segment it will not be meaningful for the non-base track. This means that we have to use the Track::find() method each time instead. Remove the wide flag, as it is a duplicate of the one in AutoSegment. Added a getTrackCount() method to tell the number of track the segment is inserted into. Needed in the Track destroy step to delete a segment only when the last track that refers it is destroyed. Added getSymmetricAxis() to correct the computation of the symmetric base track in case of wide segment as the base track is not centered but the the leftmost one. * Change: In Track::insert() insert wide segments in their whole track span. * Change: In TrackCost, create an array of costs according to the segment track span. * Change: In TrackSegment::create(), now activate the factory and create wide segments. * Bug: In Katana::AutoSegments_Perpandicular, correct the debug indentation problem (ever shifting to the right).
2017-07-28 08:30:22 -05:00
}
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
}
DRC correct on Arlet6505 / TSMC C180. Integrate new features and bug fixes so the Arlet 6502 benchs successfully passes real DRC with reference industrial tools. Short summary: * Manage minimum area for VIAs in Katana::Tracks. * Allow different wire width for wires perpandicular to the prefered routing direction. * StackedVIAs used in the clock tree no longer assume an uniform routing grid (same offset & pitch all the way up). * Some hard-coded patches in PowerRails for FlexLib. * New: In CRL/symbolic/cmos/kite.py & cmos45/kite.py, update the RoutingLayerGauges by adding the new PWireWidth parameter. Always zero in case of symbolic layout (too fine tuning). * New: In CRL::RoutingGauge, add accessor to PWireWidth parameter. Modify the clone method. * New: In CRL::RoutingLayerGauge, add new parameter "PWireWidth" to give the width of a wire when it not drawn in the prefered routing direction. If it is set to zero, the normal width is used. * New: In CRL::PyRoutingGauge, export the updated constructor interface. It is *not* backward compatible, one must add the PWireWidth parameter in the various kite.py configuration files (in etc/). * Change: In AnabaticEngine::_gutAnabatic(), disable the minimum area detection mechanism, replaced by a more complete one in Katana::Track. Left commented out for now, but will be removed in the future. * Change: In Anabatic::AutoContact::updateLayer(), now systematically calls setLayerAndWidth() to potentially resize the VIAs. This is needed in real mode as VIAs are *not* macro-generated but have their real final size. * Change: In Anabatic::AutoContact::setLayerAndWidth(), select the width and height of the contact using the gauge wire width *and* perpandicular *wire width*. * Change: In Anabatic::AutoSegment::_initialize(), the "VIA to same cap" to PWireWidth/2, this will be the size of the VIA in the non-preferred direction at the end cap (non-square in real mode). * Change: In Anabatic::AutoSegment::getExtensionCap(), makes different cases for symbolic and real. Use raw length in real, add half the wire width in symbolic. Add a flag to get the extension cap *only*, not increased of half the minimal spacing. * Change: In Anabatic::AutoSegment::bloatStackedStrap(), enhanced, but finally unused... * New: In Anabatic::AutoSegment::create(), use the PWireWidth when the segment is not in the preferred routing direction (and of minimal width). * New: In Anabatic::Configuration, add new getPWirewidth(), DPHorizontalWidth() and DPVerticalWidth() accessors. * Change: In AnabaticEngine::setupPreRouteds(), skip components in in "cut" material. We are only interested in objects containing some metal (happens in real mode when VIAs cuts are really there). * New: In Katana::PowerRailsPlanes::Rail::doLayout(), add an hard-coded patch that artificially enlarge the *wide wire* so the spacing for wide wire is enforced. For now, two pitches on each side for "FlexLib" gauge. * New: In Katana::Track, add support to find and correct small wire chunks so they respect the minimum area rules. Two helper functions: * ::hasSameLayerTurn(), to find if a a TrackElement as non-zero length perpandicular is same layer connected to it. * ::toFoundryGrid(), to ensure that all coordinates will be on the foundry grid (may move in a more shared location). * ::expandToMinArea(), try to expand, *in the routing direction* the too small wire so it respect the minimal area. Check for the free space in the track. Track::minExpandArea() go through all the TrackElements in the track to look for too small ones and correct them. * Change: In Katana::RoutingPlane, add an accessor to get the tracks. * New: In KatanaEngine::finalizeLayout(), add a post-treatment to find for minimal area violations. * Change: In cumulus/plugins.block.configuration.GaugeConf, add a routingBb attribute that will serve as a common reference to all the functions calculation track positions. We must not have two different reference for the core and the corona. The reference is always the corona when we working on a complete chip. * New: In cumulus/plugins.block.configuration.GaugeConf.getTrack(), Simplified and more reliable way of getting tracks positions. Use the routingBb. * New: In cumulus/plugins.block.configuration.GaugeConf.rpAccess(), Make use of getTrack() to get every metal strap on the right X/Y position. * New: In cumulus/plugins.block.configuration.GaugeConf.expandMinArea(), As those wires are left alone by the router, it is our responsability to abide by the minimal area rule here. Hence the code duplication from the router (bad). Mainly wires made for the clock tree, I mean. * Bug: In cumulus/plugins.chip.configuration.ChipConf.setupICore(), the core instance must be placed on the GCell grid, defined by the slice height (X *and* Y). * Bug: In cumulus/plugins.chip.corona.Builder(), forgot to use bigvia for the corners of the inner ring. * Bug: In cumulus/plugins.chip.pads.corona._createCoreWire(), hard-coded patch for LibreSOCIO, the power/ground connectors toward the core are too wide and can create DRC errors when put side by side. Shrink them by the minimal distance.
2020-11-23 16:07:15 -06:00
cdebug_log(149,0) << "verticalWidth:" << DbU::getValueString(verticalWidth).c_str() << endl;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
AutoSegment* segment;
AutoContact* reference = source;
cdebug_log(149,0) << "Source:" << source << endl;
cdebug_log(149,0) << "Target:" << target << endl;
if (target->isFixed()) {
if (source->isFixed()) {
if ( (dir == Flags::Horizontal) and (source->getY() != target->getY()))
cerr << Warning( "Straight AutoHorizontal connecting misaligned contacts:\n"
" %s\n"
" %s"
, getString(source).c_str()
, getString(target).c_str()
) << endl;
if ( (dir == Flags::Vertical) and (source->getX() != target->getX()))
cerr << Warning( "Straight AutoVertical connecting misaligned contacts:\n"
" %s\n"
" %s"
, getString(source).c_str()
, getString(target).c_str()
) << endl;
} else
reference = target;
}
if (dir & Flags::Horizontal) {
segment = create( source
, target
, Horizontal::create( source->base()
, target->base()
, horizontalLayer
, reference->getY()
, horizontalWidth ) );
} else if (dir & Flags::Vertical) {
segment = create( source
, target
, Vertical::create( source->base()
, target->base()
, verticalLayer
, reference->getX()
, verticalWidth
) );
} else
throw Error( badSegment, getString(source).c_str(), getString(target).c_str() );
Katana manage wide wires, and they can also be symmetric. * New: In Anabatic::AutoContact and the derived classes, manages wide wires. The contact self dimension itself according to the segments it is connected to. Special case for the AutoContactTerminal which also read the size of the component it is anchored upon. New refresh method "updateSize()" and flag CntInvalidatedWidth. to compute the size. In AutoContactTerminal, compute the constraint box according to the width of the segment. * New: In Anabatic::AutoSegment, flags are now implemented as "static const" attributes of the class. The flags are stored into a uint64_t as they are more than 32. Added new flag "SegWide" and associated predicates. * Change: In GCellTopology::_doHChannel() and GCellTopology::_doVChannel(), uses the simpler overload of AutoSegment::create() in order to detect the wire width automatically. * New: In Katana::Manipulator, split insertToTrack() and forceToTrack() into a one-track method and a segment level method that iterate over the track span of the segment. * New: In Katana::SegmentFsm, for each cost in the table, now allow access to a specific track. So the base functions have now two parameters: "icost" and "itrack" (has a cost can have multiple tracks in the case of wide segments). * Change: In Katana::TrackElement, remove the index of the element inside it's track, as for a wide segment it will not be meaningful for the non-base track. This means that we have to use the Track::find() method each time instead. Remove the wide flag, as it is a duplicate of the one in AutoSegment. Added a getTrackCount() method to tell the number of track the segment is inserted into. Needed in the Track destroy step to delete a segment only when the last track that refers it is destroyed. Added getSymmetricAxis() to correct the computation of the symmetric base track in case of wide segment as the base track is not centered but the the leftmost one. * Change: In Track::insert() insert wide segments in their whole track span. * Change: In TrackCost, create an array of costs according to the segment track span. * Change: In TrackSegment::create(), now activate the factory and create wide segments. * Bug: In Katana::AutoSegments_Perpandicular, correct the debug indentation problem (ever shifting to the right).
2017-07-28 08:30:22 -05:00
if (wPitch > 1) segment->setFlags( SegWide );
Improved management of AutoContactTerminal for VH gauges (real ones). * New: In Anabatic & Katana, add the new "drag" feature. With VH gauges used by real technologies (M1-H, M2-V, M3-H) a new routing configuration that was not efficiently handled did appear. While the preferred routing direction for metal1 is officially horizontal, due to the way the standard cell must be designed, their metal1 terminals are still verticals (or punctuals). Thus, when connecting to them, we face the case where the metal1 terminal (RoutingPad) is vertical *and* the metal2 wire is also vertical. With that setup, the position of the AutoContactTerminal via12 cannot be deduced, it may range all the way over the metal1 RoutingPad. What may define it's position is the metal3 the metal2 finally connects to. That, is, when we have one horizontal (the metal3) and one vertical (the metal1 RoutingPad). The intermediate wire of metal2 can be kept to a minimum size by "dragging" the via12 close to the via23 when the metal3 wire is moved. * New: In Anabatic & Katana, problem of closely vertically aligneds RoutingPads in metal1 is managed first in PreProcess by restricting the span of the connecteds metal3 and in _makeDogleg also by restricting the span even more tightly (to the RoutingPad itself). * New: In Anabatic::AutoContactTerminal, add the "drag" support. Automatically check if the connecting segment is in the same direction as the RoutingPad, if so, sets the "SegDrag" flag. The dragging state can be known with the "::canDrag()" predicate. * New: In Anabatic::AutoHorizontal, add the "drag" support. The drag state can be known with the "::isDrag()" predicate. In "::_makeDogleg()", when making a dogleg on a dragable segment pass the drag state correctly and restrict the perpandicular span of the perpandicular to the RoutingPad (though segment user constraints). If we make a dogleg on the metal2 is it likely than we cannot go straigth out vertically from the RoutingPad, so the new perpandicular *is* restricted to the RoutingPad span. Idem for AutoVertical. * New: In Katana::Manipulator, add method "::dragMinimize()" which find a hole where to minimize a draggable segment. We finally did not use it, but keep it for potential further use. * New: In Katana::PreProcess, adds a "protectAlignedaccesses()" local function to check for vertically aligned metal1 RoutingPads, in that case setup user constraints on the metal3 segments so they cannot completly cover the other RoutingPad with metal2. We also keep a "metal2protect()" function that create a fixed segment to lock/protect a RoutingPad. Not used for now. * New: In Katana::Session, add a RoutingPad locking event mechanism. This allows us to request the creation of a locking (fixed segment) over a draggable segment. Not used for now. Lock events are processeds before all others as they create new TrackElements. * New: In Katana::Track, "::getNextFree()" and "::getPreviousFree()" method to find the nearest free interval in a Track after/before a position. * Bug: In Anabatic::AutoHorizontal::getConstraints(), merge with user constraints *only* if it's not an empty interval (as we use min/max functions). Idem for AutoVertical. * Bug: In AutoSegments_OnContacts::Locator::isValid(), the boolean test must be inverted. Seems it never worked, but we never used it until now...
2018-01-25 04:58:04 -06:00
if (source->canDrag() or target->canDrag()) segment->setFlags( SegDrag );
Upgrade of Katana detailed router to support Arlet 6502. * Change: In Hurricane::SharedName, replace the incremental Id by a hash key. This is to ensure better deterministic properties. Between use cases, additional strings may have to be allocated, shitfing the ids. Even if hash can be duplicated, we should be able to ensure that the absolute order in map table should be preserved. Supplemental strings are inserted in a way that keep the previous order. * Change: In CRL/etc/symbolic/cmos/kite.conf, add "katabatic.routingGauge" default parameter value ("sxlib"). * Change: In CRL/etc/common/technology.conf, define minimal spacing for symbolic layers too (added for METAL4 only for now). * Change: In CRL::Histogram, extend support to dynamically sized histograms. Add a text pretty print with table and pseudo-curve. * Change: In Cumulus/plugins/ClockTreePlugin, create blockage under the block corona corners so the global router do not draw wire under them. This was creating deadlock for the detailed router. When the abutment has to be computed, directly use Etesian to do it instead of duplicating the computation in the Python plugin. * New: In Etesian, as Coloquinte seems reluctant to evenly spread the standard cells, we trick it by making them bigger during the placement stage. Furthermore, we do not not uniformely increase the size of the cells but create a "bloating profile" based on cell size, cell name or it's density of terminals. Currently only two profiles are defined, "disabled" which does nothing and "nsxlib" targeted on 4 metal layer technologies (aka AMS 350nm, c35b4). * Bug: In Knik::MatrixVertex, load the default routing gauge using the configuration parameter "katabatic.routingGauge" as the default one may not be the first registered one. * New: In AnabaticEngine::setupNetDatas(), build a dynamic historgram of the nets terminal numbers. * Bug: In Anabatic::AutoContact::Invalidate(), always invalidate the contact cache when topology is invalidated. In case of multiple invalidations, if the first did not invalidate the cache, later one that may need it where not allowed to do so. The end result was correct nonetheless, but it did generate annoying error messages. * Bug: In Anabatic::AutoContactTurn::updateTopology(), bad computation of the contact's depth when delta == 2. * Bug: In Anabatic::Gcell::getCapacity(), was always returning the west edge capacity, even for the westermost GCell, should be the east edge in that case. * New: In Anabatic::AutoSegment, introduce a new measure "distance to terminal". This is the minimal number of segments separating the current one from the nearest RoutingPad. This replace the previous "strong terminal" and "weak terminal" flags. This distance is used by Katana to sort the events, we route the segments *from* the RoutingPads *outward*. The idea being that if we cannot event connect to the RoutingPad, there is no points continuing as thoses segments are the more constraineds. This gives an order close to the simple ascending metals but with better results. * New: In Anabatic::AutoSegment, introduce a new flag "Unbreakable", disable dogleg making on those segments. mainly intended for local segments directly connecteds to RoutingPads (distance == 0). * New: In Anabatic::AutoSegment, more aggressive reducing of segments. Now the only case where a segment cannot be reduced is when it is one horizontal branch in a HTee or a vertical on a VTee. Check if, when not accounted the source & target VIAs are still connex, if so, allow reducing. * New: In Anabatic::AutoContact, new state flags CntVDogleg & CntHDogleg mainly to prevent making doglegs twice on a turn contact. This is to limit over-fragmentation. If one dogleg doesn't solve the problem, making a second one will make things worse only... * Bug: In Anabatic::Configuration::selectRpcomponent(), we were choosing the component with the *smallest* span instead of the *bigger* one. * New: In Anabatic::GCell, introduce a new flag "GoStraight" to tell that no turn go be made inside those GCells. Mainly used underneath a block corona. * New: In AnabaticEngine::layerAssign(), new GCellRps & RpsInRow to manage GCells with too many terminals. Slacken at least one RoutingPad access when there is more than 8 RoutingPad in the GCell (slacken or change a vertical METAL2 (non-preferred) into a METAL3). * Change: In Anabatic::NetBuilderHV, allow the use of terminal connection in non-preferred direction. That is, vertical METAL2 directly connected to the RoutingPad (then a horizontal METAL2). This alllows for short dogleg without clutering the METAL3 layer (critical for AMS c35b4). Done in NetBuilderHV::doRp_Access(), with a new UseNonPref flag. Perform some other tweaking on METAL1 access topologies, to also minimize METAL3 use. * New: In AnabaticEngine::computeNetConstraints(), also compute the distance to RoutingPad for segments. Set the Unbreakable flag, based on the distance and segment length (local, short global or long global). New local function "propagateDistanceFromRp()". * Change: In AnabaticEngine.h, the sorting class for NetData, SparsityOrder, is modificated so net with a degree superior to 10 are sorted first, whatever their sparsity. This is to work in tandem with GlobalRouting. * New: In Katana::TrackSegmentNonPref, introduce a class to manage segment in non-preferred routing direction. Mostly intended for small METAL2 vertical directly connected to RoutingPad. Modifications to manage this new variant all through Katana. * Change: In Katana::GlobalRoute, DigitalDistance honor the GoStraight flag of the GCell. Do not make bend inside thoses GCells. * Change: In KatanaEngine::runGlobalRouter(), high degree nets (>= 10) are routed first and whitout the global routing estimation. There should be few of them so they wont create saturations and we want them as straight as possible. Detour are for long be-points. Set the saerch halo to one GCell in the initial routing stage (before ripup). * Bug: In KatanaEngine & NegociateWindow, call _computeCagedconstraints() inside NegociateWindow::run(), as segments are inserted into tracks only at that point so we cannot make the computation earlier. * Change: In Katana::Manipulator::repackPerpandiculars(), add a flag to select whether to replace the perpandiculars *after* or *before* the current segment. * Change: In Katana::NegociateWindow::NegociateOverlapCost(), when the segment is fully enclosed inside a global, the longest overlap cost is set to the shortest global hoverhang (before or after). When the cost is for a global, set an infinite cost if the overlapping segment has a RP distance less or equal to 1 (this is an access segment). * Bug: In Katana::PowerRailsPlane::Rail::doLayout(), correct computation of the segments extension cap. * New: In Katana::QueryPowerRails::addToPowerRail(), add support for Pad. * Change: In Katana/PreProcess::protectCagedTerminals(), apply the contraints to any turn connected to the first segment of the RoutingPad so the perpandicular constraints got propagated to the perpandicular segment... * Change: In RoutingEvent, cache the "distance to RP" value. * Change: In RoutingEvent::Key::compare(), sort *first* on distance to RoutingPad, then layer depth. If both distance to RoutingPad is null, then sort on segment length. * Change: In RoutingEvent::_processRepair(), try a repack perpandicular with perpandiculars first (then with perpandicular last, then give up). * Change: In SegmentFsm::bindToTrack() and moveToTrack(), set an axis hint when creating the insertion event. * Change: In SegmentFsm::_slackenStrap(), add a step through slacken between minimize and maximum slack (wihch directly end up in unimplemented). * Change: In Session::_addInsertEvent(), add an axis parameter needed when the axis of the segment is not the one of the track (case of wide segments or non-preferred direction). * Bug: In Track::_preDestroy(), bad management of the TrackElement reference count. Destroy the segment only when reaching zero... * Bug: In Track::expandFreeIneterval(), forgotten to manage case when there is a set of overlaping segments at the "end" of the track, the EndIsTrackMax was not set. * Change: In TrackCost::Compare, increase the cost when an overlaping segment is at it's ripup limit. We should try *not* to rip it up if we can. Add a dedicated flag "AtRipupLimit". * Change: In TrackElement, add proxies for isUnbreakable(), new function updateTrackSpan(). * New: In TrackFixedSegment CTOR, when a supply wire of METAL2 or above is found, make the underlying GCells "GoStraight". * New: In TrackElement::canDogleg(GCell*), check for already done perpandicular dogleg on source/target (reject if so).
2019-07-28 16:20:00 -05:00
if (dir & Flags::UseNonPref) segment->setFlags( SegNonPref );
Katana manage wide wires, and they can also be symmetric. * New: In Anabatic::AutoContact and the derived classes, manages wide wires. The contact self dimension itself according to the segments it is connected to. Special case for the AutoContactTerminal which also read the size of the component it is anchored upon. New refresh method "updateSize()" and flag CntInvalidatedWidth. to compute the size. In AutoContactTerminal, compute the constraint box according to the width of the segment. * New: In Anabatic::AutoSegment, flags are now implemented as "static const" attributes of the class. The flags are stored into a uint64_t as they are more than 32. Added new flag "SegWide" and associated predicates. * Change: In GCellTopology::_doHChannel() and GCellTopology::_doVChannel(), uses the simpler overload of AutoSegment::create() in order to detect the wire width automatically. * New: In Katana::Manipulator, split insertToTrack() and forceToTrack() into a one-track method and a segment level method that iterate over the track span of the segment. * New: In Katana::SegmentFsm, for each cost in the table, now allow access to a specific track. So the base functions have now two parameters: "icost" and "itrack" (has a cost can have multiple tracks in the case of wide segments). * Change: In Katana::TrackElement, remove the index of the element inside it's track, as for a wide segment it will not be meaningful for the non-base track. This means that we have to use the Track::find() method each time instead. Remove the wide flag, as it is a duplicate of the one in AutoSegment. Added a getTrackCount() method to tell the number of track the segment is inserted into. Needed in the Track destroy step to delete a segment only when the last track that refers it is destroyed. Added getSymmetricAxis() to correct the computation of the symmetric base track in case of wide segment as the base track is not centered but the the leftmost one. * Change: In Track::insert() insert wide segments in their whole track span. * Change: In TrackCost, create an array of costs according to the segment track span. * Change: In TrackSegment::create(), now activate the factory and create wide segments. * Bug: In Katana::AutoSegments_Perpandicular, correct the debug indentation problem (ever shifting to the right).
2017-07-28 08:30:22 -05:00
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
return segment;
}
void AutoSegment::destroy ()
{
_preDestroy ();
delete this;
}
Replace "unsigned int" by "Flags" in all AutoSegments collections. * Change: In Anabatic::AutoSegments collections, change the type of all the flags that where in "unsigned int" (32 bits) to Flags (uint64_t) as there is now more than 32 flags for functions. * New: In Ababatic::Constants, added new flag Flags::WithPerpands, which makes the number of flags tip over 32 bits, thus making mandatory to uses Flags and not unsigned int. * New: In Anabatic::AutoSegments_Perpandiculars, manage a new flag Flags::WithDoglegs to allow to propagate through global segments that are connecteds via doglegs on local segments. Meaning that there is a good chance that they could be aligned. Slighly change the way we propagate on aligned segments: no longer check for VTee or HTee, but only for same direction and layer as master. * New: In Anabatic & Katana, replace all the "int", "long" and their variants by the less implementation ambiguous "int32_t", "int64_t" (and variant). This should help to better detect bit trucation in flags. Use the type to give a hint about the flags kind: - Type "Flags", for flags shared among Anabatic & Katana functions/methods (may also appear in some objects states). - Type "uint32_t" for flags belonging to an object internal state of from Hurricane functions flags (those should be grouped in a Flag subclass in a perfect world).
2017-05-16 07:53:33 -05:00
AutoSegment* AutoSegment::getGlobalThroughDogleg ( AutoSegment* dogleg, AutoContact* from )
{
AutoContact* source = dogleg->getAutoSource();
AutoContact* target = dogleg->getAutoTarget();
if (not source->isTurn() or not target->isTurn()) return NULL;
AutoSegment* fromSegment = (source == from) ? source->getPerpandicular(dogleg) : target->getPerpandicular(dogleg);
AutoSegment* toSegment = (source != from) ? source->getPerpandicular(dogleg) : target->getPerpandicular(dogleg);
if (not toSegment->isGlobal() or (toSegment->getLayer() != fromSegment->getLayer())) return NULL;
Interval fromConstraints;
Interval toConstraints;
fromSegment->getConstraints( fromConstraints );
toSegment ->getConstraints( toConstraints );
if (not fromConstraints.intersect(toConstraints)) return NULL;
return toSegment;
}
bool AutoSegment::isTopologicalBound ( AutoSegment* seed, Flags flags )
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
{
cdebug_log(145,1) << "isTopologicalBound() - " << seed << endl;
set<AutoContact*> exploreds;
vector<AutoContact*> stack;
DbU::Unit axis;
if (flags & Flags::Superior) axis = seed->getTargetU();
else axis = seed->getSourceU();
cdebug_log(145,0) << "check for bound " << DbU::getValueString(axis) << endl;
exploreds.insert( seed->getAutoSource() );
exploreds.insert( seed->getAutoTarget() );
if (seed->getLength()) {
if (flags & Flags::Superior) stack.push_back( seed->getAutoTarget() );
else stack.push_back( seed->getAutoSource() );
} else {
stack.push_back( seed->getAutoTarget() );
stack.push_back( seed->getAutoSource() );
}
while ( not stack.empty() ) {
AutoContact* currentContact = stack.back();
stack.pop_back();
cdebug_log(145,0) << "Exploring: " << (void*)currentContact << " " << currentContact << endl;
exploreds.insert( currentContact );
if (currentContact->getAnchor()) { cdebug_tabw(145,-1); return true; }
forEach ( Component*, component, currentContact->getSlaveComponents() ) {
Segment* segment = dynamic_cast<Segment*>( *component );
if (not segment) continue;
AutoSegment* autoSegment = Session::lookup( segment );
if (not autoSegment) continue;
if (not autoSegment->getLength()) {
AutoContact* contact = autoSegment->getAutoSource();
if (contact and (contact != currentContact)) {
if (exploreds.find(contact) == exploreds.end())
stack.push_back( contact );
}
contact = autoSegment->getAutoTarget();
if (contact and (contact != currentContact)) {
if (exploreds.find(contact) == exploreds.end())
stack.push_back( contact );
}
continue;
}
Added analog type on segment NetRoutingProperty. * New: In Anabatic_AutoSegments collection, added a Flag to the constructors to allow different behavior between digital and analog modes. For "Aligneds" and "Perpandiculars" collections, now manage a new Flag WithDoglegs to follow aligned globals through local doglegs (for analog nets). Adjust the log level of collections to 144 (formerly 145). * New: In Anabatic::AutoSegment, new flag SegAnalog for segments that are part of an analog net. Note that with this flag, we reach the 32 bits limit... * Change: In Anabatic::Constants, Flags are now declared as BaseFlags objects and *not* uint64_t. This avoids overload resolution problems with arithmetical overload of the operators. The BaseFlags/Flags types are now completly "isolated" from the uint64_t, it has the advantage of showing where unwanted previous implicit conversions where occuring. * Change: In Katana::Constants, Flags values are now of BaseFlags type instead of uint64_t. * Change: In Anabatic::Dijkstra, lots of log cleanup. * Change: In Anabatic::GCell::getSide(), make the "shrink" parameter visible to allow to substract the topmost and rightmost track for axis span computation in AutoSegment::computeOptimal(). Used for analog mode. * Change: In NetRoutingState, added a flag for analog mode. Use uint32_t for the flags type. * New: In Isobar, export the NetRoutingState and NetRoutingExtension objects.
2017-05-20 05:33:12 -05:00
if (autoSegment->isHorizontal() xor (bool)(flags & Flags::Horizontal)) continue;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
cdebug_log(145,0) << "| " << autoSegment << endl;
if (flags & Flags::Superior) {
if (autoSegment->getTargetU() > axis) { cdebug_tabw(145,-1); return true; }
} else {
if (autoSegment->getSourceU() < axis) { cdebug_tabw(145,-1); return true; }
}
}
}
cdebug_tabw(145,-1);
return false;
}
#if THIS_IS_DISABLED
Replace "unsigned int" by "Flags" in all AutoSegments collections. * Change: In Anabatic::AutoSegments collections, change the type of all the flags that where in "unsigned int" (32 bits) to Flags (uint64_t) as there is now more than 32 flags for functions. * New: In Ababatic::Constants, added new flag Flags::WithPerpands, which makes the number of flags tip over 32 bits, thus making mandatory to uses Flags and not unsigned int. * New: In Anabatic::AutoSegments_Perpandiculars, manage a new flag Flags::WithDoglegs to allow to propagate through global segments that are connecteds via doglegs on local segments. Meaning that there is a good chance that they could be aligned. Slighly change the way we propagate on aligned segments: no longer check for VTee or HTee, but only for same direction and layer as master. * New: In Anabatic & Katana, replace all the "int", "long" and their variants by the less implementation ambiguous "int32_t", "int64_t" (and variant). This should help to better detect bit trucation in flags. Use the type to give a hint about the flags kind: - Type "Flags", for flags shared among Anabatic & Katana functions/methods (may also appear in some objects states). - Type "uint32_t" for flags belonging to an object internal state of from Hurricane functions flags (those should be grouped in a Flag subclass in a perfect world).
2017-05-16 07:53:33 -05:00
Flags AutoSegment::getPerpandicularState ( AutoContact* contact
, AutoSegment* source
, AutoSegment* current
, bool isHorizontalMaster
, const Layer* masterLayer )
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
{
Replace "unsigned int" by "Flags" in all AutoSegments collections. * Change: In Anabatic::AutoSegments collections, change the type of all the flags that where in "unsigned int" (32 bits) to Flags (uint64_t) as there is now more than 32 flags for functions. * New: In Ababatic::Constants, added new flag Flags::WithPerpands, which makes the number of flags tip over 32 bits, thus making mandatory to uses Flags and not unsigned int. * New: In Anabatic::AutoSegments_Perpandiculars, manage a new flag Flags::WithDoglegs to allow to propagate through global segments that are connecteds via doglegs on local segments. Meaning that there is a good chance that they could be aligned. Slighly change the way we propagate on aligned segments: no longer check for VTee or HTee, but only for same direction and layer as master. * New: In Anabatic & Katana, replace all the "int", "long" and their variants by the less implementation ambiguous "int32_t", "int64_t" (and variant). This should help to better detect bit trucation in flags. Use the type to give a hint about the flags kind: - Type "Flags", for flags shared among Anabatic & Katana functions/methods (may also appear in some objects states). - Type "uint32_t" for flags belonging to an object internal state of from Hurricane functions flags (those should be grouped in a Flag subclass in a perfect world).
2017-05-16 07:53:33 -05:00
Flags state = Flags::NoFlags;
Anabatic transient commit 11. Merge Katabatic into Anabatic. * New: In Anabatic, merge in the Katabatic tool. The merge is not complete yet. Only basic capabilities have been moved.
2016-07-18 07:48:37 -05:00
bool sourcePerpandicular = arePerpandiculars ( isHorizontalMaster, source );
bool currentPerpandicular = arePerpandiculars ( isHorizontalMaster, current );
bool contactAlignate
= (contact->isHAlignate() and current->isHorizontal() and isHorizontalMaster)
or (contact->isVAlignate() and !current->isHorizontal() and !isHorizontalMaster);
if ( not currentPerpandicular and masterLayer and (masterLayer != current->getLayer()) )
state |= ParallelAndLayerChange;
if ( currentPerpandicular and !current->isCollapsed() )
state |= PerpandicularAny;
if ( sourcePerpandicular ) {
// Source segment is perpandicular to master.
if ( currentPerpandicular and !current->isCollapsed() )
state |= PerpandicularIndirect;
} else {
// Source segment is parallel to master.
if ( not (currentPerpandicular and current->isCollapsed()) and not contactAlignate ) {
// Current segment is parallel OR expanded.
state |= ParallelOrExpanded;
}
}
return state;
}
#endif
void AutoSegment::getTopologicalInfos ( AutoSegment* seed
, vector<AutoSegment*>& aligneds
, vector<AutoSegment*>& perpandiculars
, DbU::Unit& leftBound
, DbU::Unit& rightBound
)
{
cdebug_log(145,1) << "getTopologicalInfos() - " << seed << endl;
leftBound = DbU::Max;
rightBound = DbU::Min;
AutoSegmentStack stack;
stack.push( seed->getAutoSource(), seed );
stack.push( seed->getAutoTarget(), seed );
while ( not stack.isEmpty() ) {
AutoContact* sourceContact = stack.getAutoContact();
AutoSegment* sourceSegment = stack.getAutoSegment();
stack.pop();
DbU::Unit constraint;
if (seed->isHorizontal()) constraint = sourceContact->getCBXMax();
else constraint = sourceContact->getCBYMax();
if (constraint < leftBound) leftBound = constraint;
if (seed->isHorizontal()) constraint = sourceContact->getCBXMin();
else constraint = sourceContact->getCBYMin();
if (constraint > rightBound) rightBound = constraint;
cdebug_log(149,0) << "Segments of: " << sourceContact << endl;
LocatorHelper helper (sourceContact, Flags::Horizontal|Flags::WithPerpands);
for ( ; helper.isValid() ; helper.progress() ) {
AutoSegment* currentSegment = helper.getSegment();
cdebug_log(149,0) << "Looking for: " << currentSegment << endl;
if (currentSegment == sourceSegment) continue;
if (AutoSegment::areAlignedsAndDiffLayer(currentSegment,seed)) {
cerr << Error("Aligned segments not in same layer\n"
" %s\n"
" %s."
,getString(seed).c_str()
,getString(currentSegment).c_str()) << endl;
continue;
}
if (AutoSegment::areAligneds(currentSegment,seed)) {
aligneds.push_back( currentSegment );
AutoContact* targetContact = currentSegment->getOppositeAnchor( sourceContact );
cdebug_log(149,0) << "Target: " << targetContact << endl;
if (targetContact) {
if ( (seed->isHorizontal() and sourceContact->isHTee())
or (seed->isVertical () and sourceContact->isVTee()) ) {
cdebug_log(149,0) << "Stacking target. " << endl;
stack.push( targetContact, currentSegment );
}
}
} else {
cdebug_log(149,0) << "| perpandicular " << currentSegment << endl;
perpandiculars.push_back( currentSegment );
}
}
}
cdebug_tabw(145,-1);
}
int AutoSegment::getTerminalCount ( AutoSegment* seed, vector<AutoSegment*>& collapseds )
{
cdebug_log(145,0) << "getTerminalCount() - " << seed << " (+collapseds)" << endl;
int count = 0;
for ( size_t i=0 ; i < collapseds.size() ; i++ ) {
if (collapseds[i]->isStrongTerminal())
count++;
}
if (seed->getAutoSource()->isTerminal()) count++;
if (seed->getAutoTarget()->isTerminal()) count++;
return count;
}
} // End of Anabatic namespace.