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/NetBuilderHV.cpp

1913 lines
80 KiB
C++
Raw Normal View History

Modularize the global to detail routing converter. * New: In Anabatic, rename GCellTopology to NetBuilder, expose the class (no longer in a local namespace). NetBuilder become a base class performing the walkthrough the Net tree. Derived class are tasked to build routing for specific gauge. NetBuilderHV is created to manage all gauge with metal2 horizontal and metal3 vertical. In our terminolgy we consider that the first routing metal is metal2. Metal1 is used inside the standard cells.
2017-12-16 17:13:19 -06:00
// -*- 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
Modularize the global to detail routing converter. * New: In Anabatic, rename GCellTopology to NetBuilder, expose the class (no longer in a local namespace). NetBuilder become a base class performing the walkthrough the Net tree. Derived class are tasked to build routing for specific gauge. NetBuilderHV is created to manage all gauge with metal2 horizontal and metal3 vertical. In our terminolgy we consider that the first routing metal is metal2. Metal1 is used inside the standard cells.
2017-12-16 17:13:19 -06:00
//
// +-----------------------------------------------------------------+
// | 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 |
// | =============================================================== |
Complete the modularization of the detail routing converter. * New: In Anabatic::NetBuilder, set all the attributes as private and create accessors and mutators. Finish virtualising all GCell build methods and transfer them into the NetBuilderHV class. Build methods now return a boolean to tell if the GCell was processed or not, to allow cascading in the "big switch". Reorganise the "big switch" in separate sections only partially cascading. Truly fuse the *big switch* for channel routing and over-the-cells. Create a new method "_do_globalSegment()" to delagate the drawing of global segments between two GCell to the derived classes.
2017-12-18 11:15:14 -06:00
// | C++ Module : "./NetBuilderHV.cpp" |
Modularize the global to detail routing converter. * New: In Anabatic, rename GCellTopology to NetBuilder, expose the class (no longer in a local namespace). NetBuilder become a base class performing the walkthrough the Net tree. Derived class are tasked to build routing for specific gauge. NetBuilderHV is created to manage all gauge with metal2 horizontal and metal3 vertical. In our terminolgy we consider that the first routing metal is metal2. Metal1 is used inside the standard cells.
2017-12-16 17:13:19 -06:00
// +-----------------------------------------------------------------+
#include <cstdlib>
#include <sstream>
#include "hurricane/Bug.h"
#include "hurricane/Breakpoint.h"
#include "hurricane/Error.h"
#include "hurricane/Warning.h"
#include "hurricane/DebugSession.h"
#include "hurricane/Layer.h"
#include "hurricane/BasicLayer.h"
#include "hurricane/RegularLayer.h"
#include "hurricane/Technology.h"
#include "hurricane/DataBase.h"
#include "hurricane/Net.h"
#include "hurricane/NetExternalComponents.h"
#include "hurricane/NetRoutingProperty.h"
#include "hurricane/RoutingPad.h"
#include "hurricane/RoutingPads.h"
#include "hurricane/Pad.h"
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
#include "hurricane/Pin.h"
Modularize the global to detail routing converter. * New: In Anabatic, rename GCellTopology to NetBuilder, expose the class (no longer in a local namespace). NetBuilder become a base class performing the walkthrough the Net tree. Derived class are tasked to build routing for specific gauge. NetBuilderHV is created to manage all gauge with metal2 horizontal and metal3 vertical. In our terminolgy we consider that the first routing metal is metal2. Metal1 is used inside the standard cells.
2017-12-16 17:13:19 -06:00
#include "hurricane/Plug.h"
#include "hurricane/Cell.h"
#include "hurricane/Instance.h"
#include "hurricane/Vertical.h"
#include "hurricane/Horizontal.h"
#include "crlcore/AllianceFramework.h"
#include "crlcore/RoutingGauge.h"
#include "crlcore/Measures.h"
#include "anabatic/AutoContactTerminal.h"
#include "anabatic/AutoContactTurn.h"
#include "anabatic/AutoContactHTee.h"
#include "anabatic/AutoContactVTee.h"
#include "anabatic/AutoSegment.h"
#include "anabatic/NetBuilderHV.h"
#include "anabatic/AnabaticEngine.h"
namespace Anabatic {
using std::swap;
Complete the modularization of the detail routing converter. * New: In Anabatic::NetBuilder, set all the attributes as private and create accessors and mutators. Finish virtualising all GCell build methods and transfer them into the NetBuilderHV class. Build methods now return a boolean to tell if the GCell was processed or not, to allow cascading in the "big switch". Reorganise the "big switch" in separate sections only partially cascading. Truly fuse the *big switch* for channel routing and over-the-cells. Create a new method "_do_globalSegment()" to delagate the drawing of global segments between two GCell to the derived classes.
2017-12-18 11:15:14 -06:00
using Hurricane::Transformation;
using Hurricane::Warning;
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
using Hurricane::Error;
using Hurricane::Pin;
Modularize the global to detail routing converter. * New: In Anabatic, rename GCellTopology to NetBuilder, expose the class (no longer in a local namespace). NetBuilder become a base class performing the walkthrough the Net tree. Derived class are tasked to build routing for specific gauge. NetBuilderHV is created to manage all gauge with metal2 horizontal and metal3 vertical. In our terminolgy we consider that the first routing metal is metal2. Metal1 is used inside the standard cells.
2017-12-16 17:13:19 -06:00
NetBuilderHV::NetBuilderHV ()
: NetBuilder()
{ }
NetBuilderHV::~NetBuilderHV () { }
void NetBuilderHV::doRp_AutoContacts ( GCell* gcell
, Component* rp
, AutoContact*& source
, AutoContact*& target
, uint64_t flags
)
{
Complete the modularization of the detail routing converter. * New: In Anabatic::NetBuilder, set all the attributes as private and create accessors and mutators. Finish virtualising all GCell build methods and transfer them into the NetBuilderHV class. Build methods now return a boolean to tell if the GCell was processed or not, to allow cascading in the "big switch". Reorganise the "big switch" in separate sections only partially cascading. Truly fuse the *big switch* for channel routing and over-the-cells. Create a new method "_do_globalSegment()" to delagate the drawing of global segments between two GCell to the derived classes.
2017-12-18 11:15:14 -06:00
cdebug_log(145,1) << getTypeName() << "::doRp_AutoContacts()" << endl;
Modularize the global to detail routing converter. * New: In Anabatic, rename GCellTopology to NetBuilder, expose the class (no longer in a local namespace). NetBuilder become a base class performing the walkthrough the Net tree. Derived class are tasked to build routing for specific gauge. NetBuilderHV is created to manage all gauge with metal2 horizontal and metal3 vertical. In our terminolgy we consider that the first routing metal is metal2. Metal1 is used inside the standard cells.
2017-12-16 17:13:19 -06:00
cdebug_log(145,0) << rp << 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
RoutingPad* rrp = dynamic_cast<RoutingPad*>( rp );
if (not rrp) {
cerr << Warning( "%s::doRp_AutoContacts(): %s is *not* a RoutingPad."
, getTypeName().c_str()
, getString(rp).c_str() ) << endl;
}
Modularize the global to detail routing converter. * New: In Anabatic, rename GCellTopology to NetBuilder, expose the class (no longer in a local namespace). NetBuilder become a base class performing the walkthrough the Net tree. Derived class are tasked to build routing for specific gauge. NetBuilderHV is created to manage all gauge with metal2 horizontal and metal3 vertical. In our terminolgy we consider that the first routing metal is metal2. Metal1 is used inside the standard cells.
2017-12-16 17:13:19 -06:00
source = target = NULL;
Point sourcePosition;
Point targetPosition;
const Layer* rpLayer = rp->getLayer();
size_t rpDepth = Session::getLayerDepth( rp->getLayer() );
Flags direction = Session::getDirection ( rpDepth );
DbU::Unit viaSide = Session::getViaWidth ( rpDepth );
getPositions( rp, sourcePosition, targetPosition );
GCell* sourceGCell = Session::getAnabatic()->getGCellUnder( sourcePosition );
GCell* targetGCell = Session::getAnabatic()->getGCellUnder( targetPosition );
if (rpDepth == 0) {
rpLayer = Session::getContactLayer(0);
direction = Flags::Horizontal;
viaSide = Session::getViaWidth( rpDepth );
}
// Non-M1 terminal or punctual M1 protections.
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 ( ((rpDepth != 0) or (sourcePosition == targetPosition)) and not (flags & NoProtect) ) {
Modularize the global to detail routing converter. * New: In Anabatic, rename GCellTopology to NetBuilder, expose the class (no longer in a local namespace). NetBuilder become a base class performing the walkthrough the Net tree. Derived class are tasked to build routing for specific gauge. NetBuilderHV is created to manage all gauge with metal2 horizontal and metal3 vertical. In our terminolgy we consider that the first routing metal is metal2. Metal1 is used inside the standard cells.
2017-12-16 17:13:19 -06:00
map<Component*,AutoSegment*>::iterator irp = getRpLookup().find( rp );
if (irp == getRpLookup().end()) {
Correct managment of macro block regarding to P&R. * Change: In NetBuilder::getPositions(), ordering of source/target points is now integrated to this function instead of left to the caller. In case of real (non-symbolic) routing gauge, skrink the ends of half the wire width. * Change: In NetBuilderHV::doRp_AutoContacts(), in case of non-METAL1 RoutingPad, put the axis of the segment on the nearest track. Issue a warning if we have to shift, as it may be a potential source of routing problems. * New: Anabatic::Session::getNearestTrackaAxis(), proxy to compute track positions, knowing the design abutment box. * Bug: In Katana::PreProcess::propagateCagedConstraints(), when looking at all the slave components anchoreds on a RoutingPad, if they do not have an AutoSegment lookup, skip them instead of crashing. * New: In cumulus/plugins.block.Block.placeMacro(), new method to place a macro cell, partly delegating to the Macro block wrapper. Must be called *after* both core and corona abutment boxes have been set. Adjust the macro block position so the METAL2 & METAL3 pins are exactly on pitch regarding the full routing grid. The reference being the corona. A shift, less than one pitch may be applied, leading in some cases of overlapping abutment boxes. But this shouldn't be a problem. The macro to place is designated through a path of instances names, rooted at the *core* (not the corona). Meaning that the head instance must be one of the core. * Change: In cumulus/plugins.chip.Chip, the complete chip I/O pads plus corona and core placement is moved out from doPnR() and put into doChipFloorplan(). It is now mandatory to call this method *before* doPnR(). Those methods are now cleanly separated so we can perform macro block placement or any inner core floorplaning operations between them. * Change: In cumulus/plugins.macro.Macro, instead of creating large pads for the I/O pins so whatever the block position, they will be under a grid point, create a simple dogleg to put them on grid. To ensure that they are "on grid", the block pins must be in METAL2 (horizontal E/W) or METAL3 (vertical N/S) and the block is assumed to be placed so the bottom left corner of it's abutment box is exactly on one grid point for M2/M3. This should be done by Block.placeMacro().
2021-03-14 10:16:54 -05:00
Box rpBb = rp->getBoundingBox();
if (rpDepth != 0) {
DbU::Unit hborder = 0;
DbU::Unit vborder = 0;
if (direction.contains(Flags::Horizontal)) {
vborder = Session::getWireWidth(rpDepth) / 2;
rpBb.inflate( hborder, vborder );
if (not rpBb.isEmpty()) {
DbU::Unit trackAxis = Session::getNearestTrackAxis( rp->getLayer()
, sourcePosition.getY()
, Constant::Nearest );
if (trackAxis != sourcePosition.getY()) {
cerr << Warning( "NetBuilderHV::doRp_AutoContacts(): Adjust Y position to nearest H track of %s\n"
" (%s -> %s)"
, getString(rp).c_str()
, DbU::getValueString(sourcePosition.getY()).c_str()
, DbU::getValueString(trackAxis).c_str()
) << endl;
sourcePosition.setY( trackAxis );
targetPosition.setY( trackAxis );
}
} else {
cdebug_log(145,0) << "rpBb is too narrow to adjust: " << rp->getBoundingBox() << endl;
}
} else {
hborder = Session::getWireWidth(rpDepth) / 2;
rpBb.inflate( hborder, vborder );
if (not rpBb.isEmpty()) {
DbU::Unit trackAxis = Session::getNearestTrackAxis( rp->getLayer()
, sourcePosition.getX()
, Constant::Nearest );
if (trackAxis != sourcePosition.getX()) {
cerr << Warning( "NetBuilderHV::doRp_AutoContacts(): Adjust X position to nearest V track of %s\n"
" (%s -> %s)"
, getString(rp).c_str()
, DbU::getValueString(sourcePosition.getX()).c_str()
, DbU::getValueString(trackAxis).c_str()
) << endl;
sourcePosition.setX( trackAxis );
targetPosition.setX( trackAxis );
}
} else {
cdebug_log(145,0) << "rpBb is too narrow to adjust: " << rp->getBoundingBox() << endl;
}
}
}
Modularize the global to detail routing converter. * New: In Anabatic, rename GCellTopology to NetBuilder, expose the class (no longer in a local namespace). NetBuilder become a base class performing the walkthrough the Net tree. Derived class are tasked to build routing for specific gauge. NetBuilderHV is created to manage all gauge with metal2 horizontal and metal3 vertical. In our terminolgy we consider that the first routing metal is metal2. Metal1 is used inside the standard cells.
2017-12-16 17:13:19 -06:00
AutoContact* sourceProtect = AutoContactTerminal::create( sourceGCell
, rp
, rpLayer
, sourcePosition
, viaSide, viaSide
);
AutoContact* targetProtect = AutoContactTerminal::create( targetGCell
, rp
, rpLayer
, targetPosition
, viaSide, viaSide
);
sourceProtect->setFlags( CntFixed );
targetProtect->setFlags( CntFixed );
Allow the detailed routing to be build from non-bottom P&R layers. * New: In Anabatic::AutoSegment::create(), allow the created segment to be in any supported routing layer, and not only the bottom H & V. Modifications impact the *two* overload of the function. * Change: In Anabatic::NetBuilderHV::doRp_AutoContacts(), for punctual METAL1, the protection has to be in METAL2. Bump the layer depth to correctly use the updated verstion of AutoSegment::create(). * Change: In AnabaticEngine::checkPlacement(), for the Pin, check that it's layer is in the routing gauge before anything else. * New: In Katana::NegociateWindow::createTrackSegment(), if the track nearest the segment axis (refTrack) do not exists, call a breakpoint just before crashing.
2021-12-11 12:51:24 -06:00
if (rpDepth == 0) rpDepth = 1;
AutoSegment* segment = AutoSegment::create( sourceProtect, targetProtect, direction, rpDepth );
Modularize the global to detail routing converter. * New: In Anabatic, rename GCellTopology to NetBuilder, expose the class (no longer in a local namespace). NetBuilder become a base class performing the walkthrough the Net tree. Derived class are tasked to build routing for specific gauge. NetBuilderHV is created to manage all gauge with metal2 horizontal and metal3 vertical. In our terminolgy we consider that the first routing metal is metal2. Metal1 is used inside the standard cells.
2017-12-16 17:13:19 -06:00
segment->setFlags( AutoSegment::SegFixed );
getRpLookup().insert( make_pair(rp,segment) );
}
}
if (sourcePosition != targetPosition) {
if (flags & DoSourceContact)
source = AutoContactTerminal::create( sourceGCell
, rp
, rpLayer
, sourcePosition
, viaSide, viaSide
);
if (flags & DoTargetContact)
target = AutoContactTerminal::create( targetGCell
, rp
, rpLayer
, targetPosition
, viaSide, viaSide
);
}
if (not source and not target) {
source = target = AutoContactTerminal::create( gcell
, rp
, rpLayer
, rp->getCenter()
, viaSide, viaSide
);
}
cdebug_tabw(145,-1);
return;
}
AutoContact* NetBuilderHV::doRp_Access ( GCell* gcell, Component* rp, uint64_t flags )
{
Complete the modularization of the detail routing converter. * New: In Anabatic::NetBuilder, set all the attributes as private and create accessors and mutators. Finish virtualising all GCell build methods and transfer them into the NetBuilderHV class. Build methods now return a boolean to tell if the GCell was processed or not, to allow cascading in the "big switch". Reorganise the "big switch" in separate sections only partially cascading. Truly fuse the *big switch* for channel routing and over-the-cells. Create a new method "_do_globalSegment()" to delagate the drawing of global segments between two GCell to the derived classes.
2017-12-18 11:15:14 -06:00
cdebug_log(145,1) << getTypeName() << "::doRp_Access() - flags:" << flags << endl;
Modularize the global to detail routing converter. * New: In Anabatic, rename GCellTopology to NetBuilder, expose the class (no longer in a local namespace). NetBuilder become a base class performing the walkthrough the Net tree. Derived class are tasked to build routing for specific gauge. NetBuilderHV is created to manage all gauge with metal2 horizontal and metal3 vertical. In our terminolgy we consider that the first routing metal is metal2. Metal1 is used inside the standard cells.
2017-12-16 17:13:19 -06:00
Partial support for offgrid METAL3 RoutingPad in Anabatic::NetBuilderHV. * New: In Anabatic::NetBuilder::_do_1G_1M3(), RoutingPad in METAL3 from blocks are most likely to be offgrid in real mode, we must account for that case. When an offgrid METAL3 is found, a strap of METAL2 is added, as it may be less than one pitch, it will be reduced into METAL3 often. * New: In Anabatic::NetBuilder::doRp_Access(), Support for offgrid METAL2 is added but not enabled yet. Seems to need more polishing.
2021-02-17 16:34:17 -06:00
size_t rpDepth = Session::getLayerDepth( rp->getLayer() );
AutoContact* rpSourceContact = NULL;
AutoContact* rpContactTarget = NULL;
Modularize the global to detail routing converter. * New: In Anabatic, rename GCellTopology to NetBuilder, expose the class (no longer in a local namespace). NetBuilder become a base class performing the walkthrough the Net tree. Derived class are tasked to build routing for specific gauge. NetBuilderHV is created to manage all gauge with metal2 horizontal and metal3 vertical. In our terminolgy we consider that the first routing metal is metal2. Metal1 is used inside the standard cells.
2017-12-16 17:13:19 -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
Flags useNonPref = Flags::NoFlags;
if (flags & UseNonPref) useNonPref |= Flags::UseNonPref;
Modularize the global to detail routing converter. * New: In Anabatic, rename GCellTopology to NetBuilder, expose the class (no longer in a local namespace). NetBuilder become a base class performing the walkthrough the Net tree. Derived class are tasked to build routing for specific gauge. NetBuilderHV is created to manage all gauge with metal2 horizontal and metal3 vertical. In our terminolgy we consider that the first routing metal is metal2. Metal1 is used inside the standard cells.
2017-12-16 17:13:19 -06:00
flags |= checkRoutingPadSize( rp );
Various bug corrections to pass the alliance-check-toolkit reference benchs. * New: In Isobar::PyCell, export the isRouted() and setRouted() to the Python interface. * Bug: In CRL::Entity::parseEntity(), check that the closing parenthesis is the last character of the net name. Issue a more relevant error message. * In Anabatic::NetBuilder::_do_xG() and all other unimplemented methods, throw an error if called from a derived classes instead of just issuing a message in the debug stream. Avoid later core dumps... * In Anabatic::NetBuilderHV, implement the builders for GCells with one pin. Needed to support chip/corona routing. * Bug: In Cumulus/plugins/Chip.py, check that coronaCk exists before using it. * New: In Cumulus/plugins/PadsCorona/Side._placePad(), when routing a design with symbolic pads, export the chip external "pad" connectors to be able to perform a lvx (otherwise cougar do not create external nets). In Corona._createCoreWire(), set the minimal gap between the pads and the corona to 6 pitches. Empirical value to avoid DRC errors with symbolic pads (pxlib). When successufully done, mark the Corona cell as routed. * Bug: In Cumulus/plugins/Core2Chip.IoNet, the regex for vectorizet net was wrong, it was allowing only one digit in the index. * Bug: In Cumulus/plugins/Core2Chip.cmos, correct management of pad & corona clock nets. Correct connexion between vdde/vddi. * Bug: In Unicorn/cgt.py, forgot to execute scripts when in text mode.
2019-05-24 16:57:22 -05:00
doRp_AutoContacts( gcell, rp, rpSourceContact, rpContactTarget, flags );
Modularize the global to detail routing converter. * New: In Anabatic, rename GCellTopology to NetBuilder, expose the class (no longer in a local namespace). NetBuilder become a base class performing the walkthrough the Net tree. Derived class are tasked to build routing for specific gauge. NetBuilderHV is created to manage all gauge with metal2 horizontal and metal3 vertical. In our terminolgy we consider that the first routing metal is metal2. Metal1 is used inside the standard cells.
2017-12-16 17:13:19 -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
cdebug_log(145,0) << "flags: " << flags << endl;
if ((rpDepth == 0) or (rpDepth == 2)) {
cdebug_log(145,0) << "case: METAL1 or METAL3 RoutingPad." << endl;
if (flags & HAccess) {
cdebug_log(145,0) << "case: HAccess" << endl;
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 ( ((flags & VSmall) and not ((flags & UseNonPref))) or (flags & Punctual) ) {
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(145,0) << "case: VSmall and *not* UseNonPref" << endl;
AutoContact* subContact1 = AutoContactTurn::create( gcell, rp->getNet(), Session::getContactLayer(1) );
AutoSegment::create( rpSourceContact, subContact1, Flags::Horizontal );
rpSourceContact = subContact1;
flags &= ~UseNonPref;
useNonPref.reset( Flags::UseNonPref );
}
if (flags & (VSmall|UseNonPref)) {
cdebug_log(145,0) << "case: UseNonPref" << endl;
AutoContact* subContact1 = AutoContactTurn::create( gcell, rp->getNet(), Session::getContactLayer(1) );
AutoSegment::create( rpSourceContact, subContact1, Flags::Vertical|useNonPref );
rpSourceContact = subContact1;
}
} else {
if (flags & HSmall) {
cdebug_log(145,0) << "case: HSmall" << endl;
AutoContact* subContact1 = rpSourceContact;
AutoContact* subContact2 = AutoContactTurn::create( gcell, rp->getNet(), Session::getContactLayer(1) );
AutoSegment::create( subContact1, subContact2, Flags::Horizontal );
rpSourceContact = subContact2;
if (flags & Punctual) {
cdebug_log(145,0) << "case: HSmall + Punctual" << endl;
subContact1 = subContact2;
subContact2 = AutoContactTurn::create( gcell, rp->getNet(), Session::getContactLayer(1) );
AutoSegment::create( subContact1, subContact2, Flags::Vertical );
subContact1 = subContact2;
subContact2 = AutoContactTurn::create( gcell, rp->getNet(), Session::getContactLayer(1) );
AutoSegment::create( subContact1, subContact2, Flags::Horizontal );
rpSourceContact = subContact2;
}
}
Modularize the global to detail routing converter. * New: In Anabatic, rename GCellTopology to NetBuilder, expose the class (no longer in a local namespace). NetBuilder become a base class performing the walkthrough the Net tree. Derived class are tasked to build routing for specific gauge. NetBuilderHV is created to manage all gauge with metal2 horizontal and metal3 vertical. In our terminolgy we consider that the first routing metal is metal2. Metal1 is used inside the standard cells.
2017-12-16 17:13:19 -06:00
}
} 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
cdebug_log(145,0) << "case: METAL2." << endl;
AutoContact* subContact1 = NULL;
if (flags & HAccess) {
if (flags & HAccessEW)
subContact1 = AutoContactHTee::create( gcell, rp->getNet(), Session::getContactLayer(1) );
else
subContact1 = AutoContactTurn::create( gcell, rp->getNet(), Session::getContactLayer(1) );
AutoSegment::create( rpSourceContact, subContact1, Flags::Vertical );
} else {
Partial support for offgrid METAL3 RoutingPad in Anabatic::NetBuilderHV. * New: In Anabatic::NetBuilder::_do_1G_1M3(), RoutingPad in METAL3 from blocks are most likely to be offgrid in real mode, we must account for that case. When an offgrid METAL3 is found, a strap of METAL2 is added, as it may be less than one pitch, it will be reduced into METAL3 often. * New: In Anabatic::NetBuilder::doRp_Access(), Support for offgrid METAL2 is added but not enabled yet. Seems to need more polishing.
2021-02-17 16:34:17 -06:00
#if OFFGRID_M2_DISABLED
Box cellAb = getAnabatic()->getCell()->getAbutmentBox();
RoutingLayerGauge* lgM2 = Session::getLayerGauge( 1 );
DbU::Unit trackAxis = lgM2->getTrackPosition( cellAb.getYMin()
, cellAb.getYMax()
, rp->getY()
, Constant::Nearest );
bool offGrid = (trackAxis != rp->getY());
if (offGrid) {
cdebug_log(145,0) << "Off grid, Misaligned M2 RoutingPad, add vertical strap" << endl;
subContact1 = AutoContactTurn::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( rpSourceContact, subContact1, Flags::Vertical );
rpSourceContact = subContact1;
}
#endif
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
subContact1 = AutoContactTurn::create( gcell, rp->getNet(), Session::getContactLayer(1) );
Various bug corrections to pass the alliance-check-toolkit reference benchs. * New: In Isobar::PyCell, export the isRouted() and setRouted() to the Python interface. * Bug: In CRL::Entity::parseEntity(), check that the closing parenthesis is the last character of the net name. Issue a more relevant error message. * In Anabatic::NetBuilder::_do_xG() and all other unimplemented methods, throw an error if called from a derived classes instead of just issuing a message in the debug stream. Avoid later core dumps... * In Anabatic::NetBuilderHV, implement the builders for GCells with one pin. Needed to support chip/corona routing. * Bug: In Cumulus/plugins/Chip.py, check that coronaCk exists before using it. * New: In Cumulus/plugins/PadsCorona/Side._placePad(), when routing a design with symbolic pads, export the chip external "pad" connectors to be able to perform a lvx (otherwise cougar do not create external nets). In Corona._createCoreWire(), set the minimal gap between the pads and the corona to 6 pitches. Empirical value to avoid DRC errors with symbolic pads (pxlib). When successufully done, mark the Corona cell as routed. * Bug: In Cumulus/plugins/Core2Chip.IoNet, the regex for vectorizet net was wrong, it was allowing only one digit in the index. * Bug: In Cumulus/plugins/Core2Chip.cmos, correct management of pad & corona clock nets. Correct connexion between vdde/vddi. * Bug: In Unicorn/cgt.py, forgot to execute scripts when in text mode.
2019-05-24 16:57:22 -05:00
AutoSegment::create( rpSourceContact, subContact1, Flags::Horizontal );
Modularize the global to detail routing converter. * New: In Anabatic, rename GCellTopology to NetBuilder, expose the class (no longer in a local namespace). NetBuilder become a base class performing the walkthrough the Net tree. Derived class are tasked to build routing for specific gauge. NetBuilderHV is created to manage all gauge with metal2 horizontal and metal3 vertical. In our terminolgy we consider that the first routing metal is metal2. Metal1 is used inside the standard cells.
2017-12-16 17:13:19 -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
rpSourceContact = subContact1;
Modularize the global to detail routing converter. * New: In Anabatic, rename GCellTopology to NetBuilder, expose the class (no longer in a local namespace). NetBuilder become a base class performing the walkthrough the Net tree. Derived class are tasked to build routing for specific gauge. NetBuilderHV is created to manage all gauge with metal2 horizontal and metal3 vertical. In our terminolgy we consider that the first routing metal is metal2. Metal1 is used inside the standard cells.
2017-12-16 17:13:19 -06:00
}
cdebug_tabw(145,-1);
Various bug corrections to pass the alliance-check-toolkit reference benchs. * New: In Isobar::PyCell, export the isRouted() and setRouted() to the Python interface. * Bug: In CRL::Entity::parseEntity(), check that the closing parenthesis is the last character of the net name. Issue a more relevant error message. * In Anabatic::NetBuilder::_do_xG() and all other unimplemented methods, throw an error if called from a derived classes instead of just issuing a message in the debug stream. Avoid later core dumps... * In Anabatic::NetBuilderHV, implement the builders for GCells with one pin. Needed to support chip/corona routing. * Bug: In Cumulus/plugins/Chip.py, check that coronaCk exists before using it. * New: In Cumulus/plugins/PadsCorona/Side._placePad(), when routing a design with symbolic pads, export the chip external "pad" connectors to be able to perform a lvx (otherwise cougar do not create external nets). In Corona._createCoreWire(), set the minimal gap between the pads and the corona to 6 pitches. Empirical value to avoid DRC errors with symbolic pads (pxlib). When successufully done, mark the Corona cell as routed. * Bug: In Cumulus/plugins/Core2Chip.IoNet, the regex for vectorizet net was wrong, it was allowing only one digit in the index. * Bug: In Cumulus/plugins/Core2Chip.cmos, correct management of pad & corona clock nets. Correct connexion between vdde/vddi. * Bug: In Unicorn/cgt.py, forgot to execute scripts when in text mode.
2019-05-24 16:57:22 -05:00
return rpSourceContact;
Modularize the global to detail routing converter. * New: In Anabatic, rename GCellTopology to NetBuilder, expose the class (no longer in a local namespace). NetBuilder become a base class performing the walkthrough the Net tree. Derived class are tasked to build routing for specific gauge. NetBuilderHV is created to manage all gauge with metal2 horizontal and metal3 vertical. In our terminolgy we consider that the first routing metal is metal2. Metal1 is used inside the standard cells.
2017-12-16 17:13:19 -06:00
}
First working recursive place & route (Libre-SOC ALU16 benchmark). * New: In Hurricane::Cell::isLeaf(), a leaf cell was defined as one without instances. Now it can be software defined with the "FlattenLeaf" state. If it is flagged "flatten leaf" is will be considered as a leaf by the trans-hierarchical walkthrough. This flag is also set for Cells in the Alliance CATAL. * New: In Hurricane::Cell_LeafInstanceOccurrences::progress(), add a test to prune leaf cells. * Bug: In Hurricane::getComponentOccurrences(), when calling the locator constructor, the "withLeafCells" argument was forgotten and was getting the value of "doExtraction" (wich, fortunately, was the same). * Change: In Hurricane::HypreNet_LeafPlugOccurrences::progress(), prune the leaf cells instead of the terminal ones. * Change: In Hurricane::HypreNet_ComponentOccurrences::progress(), prune the leaf cells instead of the terminal ones. * New: In Hurricane::Cell, add a "useFlattenLeaf" mode to take into account or ignore the "FlattenLeaf" flag when doing trans- hierarchical walkthrough. Not sure if it shouldn't be systematic. * New: In Etesian::toColoquinte() and other, activate the "FlattenLeaf" mode for all walkthough. So we can stop at already placed sub-blocks. * Bug: In Etesian::toColoquinte(), when placed sub-block where present, the number of instances was under estimated, so the instance array was oversized and the end went unitialized. So we were getting all kind of strange behavior from Coloquinte... Now throw an exception if the number of instances differs from the computed size. * Bug: In Etesian::toColoquinte(), as for the instance array, now correctly compute the size with the ecluded nets (supplies, clocks). * Bug: In Etesian::resetPlacement(), event if the Cell is unplaced, go through the instance list to remove the potential feed cells. The netlist is written back after placement *with* the feed cells added. So, when loading again, they where cluttering the netlist and giving feed redefintions warning. They also generated erratic behaviors. * New: In Anabatic::NetBuilderHV, added configurations for: * 1G_1M1_1PinM3 * 1G_xM1_1PinM2 * 2G_xM1_1PinM2 * doRp_AccessNorthPin * New: In Anabatic::NetBuilderHV, added Pin management in SingleGCell. * Bug: In Anabatic::NetBuilderHV::_do_xG_1PinM3(), the North/South global routing configuration was forgotten (only East/West). * Bug: In Katana::TrackFixedSegment, the blockage net was a static variable initialized at the first run. So we were getting stuck with the blockage net of the first cell to be routed. Of course, it did show only when Katana was run multiple times in the same process.
2020-03-03 17:50:18 -06:00
AutoContact* NetBuilderHV::doRp_AccessNorthPin ( GCell* gcell, RoutingPad* rp )
{
cdebug_log(145,1) << getTypeName() << "::doRp_AccessNorthPin() " << rp << endl;
AutoContact* rpSourceContact = NULL;
AutoContact* rpContactTarget = NULL;
AutoContact* turn = NULL;
doRp_AutoContacts( gcell, rp, rpSourceContact, rpContactTarget, NoProtect );
Net* net = rp->getNet();
Pin* pin = dynamic_cast<Pin*>( rp->getOccurrence().getEntity() );
Pin::AccessDirection pinDir = pin->getAccessDirection();
if (pinDir == Pin::AccessDirection::NORTH) {
turn = AutoContactTurn::create( gcell, net, Session::getRoutingLayer(1) );
AutoSegment* segment = AutoSegment::create( rpSourceContact, turn, Flags::Vertical );
segment->setAxis( rp->getX(), Flags::Force );
segment->setFlags( AutoSegment::SegFixed|AutoSegment::SegFixedAxis );
rpSourceContact = turn;
turn = AutoContactTurn::create( gcell, net, Session::getContactLayer(1) );
segment = AutoSegment::create( rpSourceContact, turn, Flags::Horizontal );
rpSourceContact = turn;
DbU::Unit axis = gcell->getYMax() - Session::getDHorizontalPitch();
cdebug_log(145,0) << "axis:" << DbU::getValueString(axis) << endl;
segment->setAxis( axis, Flags::Force );
//segment->setFlags( AutoSegment::SegFixed|AutoSegment::SegFixedAxis );
cdebug_log(145,0) << segment << endl;
} else {
turn = rpSourceContact;
}
cdebug_tabw(145,-1);
return turn;
}
Added new configuration in NetBuilderHV (for M3 Pins). * New: In Anabatic::NetBuilderHV(), added configurations to manage one Pin M3 + one M1 terminal plus 2 & 3 globals: * _do_2G_1M1_1PinM3() * _do_3G_1M1_1PinM3() They were occuring for the first time in soclayout/experiment7 in the "flat" approach. * New: In Katana::runNegociate(), mark the newly routed netlist as "NetlistTerminal" so it is not placed and routed *again* when reused as an instance (mostly interract with Etesian).
2020-04-20 05:53:07 -05:00
AutoContact* NetBuilderHV::doRp_AccessEastWestPin ( GCell* gcell, RoutingPad* rp )
{
cdebug_log(145,1) << getTypeName() << "::doRp_AccessNorthPin() " << rp << endl;
Net* net = rp->getNet();
Pin* pin = dynamic_cast<Pin*>( rp->getOccurrence().getEntity() );
Pin::AccessDirection pinDir = pin->getAccessDirection();
AutoContact* rpSourceContact = NULL;
AutoContact* rpContactTarget = NULL;
AutoContact* turn = NULL;
AutoSegment* segment = NULL;
doRp_AutoContacts( gcell, rp, rpSourceContact, rpContactTarget, NoProtect );
turn = AutoContactTurn::create( gcell, net, Session::getContactLayer(1) );
segment = AutoSegment::create( rpSourceContact, turn, Flags::Horizontal );
segment->setAxis( pin->getCenter().getY(), Flags::Force );
rpSourceContact = turn;
turn = AutoContactTurn::create( gcell, net, Session::getContactLayer(1) );
segment = AutoSegment::create( rpSourceContact, turn, Flags::Vertical );
DbU::Unit axis = 0;
if (pinDir == Pin::AccessDirection::EAST)
axis = gcell->getXMax() - Session::getDVerticalPitch();
else
axis = gcell->getXMin() + Session::getDVerticalPitch();
segment->setAxis( axis );
cdebug_tabw(145,-1);
return turn;
}
Complete the modularization of the detail routing converter. * New: In Anabatic::NetBuilder, set all the attributes as private and create accessors and mutators. Finish virtualising all GCell build methods and transfer them into the NetBuilderHV class. Build methods now return a boolean to tell if the GCell was processed or not, to allow cascading in the "big switch". Reorganise the "big switch" in separate sections only partially cascading. Truly fuse the *big switch* for channel routing and over-the-cells. Create a new method "_do_globalSegment()" to delagate the drawing of global segments between two GCell to the derived classes.
2017-12-18 11:15:14 -06:00
bool NetBuilderHV::_do_1G_1M1 ()
Modularize the global to detail routing converter. * New: In Anabatic, rename GCellTopology to NetBuilder, expose the class (no longer in a local namespace). NetBuilder become a base class performing the walkthrough the Net tree. Derived class are tasked to build routing for specific gauge. NetBuilderHV is created to manage all gauge with metal2 horizontal and metal3 vertical. In our terminolgy we consider that the first routing metal is metal2. Metal1 is used inside the standard cells.
2017-12-16 17:13:19 -06:00
{
Complete the modularization of the detail routing converter. * New: In Anabatic::NetBuilder, set all the attributes as private and create accessors and mutators. Finish virtualising all GCell build methods and transfer them into the NetBuilderHV class. Build methods now return a boolean to tell if the GCell was processed or not, to allow cascading in the "big switch". Reorganise the "big switch" in separate sections only partially cascading. Truly fuse the *big switch* for channel routing and over-the-cells. Create a new method "_do_globalSegment()" to delagate the drawing of global segments between two GCell to the derived classes.
2017-12-18 11:15:14 -06:00
cdebug_log(145,1) << getTypeName() << "::_do_1G_1M1() [Managed Configuration - Optimized] " << getTopology() << endl;
Modularize the global to detail routing converter. * New: In Anabatic, rename GCellTopology to NetBuilder, expose the class (no longer in a local namespace). NetBuilder become a base class performing the walkthrough the Net tree. Derived class are tasked to build routing for specific gauge. NetBuilderHV is created to manage all gauge with metal2 horizontal and metal3 vertical. In our terminolgy we consider that the first routing metal is metal2. Metal1 is used inside the standard cells.
2017-12-16 17:13:19 -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
#if THIS_IS_DISABLED
Modularize the global to detail routing converter. * New: In Anabatic, rename GCellTopology to NetBuilder, expose the class (no longer in a local namespace). NetBuilder become a base class performing the walkthrough the Net tree. Derived class are tasked to build routing for specific gauge. NetBuilderHV is created to manage all gauge with metal2 horizontal and metal3 vertical. In our terminolgy we consider that the first routing metal is metal2. Metal1 is used inside the standard cells.
2017-12-16 17:13:19 -06:00
uint64_t flags = NoFlags;
if (east() ) { flags |= HAccess; }
else if (west() ) { flags |= HAccess; }
else if (north()) { flags |= VSmall; }
else if (south()) { flags |= VSmall; }
setBothCornerContacts( doRp_Access(getGCell(),getRoutingPads()[0],flags) );
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
#endif
if (north() or south()) {
setBothCornerContacts( doRp_Access(getGCell(),getRoutingPads()[0],VSmall) );
} else {
setBothCornerContacts( doRp_Access( getGCell(), getRoutingPads()[0], UseNonPref|HAccess ) );
}
Modularize the global to detail routing converter. * New: In Anabatic, rename GCellTopology to NetBuilder, expose the class (no longer in a local namespace). NetBuilder become a base class performing the walkthrough the Net tree. Derived class are tasked to build routing for specific gauge. NetBuilderHV is created to manage all gauge with metal2 horizontal and metal3 vertical. In our terminolgy we consider that the first routing metal is metal2. Metal1 is used inside the standard cells.
2017-12-16 17:13:19 -06:00
cdebug_tabw(145,-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
Complete the modularization of the detail routing converter. * New: In Anabatic::NetBuilder, set all the attributes as private and create accessors and mutators. Finish virtualising all GCell build methods and transfer them into the NetBuilderHV class. Build methods now return a boolean to tell if the GCell was processed or not, to allow cascading in the "big switch". Reorganise the "big switch" in separate sections only partially cascading. Truly fuse the *big switch* for channel routing and over-the-cells. Create a new method "_do_globalSegment()" to delagate the drawing of global segments between two GCell to the derived classes.
2017-12-18 11:15:14 -06:00
return true;
Modularize the global to detail routing converter. * New: In Anabatic, rename GCellTopology to NetBuilder, expose the class (no longer in a local namespace). NetBuilder become a base class performing the walkthrough the Net tree. Derived class are tasked to build routing for specific gauge. NetBuilderHV is created to manage all gauge with metal2 horizontal and metal3 vertical. In our terminolgy we consider that the first routing metal is metal2. Metal1 is used inside the standard cells.
2017-12-16 17:13:19 -06:00
}
Complete the modularization of the detail routing converter. * New: In Anabatic::NetBuilder, set all the attributes as private and create accessors and mutators. Finish virtualising all GCell build methods and transfer them into the NetBuilderHV class. Build methods now return a boolean to tell if the GCell was processed or not, to allow cascading in the "big switch". Reorganise the "big switch" in separate sections only partially cascading. Truly fuse the *big switch* for channel routing and over-the-cells. Create a new method "_do_globalSegment()" to delagate the drawing of global segments between two GCell to the derived classes.
2017-12-18 11:15:14 -06:00
bool NetBuilderHV::_do_1G_xM1 ()
Modularize the global to detail routing converter. * New: In Anabatic, rename GCellTopology to NetBuilder, expose the class (no longer in a local namespace). NetBuilder become a base class performing the walkthrough the Net tree. Derived class are tasked to build routing for specific gauge. NetBuilderHV is created to manage all gauge with metal2 horizontal and metal3 vertical. In our terminolgy we consider that the first routing metal is metal2. Metal1 is used inside the standard cells.
2017-12-16 17:13:19 -06:00
{
Complete the modularization of the detail routing converter. * New: In Anabatic::NetBuilder, set all the attributes as private and create accessors and mutators. Finish virtualising all GCell build methods and transfer them into the NetBuilderHV class. Build methods now return a boolean to tell if the GCell was processed or not, to allow cascading in the "big switch". Reorganise the "big switch" in separate sections only partially cascading. Truly fuse the *big switch* for channel routing and over-the-cells. Create a new method "_do_globalSegment()" to delagate the drawing of global segments between two GCell to the derived classes.
2017-12-18 11:15:14 -06:00
cdebug_log(145,1) << getTypeName() << "::_do_1G_" << (int)getConnexity().fields.M1 << "M1() [Managed Configuration]" << endl;
Modularize the global to detail routing converter. * New: In Anabatic, rename GCellTopology to NetBuilder, expose the class (no longer in a local namespace). NetBuilder become a base class performing the walkthrough the Net tree. Derived class are tasked to build routing for specific gauge. NetBuilderHV is created to manage all gauge with metal2 horizontal and metal3 vertical. In our terminolgy we consider that the first routing metal is metal2. Metal1 is used inside the standard cells.
2017-12-16 17:13:19 -06:00
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
sortRpByX( getRoutingPads(), NoFlags ); // increasing X.
Modularize the global to detail routing converter. * New: In Anabatic, rename GCellTopology to NetBuilder, expose the class (no longer in a local namespace). NetBuilder become a base class performing the walkthrough the Net tree. Derived class are tasked to build routing for specific gauge. NetBuilderHV is created to manage all gauge with metal2 horizontal and metal3 vertical. In our terminolgy we consider that the first routing metal is metal2. Metal1 is used inside the standard cells.
2017-12-16 17:13:19 -06:00
for ( size_t i=1 ; i<getRoutingPads().size() ; ++i ) {
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
AutoContact* leftContact = doRp_Access( getGCell(), getRoutingPads()[i-1], HAccess|UseNonPref );
AutoContact* rightContact = doRp_Access( getGCell(), getRoutingPads()[i ], HAccess|UseNonPref );
Modularize the global to detail routing converter. * New: In Anabatic, rename GCellTopology to NetBuilder, expose the class (no longer in a local namespace). NetBuilder become a base class performing the walkthrough the Net tree. Derived class are tasked to build routing for specific gauge. NetBuilderHV is created to manage all gauge with metal2 horizontal and metal3 vertical. In our terminolgy we consider that the first routing metal is metal2. Metal1 is used inside the standard cells.
2017-12-16 17:13:19 -06:00
AutoSegment::create( leftContact, rightContact, Flags::Horizontal );
}
Component* globalRp = NULL;
if (east()) globalRp = getRoutingPads()[getRoutingPads().size()-1];
else if (west()) globalRp = getRoutingPads()[0];
else {
globalRp = getRoutingPads()[0];
cdebug_log(145,0) << "| Initial N/S Global RP: " << globalRp << endl;
for ( size_t i=1 ; i<getRoutingPads().size() ; ++i ) {
if (getRoutingPads()[i]->getBoundingBox().getHeight() > globalRp->getBoundingBox().getHeight()) {
cdebug_log(145,0) << "| Better RP: " << globalRp << endl;
globalRp = getRoutingPads()[i];
}
}
}
AutoContact* globalContact = doRp_Access( getGCell(), globalRp, HAccess );
if (north() or south()) {
AutoContact* turn = globalContact;
globalContact = AutoContactTurn::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( globalContact, turn, Flags::Horizontal );
}
setBothCornerContacts( globalContact );
cdebug_tabw(145,-1);
Complete the modularization of the detail routing converter. * New: In Anabatic::NetBuilder, set all the attributes as private and create accessors and mutators. Finish virtualising all GCell build methods and transfer them into the NetBuilderHV class. Build methods now return a boolean to tell if the GCell was processed or not, to allow cascading in the "big switch". Reorganise the "big switch" in separate sections only partially cascading. Truly fuse the *big switch* for channel routing and over-the-cells. Create a new method "_do_globalSegment()" to delagate the drawing of global segments between two GCell to the derived classes.
2017-12-18 11:15:14 -06:00
return true;
Modularize the global to detail routing converter. * New: In Anabatic, rename GCellTopology to NetBuilder, expose the class (no longer in a local namespace). NetBuilder become a base class performing the walkthrough the Net tree. Derived class are tasked to build routing for specific gauge. NetBuilderHV is created to manage all gauge with metal2 horizontal and metal3 vertical. In our terminolgy we consider that the first routing metal is metal2. Metal1 is used inside the standard cells.
2017-12-16 17:13:19 -06:00
}
Complete the modularization of the detail routing converter. * New: In Anabatic::NetBuilder, set all the attributes as private and create accessors and mutators. Finish virtualising all GCell build methods and transfer them into the NetBuilderHV class. Build methods now return a boolean to tell if the GCell was processed or not, to allow cascading in the "big switch". Reorganise the "big switch" in separate sections only partially cascading. Truly fuse the *big switch* for channel routing and over-the-cells. Create a new method "_do_globalSegment()" to delagate the drawing of global segments between two GCell to the derived classes.
2017-12-18 11:15:14 -06:00
bool NetBuilderHV::_do_xG ()
{
cdebug_log(145,1) << getTypeName() << "::_do_xG()" << endl;
const Layer* viaLayer = Session::getDContactLayer();
if (getConnexity().fields.globals == 2) {
setBothCornerContacts( AutoContactTurn::create( getGCell(), getNet(), viaLayer ) );
} else if (getConnexity().fields.globals == 3) {
if (east() and west()) {
setSouthWestContact( AutoContactTurn::create( getGCell(), getNet(), viaLayer ) );
setNorthEastContact( AutoContactVTee::create( getGCell(), getNet(), viaLayer ) );
if (south()) swapCornerContacts();
AutoSegment::create( getSouthWestContact(), getNorthEastContact(), Flags::Vertical );
} else {
setSouthWestContact( AutoContactTurn::create( getGCell(), getNet(), viaLayer ) );
setNorthEastContact( AutoContactHTee::create( getGCell(), getNet(), viaLayer ) );
if (west()) swapCornerContacts();
AutoSegment::create( getSouthWestContact(), getNorthEastContact(), Flags::Horizontal );
}
} else { // fields.globals == 4.
AutoContact* turn = AutoContactTurn::create( getGCell(), getNet(), viaLayer );
setSouthWestContact( AutoContactHTee::create( getGCell(), getNet(), viaLayer ) );
setNorthEastContact( AutoContactVTee::create( getGCell(), getNet(), viaLayer ) );
AutoSegment::create( getSouthWestContact(), turn, Flags::Horizontal );
AutoSegment::create( turn, getNorthEastContact(), Flags::Vertical );
}
cdebug_tabw(145,-1);
return true;
}
bool NetBuilderHV::_do_2G ()
{
cdebug_log(145,1) << getTypeName() << "::_do_2G()" << endl;
const Layer* viaLayer = Session::getDContactLayer();
if (east() and west()) {
setSouthWestContact( AutoContactTurn::create( getGCell(), getNet(), viaLayer ) );
setNorthEastContact( AutoContactTurn::create( getGCell(), getNet(), viaLayer ) );
AutoSegment::create( getSouthWestContact(), getNorthEastContact(), Flags::Vertical );
} else if (south() and north()) {
setSouthWestContact( AutoContactTurn::create( getGCell(), getNet(), viaLayer ) );
setNorthEastContact( AutoContactTurn::create( getGCell(), getNet(), viaLayer ) );
AutoSegment::create( getSouthWestContact(), getNorthEastContact(), Flags::Horizontal );
} else {
setBothCornerContacts( AutoContactTurn::create( getGCell(), getNet(), viaLayer ) );
}
cdebug_tabw(145,-1);
return true;
}
bool NetBuilderHV::_do_xG_1Pad ()
{
cdebug_log(145,1) << getTypeName() << "::_do_xG_1Pad() [Managed Configuration - Optimized] " << getTopology() << endl;
cdebug_log(145,0) << "getConnexity().globals:" << (int)getConnexity().fields.globals << endl;
uint64_t flags = NoFlags;
bool eastPad = false;
bool westPad = false;
bool northPad = false;
bool southPad = false;
Instance* padInstance = getRoutingPads()[0]->getOccurrence().getPath().getHeadInstance();
switch ( padInstance->getTransformation().getOrientation() ) {
case Transformation::Orientation::ID: northPad = true; break;
case Transformation::Orientation::MY: southPad = true; break;
case Transformation::Orientation::YR:
case Transformation::Orientation::R3: eastPad = true; flags |= HAccess; break;
case Transformation::Orientation::R1: westPad = true; flags |= HAccess; break;
default:
cerr << Warning( "Unmanaged orientation %s for pad <%s>."
, getString(padInstance->getTransformation().getOrientation()).c_str()
, getString(padInstance).c_str() ) << endl;
break;
}
cdebug_log(145,0) << "eastPad:" << eastPad << ", "
<< "westPad:" << westPad << ", "
<< "northPad:" << northPad << ", "
<< "southPad:" << southPad
<< endl;
AutoContact* source = doRp_AccessPad( getRoutingPads()[0], flags );
// Point position = getRoutingPads()[0]->getCenter();
// AutoContact* source = NULL;
// GCell* gcell = Session::getAnabatic()->getGCellGrid()->getGCell(position);
// source = AutoContactTerminal::create ( gcell
// , getRoutingPads()[0]
// , Session::getContactLayer(3)
// , position
// , Session::getViaWidth(3), Session::getViaWidth(3)
// );
// source->setFlags( CntFixed );
// if (northPad or eastPad) {
// getSouthWestContact() = getNorthEastContact() = source;
// cdebug_tabw(145,-1);
// return;
// }
// Check for straight lines, which are not managed by _do_xG().
if (getConnexity().fields.globals == 1) {
if ( (westPad and (east() != NULL))
or (eastPad and (west() != NULL)) ) {
AutoContact* turn = AutoContactTurn::create( getGCell(), getNet(), Session::getContactLayer(1) );
setBothCornerContacts( AutoContactTurn::create( getGCell(), getNet(), Session::getContactLayer(1) ) );
AutoSegment::create( source, turn, Flags::Horizontal );
AutoSegment::create( turn, getNorthEastContact(), Flags::Vertical );
cdebug_tabw(145,-1);
return true;
} else if ( (southPad and (north() != NULL))
or (northPad and (south() != NULL)) ) {
AutoContact* turn = AutoContactTurn::create( getGCell(), getNet(), Session::getContactLayer(1) );
setBothCornerContacts( AutoContactTurn::create( getGCell(), getNet(), Session::getContactLayer(1) ) );
AutoSegment::create( source, turn, Flags::Vertical );
AutoSegment::create( turn, getNorthEastContact(), Flags::Horizontal );
cdebug_tabw(145,-1);
return true;
}
}
++getConnexity().fields.globals;
--getConnexity().fields.Pad;
if (westPad ) addToWests ( source->getBodyHook() );
if (eastPad ) addToEasts ( source->getBodyHook() );
if (southPad) addToSouths( source->getBodyHook() );
if (northPad) addToNorths( source->getBodyHook() );
_do_xG();
if (westPad) {
AutoSegment::create( source, getSouthWestContact(), Flags::Horizontal );
clearWests();
}
if (eastPad) {
AutoSegment::create( source, getNorthEastContact(), Flags::Horizontal );
clearEasts();
}
if (southPad) {
AutoSegment::create( source, getSouthWestContact(), Flags::Vertical );
clearSouths();
}
if (northPad) {
AutoSegment::create( source, getNorthEastContact(), Flags::Vertical );
clearNorths();
}
--(getConnexity().fields.globals);
cdebug_tabw(145,-1);
return true;
}
bool NetBuilderHV::_do_1G_1PinM2 ()
{
cdebug_log(145,1) << getTypeName() << "::_do_1G_1PinM2() [Managed Configuration - Optimized] " << getTopology() << endl;
Various bug corrections to pass the alliance-check-toolkit reference benchs. * New: In Isobar::PyCell, export the isRouted() and setRouted() to the Python interface. * Bug: In CRL::Entity::parseEntity(), check that the closing parenthesis is the last character of the net name. Issue a more relevant error message. * In Anabatic::NetBuilder::_do_xG() and all other unimplemented methods, throw an error if called from a derived classes instead of just issuing a message in the debug stream. Avoid later core dumps... * In Anabatic::NetBuilderHV, implement the builders for GCells with one pin. Needed to support chip/corona routing. * Bug: In Cumulus/plugins/Chip.py, check that coronaCk exists before using it. * New: In Cumulus/plugins/PadsCorona/Side._placePad(), when routing a design with symbolic pads, export the chip external "pad" connectors to be able to perform a lvx (otherwise cougar do not create external nets). In Corona._createCoreWire(), set the minimal gap between the pads and the corona to 6 pitches. Empirical value to avoid DRC errors with symbolic pads (pxlib). When successufully done, mark the Corona cell as routed. * Bug: In Cumulus/plugins/Core2Chip.IoNet, the regex for vectorizet net was wrong, it was allowing only one digit in the index. * Bug: In Cumulus/plugins/Core2Chip.cmos, correct management of pad & corona clock nets. Correct connexion between vdde/vddi. * Bug: In Unicorn/cgt.py, forgot to execute scripts when in text mode.
2019-05-24 16:57:22 -05:00
AutoContact* rpSourceContact = NULL;
Redesign of I/O pads and Corona support to mix real and symbolic. * Bug: In CRL/etc/symbolic/cmos/plugins.conf, rails dimensions are no longer expressed directly in lambda. Must be created using helper.l(). * Change: In CRL::ApParser, slightly more smart management of Pin width. Must normalize Pin behavior between Alliance & Hurricane as in Alliance they have only one dimension. * Change: In CRL::LefImport, if a net name end with "!", assume it's a global one. Have to check this naming convention. * Change: In Anabatic::NetBuilderHV::_do_1G_1PinM3() & _do_1G_1PinM2() now implemented. Needed for the corona routing support. * Change: In AnabaticEngine::setupPreRouted(), exclude segments outside the abutment box. * Change: In KatanaEngine::PowerRails, remove the I/O pad support as now we route only inside the Corona. So only one vdd/vss/ck are supported. * New: In cumulus/plugins/ChipPlugin.py, complete rewrite of the chip support: * Uncouple pad I/O ring whith real cells (foundry) from a symbolic core. A new intermediate level "corona" is introduced to handle the real/symbolic transition. * Ability to explicitly setup position of the pads on the chip side in case of uneven distribution. * Enable clock tree to be build with 3 metal only (M2 to M4) instead of (M2 to M5).
2019-05-10 04:58:05 -05:00
AutoContact* rpContactTarget = NULL;
Various bug corrections to pass the alliance-check-toolkit reference benchs. * New: In Isobar::PyCell, export the isRouted() and setRouted() to the Python interface. * Bug: In CRL::Entity::parseEntity(), check that the closing parenthesis is the last character of the net name. Issue a more relevant error message. * In Anabatic::NetBuilder::_do_xG() and all other unimplemented methods, throw an error if called from a derived classes instead of just issuing a message in the debug stream. Avoid later core dumps... * In Anabatic::NetBuilderHV, implement the builders for GCells with one pin. Needed to support chip/corona routing. * Bug: In Cumulus/plugins/Chip.py, check that coronaCk exists before using it. * New: In Cumulus/plugins/PadsCorona/Side._placePad(), when routing a design with symbolic pads, export the chip external "pad" connectors to be able to perform a lvx (otherwise cougar do not create external nets). In Corona._createCoreWire(), set the minimal gap between the pads and the corona to 6 pitches. Empirical value to avoid DRC errors with symbolic pads (pxlib). When successufully done, mark the Corona cell as routed. * Bug: In Cumulus/plugins/Core2Chip.IoNet, the regex for vectorizet net was wrong, it was allowing only one digit in the index. * Bug: In Cumulus/plugins/Core2Chip.cmos, correct management of pad & corona clock nets. Correct connexion between vdde/vddi. * Bug: In Unicorn/cgt.py, forgot to execute scripts when in text mode.
2019-05-24 16:57:22 -05:00
doRp_AutoContacts( getGCell(), getRoutingPads()[0], rpSourceContact, rpContactTarget, NoFlags );
Redesign of I/O pads and Corona support to mix real and symbolic. * Bug: In CRL/etc/symbolic/cmos/plugins.conf, rails dimensions are no longer expressed directly in lambda. Must be created using helper.l(). * Change: In CRL::ApParser, slightly more smart management of Pin width. Must normalize Pin behavior between Alliance & Hurricane as in Alliance they have only one dimension. * Change: In CRL::LefImport, if a net name end with "!", assume it's a global one. Have to check this naming convention. * Change: In Anabatic::NetBuilderHV::_do_1G_1PinM3() & _do_1G_1PinM2() now implemented. Needed for the corona routing support. * Change: In AnabaticEngine::setupPreRouted(), exclude segments outside the abutment box. * Change: In KatanaEngine::PowerRails, remove the I/O pad support as now we route only inside the Corona. So only one vdd/vss/ck are supported. * New: In cumulus/plugins/ChipPlugin.py, complete rewrite of the chip support: * Uncouple pad I/O ring whith real cells (foundry) from a symbolic core. A new intermediate level "corona" is introduced to handle the real/symbolic transition. * Ability to explicitly setup position of the pads on the chip side in case of uneven distribution. * Enable clock tree to be build with 3 metal only (M2 to M4) instead of (M2 to M5).
2019-05-10 04:58:05 -05:00
Bug fix: Bad topology in NetBuilderHV::_do_2G_xM1_1PinM2(). * Bug: In Anabatic::NetBuilderHV::_do_2G_xM1_1PinM2(), in the case of a bend of global wires, the topology was faulty. Badly connected HTees.
2020-08-02 11:11:39 -05:00
AutoContact* turn1 = AutoContactTurn::create( getGCell(), getNet(), Session::getContactLayer(1) );
Various bug corrections to pass the alliance-check-toolkit reference benchs. * New: In Isobar::PyCell, export the isRouted() and setRouted() to the Python interface. * Bug: In CRL::Entity::parseEntity(), check that the closing parenthesis is the last character of the net name. Issue a more relevant error message. * In Anabatic::NetBuilder::_do_xG() and all other unimplemented methods, throw an error if called from a derived classes instead of just issuing a message in the debug stream. Avoid later core dumps... * In Anabatic::NetBuilderHV, implement the builders for GCells with one pin. Needed to support chip/corona routing. * Bug: In Cumulus/plugins/Chip.py, check that coronaCk exists before using it. * New: In Cumulus/plugins/PadsCorona/Side._placePad(), when routing a design with symbolic pads, export the chip external "pad" connectors to be able to perform a lvx (otherwise cougar do not create external nets). In Corona._createCoreWire(), set the minimal gap between the pads and the corona to 6 pitches. Empirical value to avoid DRC errors with symbolic pads (pxlib). When successufully done, mark the Corona cell as routed. * Bug: In Cumulus/plugins/Core2Chip.IoNet, the regex for vectorizet net was wrong, it was allowing only one digit in the index. * Bug: In Cumulus/plugins/Core2Chip.cmos, correct management of pad & corona clock nets. Correct connexion between vdde/vddi. * Bug: In Unicorn/cgt.py, forgot to execute scripts when in text mode.
2019-05-24 16:57:22 -05:00
AutoSegment::create( rpSourceContact, turn1, Flags::Horizontal );
Redesign of I/O pads and Corona support to mix real and symbolic. * Bug: In CRL/etc/symbolic/cmos/plugins.conf, rails dimensions are no longer expressed directly in lambda. Must be created using helper.l(). * Change: In CRL::ApParser, slightly more smart management of Pin width. Must normalize Pin behavior between Alliance & Hurricane as in Alliance they have only one dimension. * Change: In CRL::LefImport, if a net name end with "!", assume it's a global one. Have to check this naming convention. * Change: In Anabatic::NetBuilderHV::_do_1G_1PinM3() & _do_1G_1PinM2() now implemented. Needed for the corona routing support. * Change: In AnabaticEngine::setupPreRouted(), exclude segments outside the abutment box. * Change: In KatanaEngine::PowerRails, remove the I/O pad support as now we route only inside the Corona. So only one vdd/vss/ck are supported. * New: In cumulus/plugins/ChipPlugin.py, complete rewrite of the chip support: * Uncouple pad I/O ring whith real cells (foundry) from a symbolic core. A new intermediate level "corona" is introduced to handle the real/symbolic transition. * Ability to explicitly setup position of the pads on the chip side in case of uneven distribution. * Enable clock tree to be build with 3 metal only (M2 to M4) instead of (M2 to M5).
2019-05-10 04:58:05 -05:00
if (east() or west()) {
AutoContact* turn2 = AutoContactTurn::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( turn1, turn2, Flags::Vertical );
turn1 = turn2;
}
setBothCornerContacts( turn1 );
cdebug_tabw(145,-1);
return true;
}
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
bool NetBuilderHV::_do_1G_1PinM1 ()
{
cdebug_log(145,1) << getTypeName() << "::_do_1G_1PinM1() [Managed Configuration - Optimized] " << getTopology() << endl;
AutoContact* rpSourceContact = NULL;
AutoContact* rpContactTarget = NULL;
AutoContact* turn = NULL;
doRp_AutoContacts( getGCell(), getRoutingPads()[0], rpSourceContact, rpContactTarget, NoProtect );
Pin* pin = dynamic_cast<Pin*>( getRoutingPads()[0]->getOccurrence().getEntity() );
Pin::AccessDirection pinDir = pin->getAccessDirection();
if ( (pinDir == Pin::AccessDirection::NORTH)
or (pinDir == Pin::AccessDirection::SOUTH) ) {
turn = AutoContactTurn::create( getGCell(), getNet(), Session::getRoutingLayer(1) );
AutoSegment::create( rpSourceContact, turn, Flags::Vertical|Flags::UseNonPref );
rpSourceContact = turn;
turn = AutoContactTurn::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment* horizontal = AutoSegment::create( rpSourceContact, turn, Flags::Horizontal );
rpSourceContact = turn;
DbU::Unit axis = getGCell()->getYMax() - Session::getDHorizontalPitch();
if (pinDir == Pin::AccessDirection::SOUTH)
axis = getGCell()->getYMin() + Session::getDHorizontalPitch();
cdebug_log(145,0) << "axis:" << DbU::getValueString(axis) << endl;
horizontal->setAxis( axis, Flags::Force );
horizontal->setFlags( AutoSegment::SegFixed|AutoSegment::SegFixedAxis );
cdebug_log(145,0) << horizontal << endl;
turn = AutoContactTurn::create( getGCell(), getNet(), Session::getRoutingLayer(1) );
AutoSegment* vertical = AutoSegment::create( rpSourceContact, turn, Flags::Vertical );
rpSourceContact = turn;
vertical->setAxis( pin->getX(), Flags::Force );
vertical->setFlags( AutoSegment::SegFixed|AutoSegment::SegFixedAxis );
cdebug_log(145,0) << vertical << endl;
turn = AutoContactTurn::create( getGCell(), getNet(), Session::getContactLayer(1) );
horizontal = AutoSegment::create( rpSourceContact, turn, Flags::Horizontal );
horizontal->setAxis( axis, Flags::Force );
rpSourceContact = turn;
} else {
turn = AutoContactTurn::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( rpSourceContact, turn, Flags::Horizontal );
rpSourceContact = turn;
}
if (east() or west()) {
rpSourceContact = AutoContactTurn::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment* vertical = AutoSegment::create( turn, rpSourceContact, Flags::Vertical );
DbU::Unit axis = getGCell()->getXMax() - Session::getDVerticalPitch();
if (pinDir == Pin::AccessDirection::WEST)
axis = getGCell()->getXMin() + Session::getDVerticalPitch();
cdebug_log(145,0) << "axis:" << DbU::getValueString(axis) << endl;
vertical->setAxis( axis, Flags::Force );
vertical->setFlags( AutoSegment::SegFixed|AutoSegment::SegFixedAxis );
turn = rpSourceContact;
}
setBothCornerContacts( turn );
cdebug_tabw(145,-1);
return true;
}
bool NetBuilderHV::_do_2G_1PinM1 ()
{
cdebug_log(145,1) << getTypeName() << "::_do_2G_1PinM1() [Managed Configuration - Optimized] " << getTopology() << endl;
AutoContact* rpSourceContact = NULL;
AutoContact* rpContactTarget = NULL;
AutoContact* tee = NULL;
AutoContact* turn = NULL;
doRp_AutoContacts( getGCell(), getRoutingPads()[0], rpSourceContact, rpContactTarget, NoProtect );
Pin* pin = dynamic_cast<Pin*>( getRoutingPads()[0]->getOccurrence().getEntity() );
Pin::AccessDirection pinDir = pin->getAccessDirection();
if ( (pinDir == Pin::AccessDirection::NORTH)
or (pinDir == Pin::AccessDirection::SOUTH) ) {
turn = AutoContactTurn::create( getGCell(), getNet(), Session::getRoutingLayer(1) );
AutoSegment::create( rpSourceContact, turn, Flags::Vertical|Flags::UseNonPref );
rpSourceContact = turn;
turn = AutoContactTurn::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment* horizontal = AutoSegment::create( rpSourceContact, turn, Flags::Horizontal );
rpSourceContact = turn;
DbU::Unit axis = getGCell()->getYMax() - Session::getDHorizontalPitch();
if (pinDir == Pin::AccessDirection::SOUTH)
axis = getGCell()->getYMin() + Session::getDHorizontalPitch();
horizontal->setAxis( axis, Flags::Force );
horizontal->setFlags( AutoSegment::SegFixed );
turn = AutoContactTurn::create( getGCell(), getNet(), Session::getRoutingLayer(1) );
AutoSegment* vertical = AutoSegment::create( rpSourceContact, turn, Flags::Vertical );
rpSourceContact = turn;
vertical->setFlags( AutoSegment::SegFixed );
turn = AutoContactTurn::create( getGCell(), getNet(), Session::getContactLayer(1) );
horizontal = AutoSegment::create( rpSourceContact, turn, Flags::Horizontal );
horizontal->setAxis( axis, Flags::Force );
rpSourceContact = turn;
turn = NULL;
}
if (east() and west()) {
// Pin must be North or South.
tee = AutoContactHTee::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( rpSourceContact, tee, Flags::Vertical );
} else if (north() and south()) {
// Pin must be East or West.
tee = AutoContactVTee::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( rpSourceContact, tee, Flags::Horizontal );
} else {
if ( (pinDir == Pin::AccessDirection::EAST)
or (pinDir == Pin::AccessDirection::WEST) ) {
turn = AutoContactTurn::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( rpSourceContact, turn, Flags::Horizontal );
rpSourceContact = turn;
}
tee = AutoContactVTee::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment* vertical = AutoSegment::create( rpSourceContact, tee, Flags::Vertical );
if ( (pinDir == Pin::AccessDirection::EAST)
or (pinDir == Pin::AccessDirection::WEST) ) {
DbU::Unit axis = getGCell()->getXMax() - Session::getDVerticalPitch();
if (pinDir == Pin::AccessDirection::WEST)
axis = getGCell()->getXMin() + Session::getDVerticalPitch();
cdebug_log(145,0) << "axis:" << DbU::getValueString(axis) << endl;
vertical->setAxis( axis, Flags::Force );
vertical->setFlags( AutoSegment::SegFixed|AutoSegment::SegFixedAxis );
}
}
setBothCornerContacts( tee );
cdebug_tabw(145,-1);
return true;
}
Various bug corrections to pass the alliance-check-toolkit reference benchs. * New: In Isobar::PyCell, export the isRouted() and setRouted() to the Python interface. * Bug: In CRL::Entity::parseEntity(), check that the closing parenthesis is the last character of the net name. Issue a more relevant error message. * In Anabatic::NetBuilder::_do_xG() and all other unimplemented methods, throw an error if called from a derived classes instead of just issuing a message in the debug stream. Avoid later core dumps... * In Anabatic::NetBuilderHV, implement the builders for GCells with one pin. Needed to support chip/corona routing. * Bug: In Cumulus/plugins/Chip.py, check that coronaCk exists before using it. * New: In Cumulus/plugins/PadsCorona/Side._placePad(), when routing a design with symbolic pads, export the chip external "pad" connectors to be able to perform a lvx (otherwise cougar do not create external nets). In Corona._createCoreWire(), set the minimal gap between the pads and the corona to 6 pitches. Empirical value to avoid DRC errors with symbolic pads (pxlib). When successufully done, mark the Corona cell as routed. * Bug: In Cumulus/plugins/Core2Chip.IoNet, the regex for vectorizet net was wrong, it was allowing only one digit in the index. * Bug: In Cumulus/plugins/Core2Chip.cmos, correct management of pad & corona clock nets. Correct connexion between vdde/vddi. * Bug: In Unicorn/cgt.py, forgot to execute scripts when in text mode.
2019-05-24 16:57:22 -05:00
bool NetBuilderHV::_do_xG_1PinM2 ()
{
cdebug_log(145,1) << getTypeName() << "::_do_xG_1PinM2() [Managed Configuration - Optimized] " << getTopology() << endl;
AutoContact* rpSourceContact = NULL;
AutoContact* rpContactTarget = NULL;
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
AutoContact* turn = NULL;
Various bug corrections to pass the alliance-check-toolkit reference benchs. * New: In Isobar::PyCell, export the isRouted() and setRouted() to the Python interface. * Bug: In CRL::Entity::parseEntity(), check that the closing parenthesis is the last character of the net name. Issue a more relevant error message. * In Anabatic::NetBuilder::_do_xG() and all other unimplemented methods, throw an error if called from a derived classes instead of just issuing a message in the debug stream. Avoid later core dumps... * In Anabatic::NetBuilderHV, implement the builders for GCells with one pin. Needed to support chip/corona routing. * Bug: In Cumulus/plugins/Chip.py, check that coronaCk exists before using it. * New: In Cumulus/plugins/PadsCorona/Side._placePad(), when routing a design with symbolic pads, export the chip external "pad" connectors to be able to perform a lvx (otherwise cougar do not create external nets). In Corona._createCoreWire(), set the minimal gap between the pads and the corona to 6 pitches. Empirical value to avoid DRC errors with symbolic pads (pxlib). When successufully done, mark the Corona cell as routed. * Bug: In Cumulus/plugins/Core2Chip.IoNet, the regex for vectorizet net was wrong, it was allowing only one digit in the index. * Bug: In Cumulus/plugins/Core2Chip.cmos, correct management of pad & corona clock nets. Correct connexion between vdde/vddi. * Bug: In Unicorn/cgt.py, forgot to execute scripts when in text mode.
2019-05-24 16:57:22 -05:00
doRp_AutoContacts( getGCell(), getRoutingPads()[0], rpSourceContact, rpContactTarget, NoFlags );
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
Pin* pin = dynamic_cast<Pin*>( getRoutingPads()[0]->getOccurrence().getEntity() );
Pin::AccessDirection pinDir = pin->getAccessDirection();
if ( (pinDir == Pin::AccessDirection::EAST)
or (pinDir == Pin::AccessDirection::WEST) ) {
turn = AutoContactTurn::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( rpSourceContact, turn, Flags::Horizontal );
rpSourceContact = turn;
turn = AutoContactTurn::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment* vertical = AutoSegment::create( rpSourceContact, turn, Flags::Vertical );
rpSourceContact = turn;
Bad computation of the track axis for fixed in NetBuilderHV::_do_xG_1PinM2(). * Bug: In NetBuilderHV::_do_xG_1PinM2(), for METAL2 pins on the EAST side, we were computing the track from the XMax position of the GCell. Now, with complete discoupling of the routing pitches and GCell size, this is most of the time false. Replace with a true computation of the last track of routing area.
2021-04-18 13:36:44 -05:00
Box cellAb = getAnabatic()->getCell()->getAbutmentBox();
RoutingLayerGauge* lgM3 = Session::getLayerGauge( 2 );
DbU::Unit axis = lgM3->getTrackPosition( cellAb.getXMin()
, cellAb.getXMax()
, getGCell()->getXMax() - Session::getDVerticalPitch()
, Constant::Superior );
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 (pinDir == Pin::AccessDirection::WEST)
Bad computation of the track axis for fixed in NetBuilderHV::_do_xG_1PinM2(). * Bug: In NetBuilderHV::_do_xG_1PinM2(), for METAL2 pins on the EAST side, we were computing the track from the XMax position of the GCell. Now, with complete discoupling of the routing pitches and GCell size, this is most of the time false. Replace with a true computation of the last track of routing area.
2021-04-18 13:36:44 -05:00
axis = lgM3->getTrackPosition( cellAb.getXMin()
, cellAb.getXMax()
, getGCell()->getXMin() + Session::getDVerticalPitch()
, Constant::Inferior );
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
cdebug_log(145,0) << "axis:" << DbU::getValueString(axis) << endl;
vertical->setAxis( axis, Flags::Force );
vertical->setFlags( AutoSegment::SegFixed|AutoSegment::SegFixedAxis );
}
Various bug corrections to pass the alliance-check-toolkit reference benchs. * New: In Isobar::PyCell, export the isRouted() and setRouted() to the Python interface. * Bug: In CRL::Entity::parseEntity(), check that the closing parenthesis is the last character of the net name. Issue a more relevant error message. * In Anabatic::NetBuilder::_do_xG() and all other unimplemented methods, throw an error if called from a derived classes instead of just issuing a message in the debug stream. Avoid later core dumps... * In Anabatic::NetBuilderHV, implement the builders for GCells with one pin. Needed to support chip/corona routing. * Bug: In Cumulus/plugins/Chip.py, check that coronaCk exists before using it. * New: In Cumulus/plugins/PadsCorona/Side._placePad(), when routing a design with symbolic pads, export the chip external "pad" connectors to be able to perform a lvx (otherwise cougar do not create external nets). In Corona._createCoreWire(), set the minimal gap between the pads and the corona to 6 pitches. Empirical value to avoid DRC errors with symbolic pads (pxlib). When successufully done, mark the Corona cell as routed. * Bug: In Cumulus/plugins/Core2Chip.IoNet, the regex for vectorizet net was wrong, it was allowing only one digit in the index. * Bug: In Cumulus/plugins/Core2Chip.cmos, correct management of pad & corona clock nets. Correct connexion between vdde/vddi. * Bug: In Unicorn/cgt.py, forgot to execute scripts when in text mode.
2019-05-24 16:57:22 -05:00
if (getConnexity().fields.globals == 2) {
if (west() and south()) {
AutoContact* turn1 = AutoContactTurn::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( rpSourceContact, turn1, Flags::Horizontal );
AutoContact* vtee1 = AutoContactVTee::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( turn1, vtee1, Flags::Vertical );
setBothCornerContacts( vtee1 );
} else if (west() and north()) {
AutoContact* vtee1 = AutoContactVTee::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( rpSourceContact, vtee1, Flags::Horizontal );
AutoContact* turn1 = AutoContactTurn::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( vtee1, turn1, Flags::Vertical );
setSouthWestContact( turn1 );
setNorthEastContact( vtee1 );
} else if (south() and north()) {
AutoContact* vtee1 = AutoContactVTee::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( rpSourceContact, vtee1, Flags::Horizontal );
setBothCornerContacts( vtee1 );
} else if (east() and north()) {
AutoContact* htee1 = AutoContactHTee::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( rpSourceContact, htee1, Flags::Horizontal );
setBothCornerContacts( htee1 );
} else if (east() and south()) {
AutoContact* htee1 = AutoContactHTee::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( rpSourceContact, htee1, Flags::Horizontal );
setBothCornerContacts( htee1 );
} else if (east() and west()) {
AutoContact* turn1 = AutoContactTurn::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( rpSourceContact, turn1, Flags::Horizontal );
AutoContact* htee1 = AutoContactHTee::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( turn1, htee1, Flags::Vertical );
setBothCornerContacts( htee1 );
}
} else {
AutoContact* htee1 = AutoContactHTee::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( rpSourceContact, htee1, Flags::Horizontal );
AutoContact* htee2 = AutoContactHTee::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( htee1, htee2, Flags::Horizontal );
Fix topological bug in NetBuilderHV::_do_xG_1PinM2(). * Bug: In NetBuilderHV::_do_xG_1PinM2(), when the M2 pin is on the *east* side and there are three globals, we must swap htee1 & htee2. Otherwise, we the SW/NE contacts do not have the right numbers of free segments attach. * New: In AnabaticEngine::cleanupGlobal(), check that no global component is left behind, sure sign that the global to detailed building went wrong.
2021-04-17 06:03:15 -05:00
if (pinDir == Pin::AccessDirection::EAST)
std::swap( htee1, htee2 );
Various bug corrections to pass the alliance-check-toolkit reference benchs. * New: In Isobar::PyCell, export the isRouted() and setRouted() to the Python interface. * Bug: In CRL::Entity::parseEntity(), check that the closing parenthesis is the last character of the net name. Issue a more relevant error message. * In Anabatic::NetBuilder::_do_xG() and all other unimplemented methods, throw an error if called from a derived classes instead of just issuing a message in the debug stream. Avoid later core dumps... * In Anabatic::NetBuilderHV, implement the builders for GCells with one pin. Needed to support chip/corona routing. * Bug: In Cumulus/plugins/Chip.py, check that coronaCk exists before using it. * New: In Cumulus/plugins/PadsCorona/Side._placePad(), when routing a design with symbolic pads, export the chip external "pad" connectors to be able to perform a lvx (otherwise cougar do not create external nets). In Corona._createCoreWire(), set the minimal gap between the pads and the corona to 6 pitches. Empirical value to avoid DRC errors with symbolic pads (pxlib). When successufully done, mark the Corona cell as routed. * Bug: In Cumulus/plugins/Core2Chip.IoNet, the regex for vectorizet net was wrong, it was allowing only one digit in the index. * Bug: In Cumulus/plugins/Core2Chip.cmos, correct management of pad & corona clock nets. Correct connexion between vdde/vddi. * Bug: In Unicorn/cgt.py, forgot to execute scripts when in text mode.
2019-05-24 16:57:22 -05:00
setSouthWestContact( htee1 );
setNorthEastContact( htee2 );
}
cdebug_tabw(145,-1);
return true;
}
Redesign of I/O pads and Corona support to mix real and symbolic. * Bug: In CRL/etc/symbolic/cmos/plugins.conf, rails dimensions are no longer expressed directly in lambda. Must be created using helper.l(). * Change: In CRL::ApParser, slightly more smart management of Pin width. Must normalize Pin behavior between Alliance & Hurricane as in Alliance they have only one dimension. * Change: In CRL::LefImport, if a net name end with "!", assume it's a global one. Have to check this naming convention. * Change: In Anabatic::NetBuilderHV::_do_1G_1PinM3() & _do_1G_1PinM2() now implemented. Needed for the corona routing support. * Change: In AnabaticEngine::setupPreRouted(), exclude segments outside the abutment box. * Change: In KatanaEngine::PowerRails, remove the I/O pad support as now we route only inside the Corona. So only one vdd/vss/ck are supported. * New: In cumulus/plugins/ChipPlugin.py, complete rewrite of the chip support: * Uncouple pad I/O ring whith real cells (foundry) from a symbolic core. A new intermediate level "corona" is introduced to handle the real/symbolic transition. * Ability to explicitly setup position of the pads on the chip side in case of uneven distribution. * Enable clock tree to be build with 3 metal only (M2 to M4) instead of (M2 to M5).
2019-05-10 04:58:05 -05:00
bool NetBuilderHV::_do_1G_1PinM3 ()
{
cdebug_log(145,1) << getTypeName() << "::_do_1G_1PinM3() [Managed Configuration - Optimized] " << getTopology() << endl;
Various bug corrections to pass the alliance-check-toolkit reference benchs. * New: In Isobar::PyCell, export the isRouted() and setRouted() to the Python interface. * Bug: In CRL::Entity::parseEntity(), check that the closing parenthesis is the last character of the net name. Issue a more relevant error message. * In Anabatic::NetBuilder::_do_xG() and all other unimplemented methods, throw an error if called from a derived classes instead of just issuing a message in the debug stream. Avoid later core dumps... * In Anabatic::NetBuilderHV, implement the builders for GCells with one pin. Needed to support chip/corona routing. * Bug: In Cumulus/plugins/Chip.py, check that coronaCk exists before using it. * New: In Cumulus/plugins/PadsCorona/Side._placePad(), when routing a design with symbolic pads, export the chip external "pad" connectors to be able to perform a lvx (otherwise cougar do not create external nets). In Corona._createCoreWire(), set the minimal gap between the pads and the corona to 6 pitches. Empirical value to avoid DRC errors with symbolic pads (pxlib). When successufully done, mark the Corona cell as routed. * Bug: In Cumulus/plugins/Core2Chip.IoNet, the regex for vectorizet net was wrong, it was allowing only one digit in the index. * Bug: In Cumulus/plugins/Core2Chip.cmos, correct management of pad & corona clock nets. Correct connexion between vdde/vddi. * Bug: In Unicorn/cgt.py, forgot to execute scripts when in text mode.
2019-05-24 16:57:22 -05:00
AutoContact* rpSourceContact = NULL;
Redesign of I/O pads and Corona support to mix real and symbolic. * Bug: In CRL/etc/symbolic/cmos/plugins.conf, rails dimensions are no longer expressed directly in lambda. Must be created using helper.l(). * Change: In CRL::ApParser, slightly more smart management of Pin width. Must normalize Pin behavior between Alliance & Hurricane as in Alliance they have only one dimension. * Change: In CRL::LefImport, if a net name end with "!", assume it's a global one. Have to check this naming convention. * Change: In Anabatic::NetBuilderHV::_do_1G_1PinM3() & _do_1G_1PinM2() now implemented. Needed for the corona routing support. * Change: In AnabaticEngine::setupPreRouted(), exclude segments outside the abutment box. * Change: In KatanaEngine::PowerRails, remove the I/O pad support as now we route only inside the Corona. So only one vdd/vss/ck are supported. * New: In cumulus/plugins/ChipPlugin.py, complete rewrite of the chip support: * Uncouple pad I/O ring whith real cells (foundry) from a symbolic core. A new intermediate level "corona" is introduced to handle the real/symbolic transition. * Ability to explicitly setup position of the pads on the chip side in case of uneven distribution. * Enable clock tree to be build with 3 metal only (M2 to M4) instead of (M2 to M5).
2019-05-10 04:58:05 -05:00
First working recursive place & route (Libre-SOC ALU16 benchmark). * New: In Hurricane::Cell::isLeaf(), a leaf cell was defined as one without instances. Now it can be software defined with the "FlattenLeaf" state. If it is flagged "flatten leaf" is will be considered as a leaf by the trans-hierarchical walkthrough. This flag is also set for Cells in the Alliance CATAL. * New: In Hurricane::Cell_LeafInstanceOccurrences::progress(), add a test to prune leaf cells. * Bug: In Hurricane::getComponentOccurrences(), when calling the locator constructor, the "withLeafCells" argument was forgotten and was getting the value of "doExtraction" (wich, fortunately, was the same). * Change: In Hurricane::HypreNet_LeafPlugOccurrences::progress(), prune the leaf cells instead of the terminal ones. * Change: In Hurricane::HypreNet_ComponentOccurrences::progress(), prune the leaf cells instead of the terminal ones. * New: In Hurricane::Cell, add a "useFlattenLeaf" mode to take into account or ignore the "FlattenLeaf" flag when doing trans- hierarchical walkthrough. Not sure if it shouldn't be systematic. * New: In Etesian::toColoquinte() and other, activate the "FlattenLeaf" mode for all walkthough. So we can stop at already placed sub-blocks. * Bug: In Etesian::toColoquinte(), when placed sub-block where present, the number of instances was under estimated, so the instance array was oversized and the end went unitialized. So we were getting all kind of strange behavior from Coloquinte... Now throw an exception if the number of instances differs from the computed size. * Bug: In Etesian::toColoquinte(), as for the instance array, now correctly compute the size with the ecluded nets (supplies, clocks). * Bug: In Etesian::resetPlacement(), event if the Cell is unplaced, go through the instance list to remove the potential feed cells. The netlist is written back after placement *with* the feed cells added. So, when loading again, they where cluttering the netlist and giving feed redefintions warning. They also generated erratic behaviors. * New: In Anabatic::NetBuilderHV, added configurations for: * 1G_1M1_1PinM3 * 1G_xM1_1PinM2 * 2G_xM1_1PinM2 * doRp_AccessNorthPin * New: In Anabatic::NetBuilderHV, added Pin management in SingleGCell. * Bug: In Anabatic::NetBuilderHV::_do_xG_1PinM3(), the North/South global routing configuration was forgotten (only East/West). * Bug: In Katana::TrackFixedSegment, the blockage net was a static variable initialized at the first run. So we were getting stuck with the blockage net of the first cell to be routed. Of course, it did show only when Katana was run multiple times in the same process.
2020-03-03 17:50:18 -06:00
rpSourceContact = doRp_AccessNorthPin( getGCell(), getRoutingPads()[0] );
Redesign of I/O pads and Corona support to mix real and symbolic. * Bug: In CRL/etc/symbolic/cmos/plugins.conf, rails dimensions are no longer expressed directly in lambda. Must be created using helper.l(). * Change: In CRL::ApParser, slightly more smart management of Pin width. Must normalize Pin behavior between Alliance & Hurricane as in Alliance they have only one dimension. * Change: In CRL::LefImport, if a net name end with "!", assume it's a global one. Have to check this naming convention. * Change: In Anabatic::NetBuilderHV::_do_1G_1PinM3() & _do_1G_1PinM2() now implemented. Needed for the corona routing support. * Change: In AnabaticEngine::setupPreRouted(), exclude segments outside the abutment box. * Change: In KatanaEngine::PowerRails, remove the I/O pad support as now we route only inside the Corona. So only one vdd/vss/ck are supported. * New: In cumulus/plugins/ChipPlugin.py, complete rewrite of the chip support: * Uncouple pad I/O ring whith real cells (foundry) from a symbolic core. A new intermediate level "corona" is introduced to handle the real/symbolic transition. * Ability to explicitly setup position of the pads on the chip side in case of uneven distribution. * Enable clock tree to be build with 3 metal only (M2 to M4) instead of (M2 to M5).
2019-05-10 04:58:05 -05:00
AutoContact* turn1 = AutoContactTurn::create( getGCell(), getNet(), Session::getContactLayer(1) );
Various bug corrections to pass the alliance-check-toolkit reference benchs. * New: In Isobar::PyCell, export the isRouted() and setRouted() to the Python interface. * Bug: In CRL::Entity::parseEntity(), check that the closing parenthesis is the last character of the net name. Issue a more relevant error message. * In Anabatic::NetBuilder::_do_xG() and all other unimplemented methods, throw an error if called from a derived classes instead of just issuing a message in the debug stream. Avoid later core dumps... * In Anabatic::NetBuilderHV, implement the builders for GCells with one pin. Needed to support chip/corona routing. * Bug: In Cumulus/plugins/Chip.py, check that coronaCk exists before using it. * New: In Cumulus/plugins/PadsCorona/Side._placePad(), when routing a design with symbolic pads, export the chip external "pad" connectors to be able to perform a lvx (otherwise cougar do not create external nets). In Corona._createCoreWire(), set the minimal gap between the pads and the corona to 6 pitches. Empirical value to avoid DRC errors with symbolic pads (pxlib). When successufully done, mark the Corona cell as routed. * Bug: In Cumulus/plugins/Core2Chip.IoNet, the regex for vectorizet net was wrong, it was allowing only one digit in the index. * Bug: In Cumulus/plugins/Core2Chip.cmos, correct management of pad & corona clock nets. Correct connexion between vdde/vddi. * Bug: In Unicorn/cgt.py, forgot to execute scripts when in text mode.
2019-05-24 16:57:22 -05:00
AutoSegment::create( rpSourceContact, turn1, Flags::Vertical );
Complete the modularization of the detail routing converter. * New: In Anabatic::NetBuilder, set all the attributes as private and create accessors and mutators. Finish virtualising all GCell build methods and transfer them into the NetBuilderHV class. Build methods now return a boolean to tell if the GCell was processed or not, to allow cascading in the "big switch". Reorganise the "big switch" in separate sections only partially cascading. Truly fuse the *big switch* for channel routing and over-the-cells. Create a new method "_do_globalSegment()" to delagate the drawing of global segments between two GCell to the derived classes.
2017-12-18 11:15:14 -06:00
if (north() or south()) {
AutoContact* turn2 = AutoContactTurn::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( turn1, turn2, Flags::Horizontal );
turn1 = turn2;
}
setBothCornerContacts( turn1 );
cdebug_tabw(145,-1);
return true;
}
Various bug corrections to pass the alliance-check-toolkit reference benchs. * New: In Isobar::PyCell, export the isRouted() and setRouted() to the Python interface. * Bug: In CRL::Entity::parseEntity(), check that the closing parenthesis is the last character of the net name. Issue a more relevant error message. * In Anabatic::NetBuilder::_do_xG() and all other unimplemented methods, throw an error if called from a derived classes instead of just issuing a message in the debug stream. Avoid later core dumps... * In Anabatic::NetBuilderHV, implement the builders for GCells with one pin. Needed to support chip/corona routing. * Bug: In Cumulus/plugins/Chip.py, check that coronaCk exists before using it. * New: In Cumulus/plugins/PadsCorona/Side._placePad(), when routing a design with symbolic pads, export the chip external "pad" connectors to be able to perform a lvx (otherwise cougar do not create external nets). In Corona._createCoreWire(), set the minimal gap between the pads and the corona to 6 pitches. Empirical value to avoid DRC errors with symbolic pads (pxlib). When successufully done, mark the Corona cell as routed. * Bug: In Cumulus/plugins/Core2Chip.IoNet, the regex for vectorizet net was wrong, it was allowing only one digit in the index. * Bug: In Cumulus/plugins/Core2Chip.cmos, correct management of pad & corona clock nets. Correct connexion between vdde/vddi. * Bug: In Unicorn/cgt.py, forgot to execute scripts when in text mode.
2019-05-24 16:57:22 -05:00
bool NetBuilderHV::_do_xG_1PinM3 ()
{
cdebug_log(145,1) << getTypeName() << "::_do_xG_1PinM3() [Managed Configuration - Optimized] " << getTopology() << endl;
AutoContact* rpSourceContact = NULL;
AutoContact* rpContactTarget = NULL;
doRp_AutoContacts( getGCell(), getRoutingPads()[0], rpSourceContact, rpContactTarget, NoFlags );
if (getConnexity().fields.globals == 2) {
if (west() and south()) {
AutoContact* turn1 = AutoContactTurn::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( rpSourceContact, turn1, Flags::Vertical );
AutoContact* htee1 = AutoContactHTee::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( turn1, htee1, Flags::Horizontal );
setBothCornerContacts( htee1 );
} else if (west() and north()) {
AutoContact* htee1 = AutoContactHTee::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( rpSourceContact, htee1, Flags::Vertical );
AutoContact* turn1 = AutoContactTurn::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( turn1, htee1, Flags::Horizontal );
setSouthWestContact( htee1 );
setNorthEastContact( turn1 );
} else if (east() and north()) {
AutoContact* vtee1 = AutoContactVTee::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( rpSourceContact, vtee1, Flags::Vertical );
setBothCornerContacts( vtee1 );
} else if (east() and south()) {
AutoContact* htee1 = AutoContactHTee::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( rpSourceContact, htee1, Flags::Vertical );
AutoContact* turn1 = AutoContactTurn::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( htee1, turn1, Flags::Horizontal );
setSouthWestContact( turn1 );
setNorthEastContact( htee1 );
First working recursive place & route (Libre-SOC ALU16 benchmark). * New: In Hurricane::Cell::isLeaf(), a leaf cell was defined as one without instances. Now it can be software defined with the "FlattenLeaf" state. If it is flagged "flatten leaf" is will be considered as a leaf by the trans-hierarchical walkthrough. This flag is also set for Cells in the Alliance CATAL. * New: In Hurricane::Cell_LeafInstanceOccurrences::progress(), add a test to prune leaf cells. * Bug: In Hurricane::getComponentOccurrences(), when calling the locator constructor, the "withLeafCells" argument was forgotten and was getting the value of "doExtraction" (wich, fortunately, was the same). * Change: In Hurricane::HypreNet_LeafPlugOccurrences::progress(), prune the leaf cells instead of the terminal ones. * Change: In Hurricane::HypreNet_ComponentOccurrences::progress(), prune the leaf cells instead of the terminal ones. * New: In Hurricane::Cell, add a "useFlattenLeaf" mode to take into account or ignore the "FlattenLeaf" flag when doing trans- hierarchical walkthrough. Not sure if it shouldn't be systematic. * New: In Etesian::toColoquinte() and other, activate the "FlattenLeaf" mode for all walkthough. So we can stop at already placed sub-blocks. * Bug: In Etesian::toColoquinte(), when placed sub-block where present, the number of instances was under estimated, so the instance array was oversized and the end went unitialized. So we were getting all kind of strange behavior from Coloquinte... Now throw an exception if the number of instances differs from the computed size. * Bug: In Etesian::toColoquinte(), as for the instance array, now correctly compute the size with the ecluded nets (supplies, clocks). * Bug: In Etesian::resetPlacement(), event if the Cell is unplaced, go through the instance list to remove the potential feed cells. The netlist is written back after placement *with* the feed cells added. So, when loading again, they where cluttering the netlist and giving feed redefintions warning. They also generated erratic behaviors. * New: In Anabatic::NetBuilderHV, added configurations for: * 1G_1M1_1PinM3 * 1G_xM1_1PinM2 * 2G_xM1_1PinM2 * doRp_AccessNorthPin * New: In Anabatic::NetBuilderHV, added Pin management in SingleGCell. * Bug: In Anabatic::NetBuilderHV::_do_xG_1PinM3(), the North/South global routing configuration was forgotten (only East/West). * Bug: In Katana::TrackFixedSegment, the blockage net was a static variable initialized at the first run. So we were getting stuck with the blockage net of the first cell to be routed. Of course, it did show only when Katana was run multiple times in the same process.
2020-03-03 17:50:18 -06:00
} else if (north() and south()) {
AutoContact* turn1 = AutoContactTurn::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( rpSourceContact, turn1, Flags::Vertical );
Various bug corrections to pass the alliance-check-toolkit reference benchs. * New: In Isobar::PyCell, export the isRouted() and setRouted() to the Python interface. * Bug: In CRL::Entity::parseEntity(), check that the closing parenthesis is the last character of the net name. Issue a more relevant error message. * In Anabatic::NetBuilder::_do_xG() and all other unimplemented methods, throw an error if called from a derived classes instead of just issuing a message in the debug stream. Avoid later core dumps... * In Anabatic::NetBuilderHV, implement the builders for GCells with one pin. Needed to support chip/corona routing. * Bug: In Cumulus/plugins/Chip.py, check that coronaCk exists before using it. * New: In Cumulus/plugins/PadsCorona/Side._placePad(), when routing a design with symbolic pads, export the chip external "pad" connectors to be able to perform a lvx (otherwise cougar do not create external nets). In Corona._createCoreWire(), set the minimal gap between the pads and the corona to 6 pitches. Empirical value to avoid DRC errors with symbolic pads (pxlib). When successufully done, mark the Corona cell as routed. * Bug: In Cumulus/plugins/Core2Chip.IoNet, the regex for vectorizet net was wrong, it was allowing only one digit in the index. * Bug: In Cumulus/plugins/Core2Chip.cmos, correct management of pad & corona clock nets. Correct connexion between vdde/vddi. * Bug: In Unicorn/cgt.py, forgot to execute scripts when in text mode.
2019-05-24 16:57:22 -05:00
AutoContact* vtee1 = AutoContactVTee::create( getGCell(), getNet(), Session::getContactLayer(1) );
First working recursive place & route (Libre-SOC ALU16 benchmark). * New: In Hurricane::Cell::isLeaf(), a leaf cell was defined as one without instances. Now it can be software defined with the "FlattenLeaf" state. If it is flagged "flatten leaf" is will be considered as a leaf by the trans-hierarchical walkthrough. This flag is also set for Cells in the Alliance CATAL. * New: In Hurricane::Cell_LeafInstanceOccurrences::progress(), add a test to prune leaf cells. * Bug: In Hurricane::getComponentOccurrences(), when calling the locator constructor, the "withLeafCells" argument was forgotten and was getting the value of "doExtraction" (wich, fortunately, was the same). * Change: In Hurricane::HypreNet_LeafPlugOccurrences::progress(), prune the leaf cells instead of the terminal ones. * Change: In Hurricane::HypreNet_ComponentOccurrences::progress(), prune the leaf cells instead of the terminal ones. * New: In Hurricane::Cell, add a "useFlattenLeaf" mode to take into account or ignore the "FlattenLeaf" flag when doing trans- hierarchical walkthrough. Not sure if it shouldn't be systematic. * New: In Etesian::toColoquinte() and other, activate the "FlattenLeaf" mode for all walkthough. So we can stop at already placed sub-blocks. * Bug: In Etesian::toColoquinte(), when placed sub-block where present, the number of instances was under estimated, so the instance array was oversized and the end went unitialized. So we were getting all kind of strange behavior from Coloquinte... Now throw an exception if the number of instances differs from the computed size. * Bug: In Etesian::toColoquinte(), as for the instance array, now correctly compute the size with the ecluded nets (supplies, clocks). * Bug: In Etesian::resetPlacement(), event if the Cell is unplaced, go through the instance list to remove the potential feed cells. The netlist is written back after placement *with* the feed cells added. So, when loading again, they where cluttering the netlist and giving feed redefintions warning. They also generated erratic behaviors. * New: In Anabatic::NetBuilderHV, added configurations for: * 1G_1M1_1PinM3 * 1G_xM1_1PinM2 * 2G_xM1_1PinM2 * doRp_AccessNorthPin * New: In Anabatic::NetBuilderHV, added Pin management in SingleGCell. * Bug: In Anabatic::NetBuilderHV::_do_xG_1PinM3(), the North/South global routing configuration was forgotten (only East/West). * Bug: In Katana::TrackFixedSegment, the blockage net was a static variable initialized at the first run. So we were getting stuck with the blockage net of the first cell to be routed. Of course, it did show only when Katana was run multiple times in the same process.
2020-03-03 17:50:18 -06:00
AutoSegment::create( turn1, vtee1, Flags::Horizontal );
Various bug corrections to pass the alliance-check-toolkit reference benchs. * New: In Isobar::PyCell, export the isRouted() and setRouted() to the Python interface. * Bug: In CRL::Entity::parseEntity(), check that the closing parenthesis is the last character of the net name. Issue a more relevant error message. * In Anabatic::NetBuilder::_do_xG() and all other unimplemented methods, throw an error if called from a derived classes instead of just issuing a message in the debug stream. Avoid later core dumps... * In Anabatic::NetBuilderHV, implement the builders for GCells with one pin. Needed to support chip/corona routing. * Bug: In Cumulus/plugins/Chip.py, check that coronaCk exists before using it. * New: In Cumulus/plugins/PadsCorona/Side._placePad(), when routing a design with symbolic pads, export the chip external "pad" connectors to be able to perform a lvx (otherwise cougar do not create external nets). In Corona._createCoreWire(), set the minimal gap between the pads and the corona to 6 pitches. Empirical value to avoid DRC errors with symbolic pads (pxlib). When successufully done, mark the Corona cell as routed. * Bug: In Cumulus/plugins/Core2Chip.IoNet, the regex for vectorizet net was wrong, it was allowing only one digit in the index. * Bug: In Cumulus/plugins/Core2Chip.cmos, correct management of pad & corona clock nets. Correct connexion between vdde/vddi. * Bug: In Unicorn/cgt.py, forgot to execute scripts when in text mode.
2019-05-24 16:57:22 -05:00
First working recursive place & route (Libre-SOC ALU16 benchmark). * New: In Hurricane::Cell::isLeaf(), a leaf cell was defined as one without instances. Now it can be software defined with the "FlattenLeaf" state. If it is flagged "flatten leaf" is will be considered as a leaf by the trans-hierarchical walkthrough. This flag is also set for Cells in the Alliance CATAL. * New: In Hurricane::Cell_LeafInstanceOccurrences::progress(), add a test to prune leaf cells. * Bug: In Hurricane::getComponentOccurrences(), when calling the locator constructor, the "withLeafCells" argument was forgotten and was getting the value of "doExtraction" (wich, fortunately, was the same). * Change: In Hurricane::HypreNet_LeafPlugOccurrences::progress(), prune the leaf cells instead of the terminal ones. * Change: In Hurricane::HypreNet_ComponentOccurrences::progress(), prune the leaf cells instead of the terminal ones. * New: In Hurricane::Cell, add a "useFlattenLeaf" mode to take into account or ignore the "FlattenLeaf" flag when doing trans- hierarchical walkthrough. Not sure if it shouldn't be systematic. * New: In Etesian::toColoquinte() and other, activate the "FlattenLeaf" mode for all walkthough. So we can stop at already placed sub-blocks. * Bug: In Etesian::toColoquinte(), when placed sub-block where present, the number of instances was under estimated, so the instance array was oversized and the end went unitialized. So we were getting all kind of strange behavior from Coloquinte... Now throw an exception if the number of instances differs from the computed size. * Bug: In Etesian::toColoquinte(), as for the instance array, now correctly compute the size with the ecluded nets (supplies, clocks). * Bug: In Etesian::resetPlacement(), event if the Cell is unplaced, go through the instance list to remove the potential feed cells. The netlist is written back after placement *with* the feed cells added. So, when loading again, they where cluttering the netlist and giving feed redefintions warning. They also generated erratic behaviors. * New: In Anabatic::NetBuilderHV, added configurations for: * 1G_1M1_1PinM3 * 1G_xM1_1PinM2 * 2G_xM1_1PinM2 * doRp_AccessNorthPin * New: In Anabatic::NetBuilderHV, added Pin management in SingleGCell. * Bug: In Anabatic::NetBuilderHV::_do_xG_1PinM3(), the North/South global routing configuration was forgotten (only East/West). * Bug: In Katana::TrackFixedSegment, the blockage net was a static variable initialized at the first run. So we were getting stuck with the blockage net of the first cell to be routed. Of course, it did show only when Katana was run multiple times in the same process.
2020-03-03 17:50:18 -06:00
setBothCornerContacts( vtee1 );
} else { // Remaining case is East & West.
AutoContact* htee1 = AutoContactHTee::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( rpSourceContact, htee1, Flags::Vertical );
Various bug corrections to pass the alliance-check-toolkit reference benchs. * New: In Isobar::PyCell, export the isRouted() and setRouted() to the Python interface. * Bug: In CRL::Entity::parseEntity(), check that the closing parenthesis is the last character of the net name. Issue a more relevant error message. * In Anabatic::NetBuilder::_do_xG() and all other unimplemented methods, throw an error if called from a derived classes instead of just issuing a message in the debug stream. Avoid later core dumps... * In Anabatic::NetBuilderHV, implement the builders for GCells with one pin. Needed to support chip/corona routing. * Bug: In Cumulus/plugins/Chip.py, check that coronaCk exists before using it. * New: In Cumulus/plugins/PadsCorona/Side._placePad(), when routing a design with symbolic pads, export the chip external "pad" connectors to be able to perform a lvx (otherwise cougar do not create external nets). In Corona._createCoreWire(), set the minimal gap between the pads and the corona to 6 pitches. Empirical value to avoid DRC errors with symbolic pads (pxlib). When successufully done, mark the Corona cell as routed. * Bug: In Cumulus/plugins/Core2Chip.IoNet, the regex for vectorizet net was wrong, it was allowing only one digit in the index. * Bug: In Cumulus/plugins/Core2Chip.cmos, correct management of pad & corona clock nets. Correct connexion between vdde/vddi. * Bug: In Unicorn/cgt.py, forgot to execute scripts when in text mode.
2019-05-24 16:57:22 -05:00
First working recursive place & route (Libre-SOC ALU16 benchmark). * New: In Hurricane::Cell::isLeaf(), a leaf cell was defined as one without instances. Now it can be software defined with the "FlattenLeaf" state. If it is flagged "flatten leaf" is will be considered as a leaf by the trans-hierarchical walkthrough. This flag is also set for Cells in the Alliance CATAL. * New: In Hurricane::Cell_LeafInstanceOccurrences::progress(), add a test to prune leaf cells. * Bug: In Hurricane::getComponentOccurrences(), when calling the locator constructor, the "withLeafCells" argument was forgotten and was getting the value of "doExtraction" (wich, fortunately, was the same). * Change: In Hurricane::HypreNet_LeafPlugOccurrences::progress(), prune the leaf cells instead of the terminal ones. * Change: In Hurricane::HypreNet_ComponentOccurrences::progress(), prune the leaf cells instead of the terminal ones. * New: In Hurricane::Cell, add a "useFlattenLeaf" mode to take into account or ignore the "FlattenLeaf" flag when doing trans- hierarchical walkthrough. Not sure if it shouldn't be systematic. * New: In Etesian::toColoquinte() and other, activate the "FlattenLeaf" mode for all walkthough. So we can stop at already placed sub-blocks. * Bug: In Etesian::toColoquinte(), when placed sub-block where present, the number of instances was under estimated, so the instance array was oversized and the end went unitialized. So we were getting all kind of strange behavior from Coloquinte... Now throw an exception if the number of instances differs from the computed size. * Bug: In Etesian::toColoquinte(), as for the instance array, now correctly compute the size with the ecluded nets (supplies, clocks). * Bug: In Etesian::resetPlacement(), event if the Cell is unplaced, go through the instance list to remove the potential feed cells. The netlist is written back after placement *with* the feed cells added. So, when loading again, they where cluttering the netlist and giving feed redefintions warning. They also generated erratic behaviors. * New: In Anabatic::NetBuilderHV, added configurations for: * 1G_1M1_1PinM3 * 1G_xM1_1PinM2 * 2G_xM1_1PinM2 * doRp_AccessNorthPin * New: In Anabatic::NetBuilderHV, added Pin management in SingleGCell. * Bug: In Anabatic::NetBuilderHV::_do_xG_1PinM3(), the North/South global routing configuration was forgotten (only East/West). * Bug: In Katana::TrackFixedSegment, the blockage net was a static variable initialized at the first run. So we were getting stuck with the blockage net of the first cell to be routed. Of course, it did show only when Katana was run multiple times in the same process.
2020-03-03 17:50:18 -06:00
setBothCornerContacts( htee1 );
Various bug corrections to pass the alliance-check-toolkit reference benchs. * New: In Isobar::PyCell, export the isRouted() and setRouted() to the Python interface. * Bug: In CRL::Entity::parseEntity(), check that the closing parenthesis is the last character of the net name. Issue a more relevant error message. * In Anabatic::NetBuilder::_do_xG() and all other unimplemented methods, throw an error if called from a derived classes instead of just issuing a message in the debug stream. Avoid later core dumps... * In Anabatic::NetBuilderHV, implement the builders for GCells with one pin. Needed to support chip/corona routing. * Bug: In Cumulus/plugins/Chip.py, check that coronaCk exists before using it. * New: In Cumulus/plugins/PadsCorona/Side._placePad(), when routing a design with symbolic pads, export the chip external "pad" connectors to be able to perform a lvx (otherwise cougar do not create external nets). In Corona._createCoreWire(), set the minimal gap between the pads and the corona to 6 pitches. Empirical value to avoid DRC errors with symbolic pads (pxlib). When successufully done, mark the Corona cell as routed. * Bug: In Cumulus/plugins/Core2Chip.IoNet, the regex for vectorizet net was wrong, it was allowing only one digit in the index. * Bug: In Cumulus/plugins/Core2Chip.cmos, correct management of pad & corona clock nets. Correct connexion between vdde/vddi. * Bug: In Unicorn/cgt.py, forgot to execute scripts when in text mode.
2019-05-24 16:57:22 -05:00
}
} else {
Bugs in NetBuilderHV for topologies containing metal3 pins. * In NetBuilder::_do_xG_1PinM3(), in the three global case, an incorrect connexion was made. * In NetBuilder::_do_2G_xM1_1PinM2(), in the global E+W case, segment between vtee1 <-> turn1 must be vertical.
2020-09-14 08:12:37 -05:00
AutoContact* turn1 = AutoContactTurn::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoContact* turn2 = AutoContactTurn::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( rpSourceContact, turn1, Flags::Vertical );
AutoSegment::create( turn1, turn2, Flags::Horizontal );
Various bug corrections to pass the alliance-check-toolkit reference benchs. * New: In Isobar::PyCell, export the isRouted() and setRouted() to the Python interface. * Bug: In CRL::Entity::parseEntity(), check that the closing parenthesis is the last character of the net name. Issue a more relevant error message. * In Anabatic::NetBuilder::_do_xG() and all other unimplemented methods, throw an error if called from a derived classes instead of just issuing a message in the debug stream. Avoid later core dumps... * In Anabatic::NetBuilderHV, implement the builders for GCells with one pin. Needed to support chip/corona routing. * Bug: In Cumulus/plugins/Chip.py, check that coronaCk exists before using it. * New: In Cumulus/plugins/PadsCorona/Side._placePad(), when routing a design with symbolic pads, export the chip external "pad" connectors to be able to perform a lvx (otherwise cougar do not create external nets). In Corona._createCoreWire(), set the minimal gap between the pads and the corona to 6 pitches. Empirical value to avoid DRC errors with symbolic pads (pxlib). When successufully done, mark the Corona cell as routed. * Bug: In Cumulus/plugins/Core2Chip.IoNet, the regex for vectorizet net was wrong, it was allowing only one digit in the index. * Bug: In Cumulus/plugins/Core2Chip.cmos, correct management of pad & corona clock nets. Correct connexion between vdde/vddi. * Bug: In Unicorn/cgt.py, forgot to execute scripts when in text mode.
2019-05-24 16:57:22 -05:00
AutoContact* vtee1 = AutoContactVTee::create( getGCell(), getNet(), Session::getContactLayer(1) );
Bugs in NetBuilderHV for topologies containing metal3 pins. * In NetBuilder::_do_xG_1PinM3(), in the three global case, an incorrect connexion was made. * In NetBuilder::_do_2G_xM1_1PinM2(), in the global E+W case, segment between vtee1 <-> turn1 must be vertical.
2020-09-14 08:12:37 -05:00
AutoSegment::create( turn2, vtee1, Flags::Vertical );
Various bug corrections to pass the alliance-check-toolkit reference benchs. * New: In Isobar::PyCell, export the isRouted() and setRouted() to the Python interface. * Bug: In CRL::Entity::parseEntity(), check that the closing parenthesis is the last character of the net name. Issue a more relevant error message. * In Anabatic::NetBuilder::_do_xG() and all other unimplemented methods, throw an error if called from a derived classes instead of just issuing a message in the debug stream. Avoid later core dumps... * In Anabatic::NetBuilderHV, implement the builders for GCells with one pin. Needed to support chip/corona routing. * Bug: In Cumulus/plugins/Chip.py, check that coronaCk exists before using it. * New: In Cumulus/plugins/PadsCorona/Side._placePad(), when routing a design with symbolic pads, export the chip external "pad" connectors to be able to perform a lvx (otherwise cougar do not create external nets). In Corona._createCoreWire(), set the minimal gap between the pads and the corona to 6 pitches. Empirical value to avoid DRC errors with symbolic pads (pxlib). When successufully done, mark the Corona cell as routed. * Bug: In Cumulus/plugins/Core2Chip.IoNet, the regex for vectorizet net was wrong, it was allowing only one digit in the index. * Bug: In Cumulus/plugins/Core2Chip.cmos, correct management of pad & corona clock nets. Correct connexion between vdde/vddi. * Bug: In Unicorn/cgt.py, forgot to execute scripts when in text mode.
2019-05-24 16:57:22 -05:00
AutoContact* vtee2 = AutoContactVTee::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( vtee1, vtee2, Flags::Vertical );
setSouthWestContact( vtee1 );
setNorthEastContact( vtee2 );
}
cdebug_tabw(145,-1);
return true;
}
Support for multiple M1 connectors in GCells with M3 pins. * Change: In Anabatic::NetBuilderHV(), in 2-3 globals plus one M3 pins, support up to 3 M1 connectors.
2020-04-20 13:53:49 -05:00
bool NetBuilderHV::_do_2G_xM1_1PinM3 ()
Added new configuration in NetBuilderHV (for M3 Pins). * New: In Anabatic::NetBuilderHV(), added configurations to manage one Pin M3 + one M1 terminal plus 2 & 3 globals: * _do_2G_1M1_1PinM3() * _do_3G_1M1_1PinM3() They were occuring for the first time in soclayout/experiment7 in the "flat" approach. * New: In Katana::runNegociate(), mark the newly routed netlist as "NetlistTerminal" so it is not placed and routed *again* when reused as an instance (mostly interract with Etesian).
2020-04-20 05:53:07 -05:00
{
Support for multiple M1 connectors in GCells with M3 pins. * Change: In Anabatic::NetBuilderHV(), in 2-3 globals plus one M3 pins, support up to 3 M1 connectors.
2020-04-20 13:53:49 -05:00
cdebug_log(145,1) << getTypeName() << "::_do_2G_xM1_1PinM3() [Managed Configuration - Optimized] " << getTopology() << endl;
Added new configuration in NetBuilderHV (for M3 Pins). * New: In Anabatic::NetBuilderHV(), added configurations to manage one Pin M3 + one M1 terminal plus 2 & 3 globals: * _do_2G_1M1_1PinM3() * _do_3G_1M1_1PinM3() They were occuring for the first time in soclayout/experiment7 in the "flat" approach. * New: In Katana::runNegociate(), mark the newly routed netlist as "NetlistTerminal" so it is not placed and routed *again* when reused as an instance (mostly interract with Etesian).
2020-04-20 05:53:07 -05:00
RoutingPad* pinM3 = NULL;
AutoContact* pinContact = NULL;
AutoContact* dummy = NULL;
Support for multiple M1 connectors in GCells with M3 pins. * Change: In Anabatic::NetBuilderHV(), in 2-3 globals plus one M3 pins, support up to 3 M1 connectors.
2020-04-20 13:53:49 -05:00
sortRpByX( getRoutingPads(), NoFlags ); // increasing X.
vector<RoutingPad*> rpsM1;
Added new configuration in NetBuilderHV (for M3 Pins). * New: In Anabatic::NetBuilderHV(), added configurations to manage one Pin M3 + one M1 terminal plus 2 & 3 globals: * _do_2G_1M1_1PinM3() * _do_3G_1M1_1PinM3() They were occuring for the first time in soclayout/experiment7 in the "flat" approach. * New: In Katana::runNegociate(), mark the newly routed netlist as "NetlistTerminal" so it is not placed and routed *again* when reused as an instance (mostly interract with Etesian).
2020-04-20 05:53:07 -05:00
for ( RoutingPad* rp : getRoutingPads() ) {
if (dynamic_cast<Pin*>(rp->getOccurrence().getEntity())) pinM3 = rp;
Support for multiple M1 connectors in GCells with M3 pins. * Change: In Anabatic::NetBuilderHV(), in 2-3 globals plus one M3 pins, support up to 3 M1 connectors.
2020-04-20 13:53:49 -05:00
else rpsM1.push_back( rp );
}
for ( size_t i=1 ; i<rpsM1.size() ; ++i ) {
AutoContact* leftContact = doRp_Access( getGCell(), getRoutingPads()[i-1], HAccess );
AutoContact* rightContact = doRp_Access( getGCell(), getRoutingPads()[i ], HAccess );
AutoSegment::create( leftContact, rightContact, Flags::Horizontal );
Added new configuration in NetBuilderHV (for M3 Pins). * New: In Anabatic::NetBuilderHV(), added configurations to manage one Pin M3 + one M1 terminal plus 2 & 3 globals: * _do_2G_1M1_1PinM3() * _do_3G_1M1_1PinM3() They were occuring for the first time in soclayout/experiment7 in the "flat" approach. * New: In Katana::runNegociate(), mark the newly routed netlist as "NetlistTerminal" so it is not placed and routed *again* when reused as an instance (mostly interract with Etesian).
2020-04-20 05:53:07 -05:00
}
doRp_AutoContacts( getGCell(), pinM3, pinContact, dummy, NoFlags );
Support for multiple M1 connectors in GCells with M3 pins. * Change: In Anabatic::NetBuilderHV(), in 2-3 globals plus one M3 pins, support up to 3 M1 connectors.
2020-04-20 13:53:49 -05:00
AutoContact* m1contact = doRp_Access( getGCell(), rpsM1.back(), HAccess );
Added new configuration in NetBuilderHV (for M3 Pins). * New: In Anabatic::NetBuilderHV(), added configurations to manage one Pin M3 + one M1 terminal plus 2 & 3 globals: * _do_2G_1M1_1PinM3() * _do_3G_1M1_1PinM3() They were occuring for the first time in soclayout/experiment7 in the "flat" approach. * New: In Katana::runNegociate(), mark the newly routed netlist as "NetlistTerminal" so it is not placed and routed *again* when reused as an instance (mostly interract with Etesian).
2020-04-20 05:53:07 -05:00
AutoContact* vtee1 = AutoContactVTee::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( m1contact , vtee1, Flags::Horizontal );
Support for multiple M1 connectors in GCells with M3 pins. * Change: In Anabatic::NetBuilderHV(), in 2-3 globals plus one M3 pins, support up to 3 M1 connectors.
2020-04-20 13:53:49 -05:00
AutoSegment::create( pinContact, vtee1, Flags::Vertical );
Added new configuration in NetBuilderHV (for M3 Pins). * New: In Anabatic::NetBuilderHV(), added configurations to manage one Pin M3 + one M1 terminal plus 2 & 3 globals: * _do_2G_1M1_1PinM3() * _do_3G_1M1_1PinM3() They were occuring for the first time in soclayout/experiment7 in the "flat" approach. * New: In Katana::runNegociate(), mark the newly routed netlist as "NetlistTerminal" so it is not placed and routed *again* when reused as an instance (mostly interract with Etesian).
2020-04-20 05:53:07 -05:00
AutoContact* vtee2 = AutoContactVTee::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( vtee1, vtee2, Flags::Vertical );
Support for multiple M1 connectors in GCells with M3 pins. * Change: In Anabatic::NetBuilderHV(), in 2-3 globals plus one M3 pins, support up to 3 M1 connectors.
2020-04-20 13:53:49 -05:00
setBothCornerContacts( vtee2 );
Added new configuration in NetBuilderHV (for M3 Pins). * New: In Anabatic::NetBuilderHV(), added configurations to manage one Pin M3 + one M1 terminal plus 2 & 3 globals: * _do_2G_1M1_1PinM3() * _do_3G_1M1_1PinM3() They were occuring for the first time in soclayout/experiment7 in the "flat" approach. * New: In Katana::runNegociate(), mark the newly routed netlist as "NetlistTerminal" so it is not placed and routed *again* when reused as an instance (mostly interract with Etesian).
2020-04-20 05:53:07 -05:00
cdebug_tabw(145,-1);
return true;
}
Support for multiple M1 connectors in GCells with M3 pins. * Change: In Anabatic::NetBuilderHV(), in 2-3 globals plus one M3 pins, support up to 3 M1 connectors.
2020-04-20 13:53:49 -05:00
bool NetBuilderHV::_do_3G_xM1_1PinM3 ()
Added new configuration in NetBuilderHV (for M3 Pins). * New: In Anabatic::NetBuilderHV(), added configurations to manage one Pin M3 + one M1 terminal plus 2 & 3 globals: * _do_2G_1M1_1PinM3() * _do_3G_1M1_1PinM3() They were occuring for the first time in soclayout/experiment7 in the "flat" approach. * New: In Katana::runNegociate(), mark the newly routed netlist as "NetlistTerminal" so it is not placed and routed *again* when reused as an instance (mostly interract with Etesian).
2020-04-20 05:53:07 -05:00
{
Support for multiple M1 connectors in GCells with M3 pins. * Change: In Anabatic::NetBuilderHV(), in 2-3 globals plus one M3 pins, support up to 3 M1 connectors.
2020-04-20 13:53:49 -05:00
cdebug_log(145,1) << getTypeName() << "::_do_3G_1M1_1PinM3() [Managed Configuration - Optimized] " << getTopology() << endl;
Added new configuration in NetBuilderHV (for M3 Pins). * New: In Anabatic::NetBuilderHV(), added configurations to manage one Pin M3 + one M1 terminal plus 2 & 3 globals: * _do_2G_1M1_1PinM3() * _do_3G_1M1_1PinM3() They were occuring for the first time in soclayout/experiment7 in the "flat" approach. * New: In Katana::runNegociate(), mark the newly routed netlist as "NetlistTerminal" so it is not placed and routed *again* when reused as an instance (mostly interract with Etesian).
2020-04-20 05:53:07 -05:00
RoutingPad* pinM3 = NULL;
AutoContact* pinContact = NULL;
AutoContact* dummy = NULL;
Support for multiple M1 connectors in GCells with M3 pins. * Change: In Anabatic::NetBuilderHV(), in 2-3 globals plus one M3 pins, support up to 3 M1 connectors.
2020-04-20 13:53:49 -05:00
sortRpByX( getRoutingPads(), NoFlags ); // increasing X.
vector<RoutingPad*> rpsM1;
Added new configuration in NetBuilderHV (for M3 Pins). * New: In Anabatic::NetBuilderHV(), added configurations to manage one Pin M3 + one M1 terminal plus 2 & 3 globals: * _do_2G_1M1_1PinM3() * _do_3G_1M1_1PinM3() They were occuring for the first time in soclayout/experiment7 in the "flat" approach. * New: In Katana::runNegociate(), mark the newly routed netlist as "NetlistTerminal" so it is not placed and routed *again* when reused as an instance (mostly interract with Etesian).
2020-04-20 05:53:07 -05:00
for ( RoutingPad* rp : getRoutingPads() ) {
if (dynamic_cast<Pin*>(rp->getOccurrence().getEntity())) pinM3 = rp;
Support for multiple M1 connectors in GCells with M3 pins. * Change: In Anabatic::NetBuilderHV(), in 2-3 globals plus one M3 pins, support up to 3 M1 connectors.
2020-04-20 13:53:49 -05:00
else rpsM1.push_back( rp );
}
for ( size_t i=1 ; i<rpsM1.size() ; ++i ) {
AutoContact* leftContact = doRp_Access( getGCell(), getRoutingPads()[i-1], HAccess );
AutoContact* rightContact = doRp_Access( getGCell(), getRoutingPads()[i ], HAccess );
AutoSegment::create( leftContact, rightContact, Flags::Horizontal );
Added new configuration in NetBuilderHV (for M3 Pins). * New: In Anabatic::NetBuilderHV(), added configurations to manage one Pin M3 + one M1 terminal plus 2 & 3 globals: * _do_2G_1M1_1PinM3() * _do_3G_1M1_1PinM3() They were occuring for the first time in soclayout/experiment7 in the "flat" approach. * New: In Katana::runNegociate(), mark the newly routed netlist as "NetlistTerminal" so it is not placed and routed *again* when reused as an instance (mostly interract with Etesian).
2020-04-20 05:53:07 -05:00
}
doRp_AutoContacts( getGCell(), pinM3, pinContact, dummy, NoFlags );
Support for multiple M1 connectors in GCells with M3 pins. * Change: In Anabatic::NetBuilderHV(), in 2-3 globals plus one M3 pins, support up to 3 M1 connectors.
2020-04-20 13:53:49 -05:00
AutoContact* m1contact = doRp_Access( getGCell(), rpsM1.back(), HAccess );
Added new configuration in NetBuilderHV (for M3 Pins). * New: In Anabatic::NetBuilderHV(), added configurations to manage one Pin M3 + one M1 terminal plus 2 & 3 globals: * _do_2G_1M1_1PinM3() * _do_3G_1M1_1PinM3() They were occuring for the first time in soclayout/experiment7 in the "flat" approach. * New: In Katana::runNegociate(), mark the newly routed netlist as "NetlistTerminal" so it is not placed and routed *again* when reused as an instance (mostly interract with Etesian).
2020-04-20 05:53:07 -05:00
AutoContact* vtee1 = AutoContactVTee::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( m1contact , vtee1, Flags::Horizontal );
AutoContact* vtee2 = AutoContactVTee::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( vtee1, vtee2, Flags::Vertical );
Support for multiple M1 connectors in GCells with M3 pins. * Change: In Anabatic::NetBuilderHV(), in 2-3 globals plus one M3 pins, support up to 3 M1 connectors.
2020-04-20 13:53:49 -05:00
AutoContact* vtee3 = AutoContactVTee::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( vtee2, vtee3, Flags::Vertical );
if (not south() or not north()) {
AutoSegment::create( pinContact, vtee1, Flags::Vertical );
}
if (not east() or not west()) {
AutoSegment::create( pinContact, vtee2, Flags::Horizontal );
}
if (not south()) {
setSouthWestContact( vtee2 );
setNorthEastContact( vtee3 );
} else if (not north()) {
setSouthWestContact( vtee3 );
setNorthEastContact( vtee2 );
} else if (not east()) {
setSouthWestContact( vtee3 );
setNorthEastContact( vtee1 );
} else if (not west()) {
setSouthWestContact( vtee1 );
setNorthEastContact( vtee3 );
}
Added new configuration in NetBuilderHV (for M3 Pins). * New: In Anabatic::NetBuilderHV(), added configurations to manage one Pin M3 + one M1 terminal plus 2 & 3 globals: * _do_2G_1M1_1PinM3() * _do_3G_1M1_1PinM3() They were occuring for the first time in soclayout/experiment7 in the "flat" approach. * New: In Katana::runNegociate(), mark the newly routed netlist as "NetlistTerminal" so it is not placed and routed *again* when reused as an instance (mostly interract with Etesian).
2020-04-20 05:53:07 -05:00
cdebug_tabw(145,-1);
return true;
}
First working recursive place & route (Libre-SOC ALU16 benchmark). * New: In Hurricane::Cell::isLeaf(), a leaf cell was defined as one without instances. Now it can be software defined with the "FlattenLeaf" state. If it is flagged "flatten leaf" is will be considered as a leaf by the trans-hierarchical walkthrough. This flag is also set for Cells in the Alliance CATAL. * New: In Hurricane::Cell_LeafInstanceOccurrences::progress(), add a test to prune leaf cells. * Bug: In Hurricane::getComponentOccurrences(), when calling the locator constructor, the "withLeafCells" argument was forgotten and was getting the value of "doExtraction" (wich, fortunately, was the same). * Change: In Hurricane::HypreNet_LeafPlugOccurrences::progress(), prune the leaf cells instead of the terminal ones. * Change: In Hurricane::HypreNet_ComponentOccurrences::progress(), prune the leaf cells instead of the terminal ones. * New: In Hurricane::Cell, add a "useFlattenLeaf" mode to take into account or ignore the "FlattenLeaf" flag when doing trans- hierarchical walkthrough. Not sure if it shouldn't be systematic. * New: In Etesian::toColoquinte() and other, activate the "FlattenLeaf" mode for all walkthough. So we can stop at already placed sub-blocks. * Bug: In Etesian::toColoquinte(), when placed sub-block where present, the number of instances was under estimated, so the instance array was oversized and the end went unitialized. So we were getting all kind of strange behavior from Coloquinte... Now throw an exception if the number of instances differs from the computed size. * Bug: In Etesian::toColoquinte(), as for the instance array, now correctly compute the size with the ecluded nets (supplies, clocks). * Bug: In Etesian::resetPlacement(), event if the Cell is unplaced, go through the instance list to remove the potential feed cells. The netlist is written back after placement *with* the feed cells added. So, when loading again, they where cluttering the netlist and giving feed redefintions warning. They also generated erratic behaviors. * New: In Anabatic::NetBuilderHV, added configurations for: * 1G_1M1_1PinM3 * 1G_xM1_1PinM2 * 2G_xM1_1PinM2 * doRp_AccessNorthPin * New: In Anabatic::NetBuilderHV, added Pin management in SingleGCell. * Bug: In Anabatic::NetBuilderHV::_do_xG_1PinM3(), the North/South global routing configuration was forgotten (only East/West). * Bug: In Katana::TrackFixedSegment, the blockage net was a static variable initialized at the first run. So we were getting stuck with the blockage net of the first cell to be routed. Of course, it did show only when Katana was run multiple times in the same process.
2020-03-03 17:50:18 -06:00
bool NetBuilderHV::_do_1G_xM1_1PinM2 ()
{
cdebug_log(145,1) << getTypeName() << "::_do_1G_xM1_1PinM2() [Managed Configuration - Optimized] " << getTopology() << endl;
sortRpByX( getRoutingPads(), NoFlags ); // increasing X.
vector<RoutingPad*> rpsM1;
RoutingPad* pinM2 = NULL;
for ( RoutingPad* rp : getRoutingPads() ) {
if (dynamic_cast<Pin*>(rp->getOccurrence().getEntity())) pinM2 = rp;
else rpsM1.push_back( rp );
}
Support for multiple M1 connectors in GCells with M3 pins. * Change: In Anabatic::NetBuilderHV(), in 2-3 globals plus one M3 pins, support up to 3 M1 connectors.
2020-04-20 13:53:49 -05:00
for ( size_t i=1 ; i<rpsM1.size() ; ++i ) {
AutoContact* leftContact = doRp_Access( getGCell(), getRoutingPads()[i-1], HAccess );
AutoContact* rightContact = doRp_Access( getGCell(), getRoutingPads()[i ], HAccess );
AutoSegment::create( leftContact, rightContact, Flags::Horizontal );
}
First working recursive place & route (Libre-SOC ALU16 benchmark). * New: In Hurricane::Cell::isLeaf(), a leaf cell was defined as one without instances. Now it can be software defined with the "FlattenLeaf" state. If it is flagged "flatten leaf" is will be considered as a leaf by the trans-hierarchical walkthrough. This flag is also set for Cells in the Alliance CATAL. * New: In Hurricane::Cell_LeafInstanceOccurrences::progress(), add a test to prune leaf cells. * Bug: In Hurricane::getComponentOccurrences(), when calling the locator constructor, the "withLeafCells" argument was forgotten and was getting the value of "doExtraction" (wich, fortunately, was the same). * Change: In Hurricane::HypreNet_LeafPlugOccurrences::progress(), prune the leaf cells instead of the terminal ones. * Change: In Hurricane::HypreNet_ComponentOccurrences::progress(), prune the leaf cells instead of the terminal ones. * New: In Hurricane::Cell, add a "useFlattenLeaf" mode to take into account or ignore the "FlattenLeaf" flag when doing trans- hierarchical walkthrough. Not sure if it shouldn't be systematic. * New: In Etesian::toColoquinte() and other, activate the "FlattenLeaf" mode for all walkthough. So we can stop at already placed sub-blocks. * Bug: In Etesian::toColoquinte(), when placed sub-block where present, the number of instances was under estimated, so the instance array was oversized and the end went unitialized. So we were getting all kind of strange behavior from Coloquinte... Now throw an exception if the number of instances differs from the computed size. * Bug: In Etesian::toColoquinte(), as for the instance array, now correctly compute the size with the ecluded nets (supplies, clocks). * Bug: In Etesian::resetPlacement(), event if the Cell is unplaced, go through the instance list to remove the potential feed cells. The netlist is written back after placement *with* the feed cells added. So, when loading again, they where cluttering the netlist and giving feed redefintions warning. They also generated erratic behaviors. * New: In Anabatic::NetBuilderHV, added configurations for: * 1G_1M1_1PinM3 * 1G_xM1_1PinM2 * 2G_xM1_1PinM2 * doRp_AccessNorthPin * New: In Anabatic::NetBuilderHV, added Pin management in SingleGCell. * Bug: In Anabatic::NetBuilderHV::_do_xG_1PinM3(), the North/South global routing configuration was forgotten (only East/West). * Bug: In Katana::TrackFixedSegment, the blockage net was a static variable initialized at the first run. So we were getting stuck with the blockage net of the first cell to be routed. Of course, it did show only when Katana was run multiple times in the same process.
2020-03-03 17:50:18 -06:00
AutoContact* turn = NULL;
AutoContact* tee = NULL;
AutoContact* pinM2Contact = NULL;
AutoContact* rpM1Contact = NULL;
doRp_AutoContacts( getGCell(), pinM2, pinM2Contact, rpM1Contact, NoProtect );
Correct too rigid topology in NetBuilderHV::_do_1G_xM1_1PinM2 (E/W). * Change: In Anabatic::NetBuilderHV::_do_1G_xM1_1PinM2(), the M2 pin (horizonal) was directly aligned with the outgoing global through a H-Tee. This was too rigid. Now it can be misaligned.
2020-08-06 08:04:45 -05:00
rpM1Contact = doRp_Access( getGCell(), rpsM1[0], (north() or south()) ? HAccess : NoFlags );
First working recursive place & route (Libre-SOC ALU16 benchmark). * New: In Hurricane::Cell::isLeaf(), a leaf cell was defined as one without instances. Now it can be software defined with the "FlattenLeaf" state. If it is flagged "flatten leaf" is will be considered as a leaf by the trans-hierarchical walkthrough. This flag is also set for Cells in the Alliance CATAL. * New: In Hurricane::Cell_LeafInstanceOccurrences::progress(), add a test to prune leaf cells. * Bug: In Hurricane::getComponentOccurrences(), when calling the locator constructor, the "withLeafCells" argument was forgotten and was getting the value of "doExtraction" (wich, fortunately, was the same). * Change: In Hurricane::HypreNet_LeafPlugOccurrences::progress(), prune the leaf cells instead of the terminal ones. * Change: In Hurricane::HypreNet_ComponentOccurrences::progress(), prune the leaf cells instead of the terminal ones. * New: In Hurricane::Cell, add a "useFlattenLeaf" mode to take into account or ignore the "FlattenLeaf" flag when doing trans- hierarchical walkthrough. Not sure if it shouldn't be systematic. * New: In Etesian::toColoquinte() and other, activate the "FlattenLeaf" mode for all walkthough. So we can stop at already placed sub-blocks. * Bug: In Etesian::toColoquinte(), when placed sub-block where present, the number of instances was under estimated, so the instance array was oversized and the end went unitialized. So we were getting all kind of strange behavior from Coloquinte... Now throw an exception if the number of instances differs from the computed size. * Bug: In Etesian::toColoquinte(), as for the instance array, now correctly compute the size with the ecluded nets (supplies, clocks). * Bug: In Etesian::resetPlacement(), event if the Cell is unplaced, go through the instance list to remove the potential feed cells. The netlist is written back after placement *with* the feed cells added. So, when loading again, they where cluttering the netlist and giving feed redefintions warning. They also generated erratic behaviors. * New: In Anabatic::NetBuilderHV, added configurations for: * 1G_1M1_1PinM3 * 1G_xM1_1PinM2 * 2G_xM1_1PinM2 * doRp_AccessNorthPin * New: In Anabatic::NetBuilderHV, added Pin management in SingleGCell. * Bug: In Anabatic::NetBuilderHV::_do_xG_1PinM3(), the North/South global routing configuration was forgotten (only East/West). * Bug: In Katana::TrackFixedSegment, the blockage net was a static variable initialized at the first run. So we were getting stuck with the blockage net of the first cell to be routed. Of course, it did show only when Katana was run multiple times in the same process.
2020-03-03 17:50:18 -06:00
if (north() or south()) {
turn = AutoContactTurn::create( getGCell(), getNet(), Session::getContactLayer(1) );
tee = AutoContactVTee::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( rpM1Contact , tee , Flags::Horizontal );
AutoSegment::create( pinM2Contact, turn, Flags::Horizontal );
AutoSegment::create( tee , turn, Flags::Vertical );
Correct too rigid topology in NetBuilderHV::_do_1G_xM1_1PinM2 (E/W). * Change: In Anabatic::NetBuilderHV::_do_1G_xM1_1PinM2(), the M2 pin (horizonal) was directly aligned with the outgoing global through a H-Tee. This was too rigid. Now it can be misaligned.
2020-08-06 08:04:45 -05:00
setBothCornerContacts( tee );
First working recursive place & route (Libre-SOC ALU16 benchmark). * New: In Hurricane::Cell::isLeaf(), a leaf cell was defined as one without instances. Now it can be software defined with the "FlattenLeaf" state. If it is flagged "flatten leaf" is will be considered as a leaf by the trans-hierarchical walkthrough. This flag is also set for Cells in the Alliance CATAL. * New: In Hurricane::Cell_LeafInstanceOccurrences::progress(), add a test to prune leaf cells. * Bug: In Hurricane::getComponentOccurrences(), when calling the locator constructor, the "withLeafCells" argument was forgotten and was getting the value of "doExtraction" (wich, fortunately, was the same). * Change: In Hurricane::HypreNet_LeafPlugOccurrences::progress(), prune the leaf cells instead of the terminal ones. * Change: In Hurricane::HypreNet_ComponentOccurrences::progress(), prune the leaf cells instead of the terminal ones. * New: In Hurricane::Cell, add a "useFlattenLeaf" mode to take into account or ignore the "FlattenLeaf" flag when doing trans- hierarchical walkthrough. Not sure if it shouldn't be systematic. * New: In Etesian::toColoquinte() and other, activate the "FlattenLeaf" mode for all walkthough. So we can stop at already placed sub-blocks. * Bug: In Etesian::toColoquinte(), when placed sub-block where present, the number of instances was under estimated, so the instance array was oversized and the end went unitialized. So we were getting all kind of strange behavior from Coloquinte... Now throw an exception if the number of instances differs from the computed size. * Bug: In Etesian::toColoquinte(), as for the instance array, now correctly compute the size with the ecluded nets (supplies, clocks). * Bug: In Etesian::resetPlacement(), event if the Cell is unplaced, go through the instance list to remove the potential feed cells. The netlist is written back after placement *with* the feed cells added. So, when loading again, they where cluttering the netlist and giving feed redefintions warning. They also generated erratic behaviors. * New: In Anabatic::NetBuilderHV, added configurations for: * 1G_1M1_1PinM3 * 1G_xM1_1PinM2 * 2G_xM1_1PinM2 * doRp_AccessNorthPin * New: In Anabatic::NetBuilderHV, added Pin management in SingleGCell. * Bug: In Anabatic::NetBuilderHV::_do_xG_1PinM3(), the North/South global routing configuration was forgotten (only East/West). * Bug: In Katana::TrackFixedSegment, the blockage net was a static variable initialized at the first run. So we were getting stuck with the blockage net of the first cell to be routed. Of course, it did show only when Katana was run multiple times in the same process.
2020-03-03 17:50:18 -06:00
} else {
turn = AutoContactTurn::create( getGCell(), getNet(), Session::getContactLayer(1) );
Correct too rigid topology in NetBuilderHV::_do_1G_xM1_1PinM2 (E/W). * Change: In Anabatic::NetBuilderHV::_do_1G_xM1_1PinM2(), the M2 pin (horizonal) was directly aligned with the outgoing global through a H-Tee. This was too rigid. Now it can be misaligned.
2020-08-06 08:04:45 -05:00
tee = AutoContactVTee::create( getGCell(), getNet(), Session::getContactLayer(1) );
First working recursive place & route (Libre-SOC ALU16 benchmark). * New: In Hurricane::Cell::isLeaf(), a leaf cell was defined as one without instances. Now it can be software defined with the "FlattenLeaf" state. If it is flagged "flatten leaf" is will be considered as a leaf by the trans-hierarchical walkthrough. This flag is also set for Cells in the Alliance CATAL. * New: In Hurricane::Cell_LeafInstanceOccurrences::progress(), add a test to prune leaf cells. * Bug: In Hurricane::getComponentOccurrences(), when calling the locator constructor, the "withLeafCells" argument was forgotten and was getting the value of "doExtraction" (wich, fortunately, was the same). * Change: In Hurricane::HypreNet_LeafPlugOccurrences::progress(), prune the leaf cells instead of the terminal ones. * Change: In Hurricane::HypreNet_ComponentOccurrences::progress(), prune the leaf cells instead of the terminal ones. * New: In Hurricane::Cell, add a "useFlattenLeaf" mode to take into account or ignore the "FlattenLeaf" flag when doing trans- hierarchical walkthrough. Not sure if it shouldn't be systematic. * New: In Etesian::toColoquinte() and other, activate the "FlattenLeaf" mode for all walkthough. So we can stop at already placed sub-blocks. * Bug: In Etesian::toColoquinte(), when placed sub-block where present, the number of instances was under estimated, so the instance array was oversized and the end went unitialized. So we were getting all kind of strange behavior from Coloquinte... Now throw an exception if the number of instances differs from the computed size. * Bug: In Etesian::toColoquinte(), as for the instance array, now correctly compute the size with the ecluded nets (supplies, clocks). * Bug: In Etesian::resetPlacement(), event if the Cell is unplaced, go through the instance list to remove the potential feed cells. The netlist is written back after placement *with* the feed cells added. So, when loading again, they where cluttering the netlist and giving feed redefintions warning. They also generated erratic behaviors. * New: In Anabatic::NetBuilderHV, added configurations for: * 1G_1M1_1PinM3 * 1G_xM1_1PinM2 * 2G_xM1_1PinM2 * doRp_AccessNorthPin * New: In Anabatic::NetBuilderHV, added Pin management in SingleGCell. * Bug: In Anabatic::NetBuilderHV::_do_xG_1PinM3(), the North/South global routing configuration was forgotten (only East/West). * Bug: In Katana::TrackFixedSegment, the blockage net was a static variable initialized at the first run. So we were getting stuck with the blockage net of the first cell to be routed. Of course, it did show only when Katana was run multiple times in the same process.
2020-03-03 17:50:18 -06:00
Correct too rigid topology in NetBuilderHV::_do_1G_xM1_1PinM2 (E/W). * Change: In Anabatic::NetBuilderHV::_do_1G_xM1_1PinM2(), the M2 pin (horizonal) was directly aligned with the outgoing global through a H-Tee. This was too rigid. Now it can be misaligned.
2020-08-06 08:04:45 -05:00
AutoSegment::create( rpM1Contact , tee , Flags::Vertical );
First working recursive place & route (Libre-SOC ALU16 benchmark). * New: In Hurricane::Cell::isLeaf(), a leaf cell was defined as one without instances. Now it can be software defined with the "FlattenLeaf" state. If it is flagged "flatten leaf" is will be considered as a leaf by the trans-hierarchical walkthrough. This flag is also set for Cells in the Alliance CATAL. * New: In Hurricane::Cell_LeafInstanceOccurrences::progress(), add a test to prune leaf cells. * Bug: In Hurricane::getComponentOccurrences(), when calling the locator constructor, the "withLeafCells" argument was forgotten and was getting the value of "doExtraction" (wich, fortunately, was the same). * Change: In Hurricane::HypreNet_LeafPlugOccurrences::progress(), prune the leaf cells instead of the terminal ones. * Change: In Hurricane::HypreNet_ComponentOccurrences::progress(), prune the leaf cells instead of the terminal ones. * New: In Hurricane::Cell, add a "useFlattenLeaf" mode to take into account or ignore the "FlattenLeaf" flag when doing trans- hierarchical walkthrough. Not sure if it shouldn't be systematic. * New: In Etesian::toColoquinte() and other, activate the "FlattenLeaf" mode for all walkthough. So we can stop at already placed sub-blocks. * Bug: In Etesian::toColoquinte(), when placed sub-block where present, the number of instances was under estimated, so the instance array was oversized and the end went unitialized. So we were getting all kind of strange behavior from Coloquinte... Now throw an exception if the number of instances differs from the computed size. * Bug: In Etesian::toColoquinte(), as for the instance array, now correctly compute the size with the ecluded nets (supplies, clocks). * Bug: In Etesian::resetPlacement(), event if the Cell is unplaced, go through the instance list to remove the potential feed cells. The netlist is written back after placement *with* the feed cells added. So, when loading again, they where cluttering the netlist and giving feed redefintions warning. They also generated erratic behaviors. * New: In Anabatic::NetBuilderHV, added configurations for: * 1G_1M1_1PinM3 * 1G_xM1_1PinM2 * 2G_xM1_1PinM2 * doRp_AccessNorthPin * New: In Anabatic::NetBuilderHV, added Pin management in SingleGCell. * Bug: In Anabatic::NetBuilderHV::_do_xG_1PinM3(), the North/South global routing configuration was forgotten (only East/West). * Bug: In Katana::TrackFixedSegment, the blockage net was a static variable initialized at the first run. So we were getting stuck with the blockage net of the first cell to be routed. Of course, it did show only when Katana was run multiple times in the same process.
2020-03-03 17:50:18 -06:00
AutoSegment::create( pinM2Contact, tee , Flags::Horizontal );
AutoSegment::create( tee , turn, Flags::Vertical );
Correct too rigid topology in NetBuilderHV::_do_1G_xM1_1PinM2 (E/W). * Change: In Anabatic::NetBuilderHV::_do_1G_xM1_1PinM2(), the M2 pin (horizonal) was directly aligned with the outgoing global through a H-Tee. This was too rigid. Now it can be misaligned.
2020-08-06 08:04:45 -05:00
setBothCornerContacts( turn );
First working recursive place & route (Libre-SOC ALU16 benchmark). * New: In Hurricane::Cell::isLeaf(), a leaf cell was defined as one without instances. Now it can be software defined with the "FlattenLeaf" state. If it is flagged "flatten leaf" is will be considered as a leaf by the trans-hierarchical walkthrough. This flag is also set for Cells in the Alliance CATAL. * New: In Hurricane::Cell_LeafInstanceOccurrences::progress(), add a test to prune leaf cells. * Bug: In Hurricane::getComponentOccurrences(), when calling the locator constructor, the "withLeafCells" argument was forgotten and was getting the value of "doExtraction" (wich, fortunately, was the same). * Change: In Hurricane::HypreNet_LeafPlugOccurrences::progress(), prune the leaf cells instead of the terminal ones. * Change: In Hurricane::HypreNet_ComponentOccurrences::progress(), prune the leaf cells instead of the terminal ones. * New: In Hurricane::Cell, add a "useFlattenLeaf" mode to take into account or ignore the "FlattenLeaf" flag when doing trans- hierarchical walkthrough. Not sure if it shouldn't be systematic. * New: In Etesian::toColoquinte() and other, activate the "FlattenLeaf" mode for all walkthough. So we can stop at already placed sub-blocks. * Bug: In Etesian::toColoquinte(), when placed sub-block where present, the number of instances was under estimated, so the instance array was oversized and the end went unitialized. So we were getting all kind of strange behavior from Coloquinte... Now throw an exception if the number of instances differs from the computed size. * Bug: In Etesian::toColoquinte(), as for the instance array, now correctly compute the size with the ecluded nets (supplies, clocks). * Bug: In Etesian::resetPlacement(), event if the Cell is unplaced, go through the instance list to remove the potential feed cells. The netlist is written back after placement *with* the feed cells added. So, when loading again, they where cluttering the netlist and giving feed redefintions warning. They also generated erratic behaviors. * New: In Anabatic::NetBuilderHV, added configurations for: * 1G_1M1_1PinM3 * 1G_xM1_1PinM2 * 2G_xM1_1PinM2 * doRp_AccessNorthPin * New: In Anabatic::NetBuilderHV, added Pin management in SingleGCell. * Bug: In Anabatic::NetBuilderHV::_do_xG_1PinM3(), the North/South global routing configuration was forgotten (only East/West). * Bug: In Katana::TrackFixedSegment, the blockage net was a static variable initialized at the first run. So we were getting stuck with the blockage net of the first cell to be routed. Of course, it did show only when Katana was run multiple times in the same process.
2020-03-03 17:50:18 -06:00
}
for ( size_t i=1 ; i<rpsM1.size() ; ++i ) {
AutoContact* leftContact = doRp_Access( getGCell(), getRoutingPads()[i-1], HAccess );
AutoContact* rightContact = doRp_Access( getGCell(), getRoutingPads()[i ], HAccess );
AutoSegment::create( leftContact, rightContact, Flags::Horizontal );
}
cdebug_tabw(145,-1);
return true;
}
bool NetBuilderHV::_do_2G_xM1_1PinM2 ()
{
cdebug_log(145,1) << getTypeName() << "::_do_2G_xM1_1PinM2() [Managed Configuration - Optimized] " << getTopology() << endl;
sortRpByX( getRoutingPads(), NoFlags ); // increasing X.
vector<RoutingPad*> rpsM1;
RoutingPad* pinM2 = NULL;
for ( RoutingPad* rp : getRoutingPads() ) {
if (dynamic_cast<Pin*>(rp->getOccurrence().getEntity())) pinM2 = rp;
else rpsM1.push_back( rp );
}
AutoContact* pinM2Contact = NULL;
AutoContact* rpM1Contact = NULL;
doRp_AutoContacts( getGCell(), pinM2, pinM2Contact, rpM1Contact, NoFlags );
rpM1Contact = doRp_Access( getGCell(), rpsM1[0], NoFlags );
if (north() and south()) {
Bug fix: Bad topology in NetBuilderHV::_do_2G_xM1_1PinM2(). * Bug: In Anabatic::NetBuilderHV::_do_2G_xM1_1PinM2(), in the case of a bend of global wires, the topology was faulty. Badly connected HTees.
2020-08-02 11:11:39 -05:00
cdebug_log(145,0) << "Case north & south." << endl;
First working recursive place & route (Libre-SOC ALU16 benchmark). * New: In Hurricane::Cell::isLeaf(), a leaf cell was defined as one without instances. Now it can be software defined with the "FlattenLeaf" state. If it is flagged "flatten leaf" is will be considered as a leaf by the trans-hierarchical walkthrough. This flag is also set for Cells in the Alliance CATAL. * New: In Hurricane::Cell_LeafInstanceOccurrences::progress(), add a test to prune leaf cells. * Bug: In Hurricane::getComponentOccurrences(), when calling the locator constructor, the "withLeafCells" argument was forgotten and was getting the value of "doExtraction" (wich, fortunately, was the same). * Change: In Hurricane::HypreNet_LeafPlugOccurrences::progress(), prune the leaf cells instead of the terminal ones. * Change: In Hurricane::HypreNet_ComponentOccurrences::progress(), prune the leaf cells instead of the terminal ones. * New: In Hurricane::Cell, add a "useFlattenLeaf" mode to take into account or ignore the "FlattenLeaf" flag when doing trans- hierarchical walkthrough. Not sure if it shouldn't be systematic. * New: In Etesian::toColoquinte() and other, activate the "FlattenLeaf" mode for all walkthough. So we can stop at already placed sub-blocks. * Bug: In Etesian::toColoquinte(), when placed sub-block where present, the number of instances was under estimated, so the instance array was oversized and the end went unitialized. So we were getting all kind of strange behavior from Coloquinte... Now throw an exception if the number of instances differs from the computed size. * Bug: In Etesian::toColoquinte(), as for the instance array, now correctly compute the size with the ecluded nets (supplies, clocks). * Bug: In Etesian::resetPlacement(), event if the Cell is unplaced, go through the instance list to remove the potential feed cells. The netlist is written back after placement *with* the feed cells added. So, when loading again, they where cluttering the netlist and giving feed redefintions warning. They also generated erratic behaviors. * New: In Anabatic::NetBuilderHV, added configurations for: * 1G_1M1_1PinM3 * 1G_xM1_1PinM2 * 2G_xM1_1PinM2 * doRp_AccessNorthPin * New: In Anabatic::NetBuilderHV, added Pin management in SingleGCell. * Bug: In Anabatic::NetBuilderHV::_do_xG_1PinM3(), the North/South global routing configuration was forgotten (only East/West). * Bug: In Katana::TrackFixedSegment, the blockage net was a static variable initialized at the first run. So we were getting stuck with the blockage net of the first cell to be routed. Of course, it did show only when Katana was run multiple times in the same process.
2020-03-03 17:50:18 -06:00
AutoContact* htee = AutoContactHTee::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoContact* vtee = AutoContactVTee::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( pinM2Contact, htee, Flags::Horizontal );
AutoSegment::create( rpM1Contact , htee, Flags::Vertical );
AutoSegment::create( htee , vtee, Flags::Horizontal );
setBothCornerContacts( vtee );
} else if (east() and west()) {
Bug fix: Bad topology in NetBuilderHV::_do_2G_xM1_1PinM2(). * Bug: In Anabatic::NetBuilderHV::_do_2G_xM1_1PinM2(), in the case of a bend of global wires, the topology was faulty. Badly connected HTees.
2020-08-02 11:11:39 -05:00
cdebug_log(145,0) << "Case east & west." << endl;
First working recursive place & route (Libre-SOC ALU16 benchmark). * New: In Hurricane::Cell::isLeaf(), a leaf cell was defined as one without instances. Now it can be software defined with the "FlattenLeaf" state. If it is flagged "flatten leaf" is will be considered as a leaf by the trans-hierarchical walkthrough. This flag is also set for Cells in the Alliance CATAL. * New: In Hurricane::Cell_LeafInstanceOccurrences::progress(), add a test to prune leaf cells. * Bug: In Hurricane::getComponentOccurrences(), when calling the locator constructor, the "withLeafCells" argument was forgotten and was getting the value of "doExtraction" (wich, fortunately, was the same). * Change: In Hurricane::HypreNet_LeafPlugOccurrences::progress(), prune the leaf cells instead of the terminal ones. * Change: In Hurricane::HypreNet_ComponentOccurrences::progress(), prune the leaf cells instead of the terminal ones. * New: In Hurricane::Cell, add a "useFlattenLeaf" mode to take into account or ignore the "FlattenLeaf" flag when doing trans- hierarchical walkthrough. Not sure if it shouldn't be systematic. * New: In Etesian::toColoquinte() and other, activate the "FlattenLeaf" mode for all walkthough. So we can stop at already placed sub-blocks. * Bug: In Etesian::toColoquinte(), when placed sub-block where present, the number of instances was under estimated, so the instance array was oversized and the end went unitialized. So we were getting all kind of strange behavior from Coloquinte... Now throw an exception if the number of instances differs from the computed size. * Bug: In Etesian::toColoquinte(), as for the instance array, now correctly compute the size with the ecluded nets (supplies, clocks). * Bug: In Etesian::resetPlacement(), event if the Cell is unplaced, go through the instance list to remove the potential feed cells. The netlist is written back after placement *with* the feed cells added. So, when loading again, they where cluttering the netlist and giving feed redefintions warning. They also generated erratic behaviors. * New: In Anabatic::NetBuilderHV, added configurations for: * 1G_1M1_1PinM3 * 1G_xM1_1PinM2 * 2G_xM1_1PinM2 * doRp_AccessNorthPin * New: In Anabatic::NetBuilderHV, added Pin management in SingleGCell. * Bug: In Anabatic::NetBuilderHV::_do_xG_1PinM3(), the North/South global routing configuration was forgotten (only East/West). * Bug: In Katana::TrackFixedSegment, the blockage net was a static variable initialized at the first run. So we were getting stuck with the blockage net of the first cell to be routed. Of course, it did show only when Katana was run multiple times in the same process.
2020-03-03 17:50:18 -06:00
AutoContact* htee1 = AutoContactHTee::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoContact* htee2 = AutoContactHTee::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoContact* turn = AutoContactTurn::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( pinM2Contact, htee1, Flags::Horizontal );
AutoSegment::create( rpM1Contact , htee1, Flags::Vertical );
AutoSegment::create( htee1 , turn , Flags::Horizontal );
AutoSegment::create( htee2 , turn , Flags::Vertical );
setBothCornerContacts( htee2 );
} else {
Bug fix: Bad topology in NetBuilderHV::_do_2G_xM1_1PinM2(). * Bug: In Anabatic::NetBuilderHV::_do_2G_xM1_1PinM2(), in the case of a bend of global wires, the topology was faulty. Badly connected HTees.
2020-08-02 11:11:39 -05:00
cdebug_log(145,0) << "Case bend." << endl;
Bugs in NetBuilderHV for topologies containing metal3 pins. * In NetBuilder::_do_xG_1PinM3(), in the three global case, an incorrect connexion was made. * In NetBuilder::_do_2G_xM1_1PinM2(), in the global E+W case, segment between vtee1 <-> turn1 must be vertical.
2020-09-14 08:12:37 -05:00
AutoContact* vtee1 = AutoContactVTee::create( getGCell(), getNet(), Session::getContactLayer(1) );
First working recursive place & route (Libre-SOC ALU16 benchmark). * New: In Hurricane::Cell::isLeaf(), a leaf cell was defined as one without instances. Now it can be software defined with the "FlattenLeaf" state. If it is flagged "flatten leaf" is will be considered as a leaf by the trans-hierarchical walkthrough. This flag is also set for Cells in the Alliance CATAL. * New: In Hurricane::Cell_LeafInstanceOccurrences::progress(), add a test to prune leaf cells. * Bug: In Hurricane::getComponentOccurrences(), when calling the locator constructor, the "withLeafCells" argument was forgotten and was getting the value of "doExtraction" (wich, fortunately, was the same). * Change: In Hurricane::HypreNet_LeafPlugOccurrences::progress(), prune the leaf cells instead of the terminal ones. * Change: In Hurricane::HypreNet_ComponentOccurrences::progress(), prune the leaf cells instead of the terminal ones. * New: In Hurricane::Cell, add a "useFlattenLeaf" mode to take into account or ignore the "FlattenLeaf" flag when doing trans- hierarchical walkthrough. Not sure if it shouldn't be systematic. * New: In Etesian::toColoquinte() and other, activate the "FlattenLeaf" mode for all walkthough. So we can stop at already placed sub-blocks. * Bug: In Etesian::toColoquinte(), when placed sub-block where present, the number of instances was under estimated, so the instance array was oversized and the end went unitialized. So we were getting all kind of strange behavior from Coloquinte... Now throw an exception if the number of instances differs from the computed size. * Bug: In Etesian::toColoquinte(), as for the instance array, now correctly compute the size with the ecluded nets (supplies, clocks). * Bug: In Etesian::resetPlacement(), event if the Cell is unplaced, go through the instance list to remove the potential feed cells. The netlist is written back after placement *with* the feed cells added. So, when loading again, they where cluttering the netlist and giving feed redefintions warning. They also generated erratic behaviors. * New: In Anabatic::NetBuilderHV, added configurations for: * 1G_1M1_1PinM3 * 1G_xM1_1PinM2 * 2G_xM1_1PinM2 * doRp_AccessNorthPin * New: In Anabatic::NetBuilderHV, added Pin management in SingleGCell. * Bug: In Anabatic::NetBuilderHV::_do_xG_1PinM3(), the North/South global routing configuration was forgotten (only East/West). * Bug: In Katana::TrackFixedSegment, the blockage net was a static variable initialized at the first run. So we were getting stuck with the blockage net of the first cell to be routed. Of course, it did show only when Katana was run multiple times in the same process.
2020-03-03 17:50:18 -06:00
AutoContact* htee2 = AutoContactHTee::create( getGCell(), getNet(), Session::getContactLayer(1) );
Bug fix: Bad topology in NetBuilderHV::_do_2G_xM1_1PinM2() again. * Bug: In Anabatic::NetBuilderHV::_do_2G_xM1_1PinM2(), in the case of a bend of global wires, the topology was faulty. Too rigid topology, the global wire was forcibly aligned on the external Pin.
2020-08-03 12:28:16 -05:00
AutoContact* turn1 = AutoContactTurn::create( getGCell(), getNet(), Session::getContactLayer(1) );
First working recursive place & route (Libre-SOC ALU16 benchmark). * New: In Hurricane::Cell::isLeaf(), a leaf cell was defined as one without instances. Now it can be software defined with the "FlattenLeaf" state. If it is flagged "flatten leaf" is will be considered as a leaf by the trans-hierarchical walkthrough. This flag is also set for Cells in the Alliance CATAL. * New: In Hurricane::Cell_LeafInstanceOccurrences::progress(), add a test to prune leaf cells. * Bug: In Hurricane::getComponentOccurrences(), when calling the locator constructor, the "withLeafCells" argument was forgotten and was getting the value of "doExtraction" (wich, fortunately, was the same). * Change: In Hurricane::HypreNet_LeafPlugOccurrences::progress(), prune the leaf cells instead of the terminal ones. * Change: In Hurricane::HypreNet_ComponentOccurrences::progress(), prune the leaf cells instead of the terminal ones. * New: In Hurricane::Cell, add a "useFlattenLeaf" mode to take into account or ignore the "FlattenLeaf" flag when doing trans- hierarchical walkthrough. Not sure if it shouldn't be systematic. * New: In Etesian::toColoquinte() and other, activate the "FlattenLeaf" mode for all walkthough. So we can stop at already placed sub-blocks. * Bug: In Etesian::toColoquinte(), when placed sub-block where present, the number of instances was under estimated, so the instance array was oversized and the end went unitialized. So we were getting all kind of strange behavior from Coloquinte... Now throw an exception if the number of instances differs from the computed size. * Bug: In Etesian::toColoquinte(), as for the instance array, now correctly compute the size with the ecluded nets (supplies, clocks). * Bug: In Etesian::resetPlacement(), event if the Cell is unplaced, go through the instance list to remove the potential feed cells. The netlist is written back after placement *with* the feed cells added. So, when loading again, they where cluttering the netlist and giving feed redefintions warning. They also generated erratic behaviors. * New: In Anabatic::NetBuilderHV, added configurations for: * 1G_1M1_1PinM3 * 1G_xM1_1PinM2 * 2G_xM1_1PinM2 * doRp_AccessNorthPin * New: In Anabatic::NetBuilderHV, added Pin management in SingleGCell. * Bug: In Anabatic::NetBuilderHV::_do_xG_1PinM3(), the North/South global routing configuration was forgotten (only East/West). * Bug: In Katana::TrackFixedSegment, the blockage net was a static variable initialized at the first run. So we were getting stuck with the blockage net of the first cell to be routed. Of course, it did show only when Katana was run multiple times in the same process.
2020-03-03 17:50:18 -06:00
Bug fix: Bad topology in NetBuilderHV::_do_2G_xM1_1PinM2() again. * Bug: In Anabatic::NetBuilderHV::_do_2G_xM1_1PinM2(), in the case of a bend of global wires, the topology was faulty. Too rigid topology, the global wire was forcibly aligned on the external Pin.
2020-08-03 12:28:16 -05:00
AutoSegment::create( pinM2Contact, vtee1, Flags::Horizontal );
AutoSegment::create( rpM1Contact , vtee1, Flags::Vertical );
Bugs in NetBuilderHV for topologies containing metal3 pins. * In NetBuilder::_do_xG_1PinM3(), in the three global case, an incorrect connexion was made. * In NetBuilder::_do_2G_xM1_1PinM2(), in the global E+W case, segment between vtee1 <-> turn1 must be vertical.
2020-09-14 08:12:37 -05:00
AutoSegment::create( vtee1 , turn1, Flags::Vertical );
Bug fix: Bad topology in NetBuilderHV::_do_2G_xM1_1PinM2() again. * Bug: In Anabatic::NetBuilderHV::_do_2G_xM1_1PinM2(), in the case of a bend of global wires, the topology was faulty. Too rigid topology, the global wire was forcibly aligned on the external Pin.
2020-08-03 12:28:16 -05:00
AutoSegment::create( turn1 , htee2, Flags::Horizontal );
Bug fix: Bad topology in NetBuilderHV::_do_2G_xM1_1PinM2(). * Bug: In Anabatic::NetBuilderHV::_do_2G_xM1_1PinM2(), in the case of a bend of global wires, the topology was faulty. Badly connected HTees.
2020-08-02 11:11:39 -05:00
setBothCornerContacts( htee2 );
First working recursive place & route (Libre-SOC ALU16 benchmark). * New: In Hurricane::Cell::isLeaf(), a leaf cell was defined as one without instances. Now it can be software defined with the "FlattenLeaf" state. If it is flagged "flatten leaf" is will be considered as a leaf by the trans-hierarchical walkthrough. This flag is also set for Cells in the Alliance CATAL. * New: In Hurricane::Cell_LeafInstanceOccurrences::progress(), add a test to prune leaf cells. * Bug: In Hurricane::getComponentOccurrences(), when calling the locator constructor, the "withLeafCells" argument was forgotten and was getting the value of "doExtraction" (wich, fortunately, was the same). * Change: In Hurricane::HypreNet_LeafPlugOccurrences::progress(), prune the leaf cells instead of the terminal ones. * Change: In Hurricane::HypreNet_ComponentOccurrences::progress(), prune the leaf cells instead of the terminal ones. * New: In Hurricane::Cell, add a "useFlattenLeaf" mode to take into account or ignore the "FlattenLeaf" flag when doing trans- hierarchical walkthrough. Not sure if it shouldn't be systematic. * New: In Etesian::toColoquinte() and other, activate the "FlattenLeaf" mode for all walkthough. So we can stop at already placed sub-blocks. * Bug: In Etesian::toColoquinte(), when placed sub-block where present, the number of instances was under estimated, so the instance array was oversized and the end went unitialized. So we were getting all kind of strange behavior from Coloquinte... Now throw an exception if the number of instances differs from the computed size. * Bug: In Etesian::toColoquinte(), as for the instance array, now correctly compute the size with the ecluded nets (supplies, clocks). * Bug: In Etesian::resetPlacement(), event if the Cell is unplaced, go through the instance list to remove the potential feed cells. The netlist is written back after placement *with* the feed cells added. So, when loading again, they where cluttering the netlist and giving feed redefintions warning. They also generated erratic behaviors. * New: In Anabatic::NetBuilderHV, added configurations for: * 1G_1M1_1PinM3 * 1G_xM1_1PinM2 * 2G_xM1_1PinM2 * doRp_AccessNorthPin * New: In Anabatic::NetBuilderHV, added Pin management in SingleGCell. * Bug: In Anabatic::NetBuilderHV::_do_xG_1PinM3(), the North/South global routing configuration was forgotten (only East/West). * Bug: In Katana::TrackFixedSegment, the blockage net was a static variable initialized at the first run. So we were getting stuck with the blockage net of the first cell to be routed. Of course, it did show only when Katana was run multiple times in the same process.
2020-03-03 17:50:18 -06:00
}
Bug fix: Bad topology in NetBuilderHV::_do_2G_xM1_1PinM2(). * Bug: In Anabatic::NetBuilderHV::_do_2G_xM1_1PinM2(), in the case of a bend of global wires, the topology was faulty. Badly connected HTees.
2020-08-02 11:11:39 -05:00
cdebug_log(145,0) << "Wiring all M1 together." << endl;
First working recursive place & route (Libre-SOC ALU16 benchmark). * New: In Hurricane::Cell::isLeaf(), a leaf cell was defined as one without instances. Now it can be software defined with the "FlattenLeaf" state. If it is flagged "flatten leaf" is will be considered as a leaf by the trans-hierarchical walkthrough. This flag is also set for Cells in the Alliance CATAL. * New: In Hurricane::Cell_LeafInstanceOccurrences::progress(), add a test to prune leaf cells. * Bug: In Hurricane::getComponentOccurrences(), when calling the locator constructor, the "withLeafCells" argument was forgotten and was getting the value of "doExtraction" (wich, fortunately, was the same). * Change: In Hurricane::HypreNet_LeafPlugOccurrences::progress(), prune the leaf cells instead of the terminal ones. * Change: In Hurricane::HypreNet_ComponentOccurrences::progress(), prune the leaf cells instead of the terminal ones. * New: In Hurricane::Cell, add a "useFlattenLeaf" mode to take into account or ignore the "FlattenLeaf" flag when doing trans- hierarchical walkthrough. Not sure if it shouldn't be systematic. * New: In Etesian::toColoquinte() and other, activate the "FlattenLeaf" mode for all walkthough. So we can stop at already placed sub-blocks. * Bug: In Etesian::toColoquinte(), when placed sub-block where present, the number of instances was under estimated, so the instance array was oversized and the end went unitialized. So we were getting all kind of strange behavior from Coloquinte... Now throw an exception if the number of instances differs from the computed size. * Bug: In Etesian::toColoquinte(), as for the instance array, now correctly compute the size with the ecluded nets (supplies, clocks). * Bug: In Etesian::resetPlacement(), event if the Cell is unplaced, go through the instance list to remove the potential feed cells. The netlist is written back after placement *with* the feed cells added. So, when loading again, they where cluttering the netlist and giving feed redefintions warning. They also generated erratic behaviors. * New: In Anabatic::NetBuilderHV, added configurations for: * 1G_1M1_1PinM3 * 1G_xM1_1PinM2 * 2G_xM1_1PinM2 * doRp_AccessNorthPin * New: In Anabatic::NetBuilderHV, added Pin management in SingleGCell. * Bug: In Anabatic::NetBuilderHV::_do_xG_1PinM3(), the North/South global routing configuration was forgotten (only East/West). * Bug: In Katana::TrackFixedSegment, the blockage net was a static variable initialized at the first run. So we were getting stuck with the blockage net of the first cell to be routed. Of course, it did show only when Katana was run multiple times in the same process.
2020-03-03 17:50:18 -06:00
for ( size_t i=1 ; i<rpsM1.size() ; ++i ) {
AutoContact* leftContact = doRp_Access( getGCell(), getRoutingPads()[i-1], HAccess );
AutoContact* rightContact = doRp_Access( getGCell(), getRoutingPads()[i ], HAccess );
AutoSegment::create( leftContact, rightContact, Flags::Horizontal );
}
cdebug_tabw(145,-1);
return true;
}
bool NetBuilderHV::_do_1G_1M1_1PinM3 ()
{
cdebug_log(145,1) << getTypeName() << "::_do_1G_1M1_1PinM3() [Managed Configuration - Optimized] " << getTopology() << endl;
AutoContact* rpSourceContact = NULL;
RoutingPad* pinM3 = getRoutingPads()[0];
RoutingPad* rpM1 = getRoutingPads()[1];
if (dynamic_cast<Pin*>(rpM1->getOccurrence().getEntity())) std::swap( rpM1, pinM3 );
AutoContact* contact1 = doRp_Access( getGCell(), rpM1, HAccess );
AutoContact* htee = AutoContactHTee::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( contact1, htee, Flags::Horizontal );
rpSourceContact = doRp_AccessNorthPin( getGCell(), pinM3 );
if (north() or south()) {
AutoContact* turn = AutoContactTurn::create( getGCell(), getNet(), Session::getContactLayer(1) );
AutoSegment::create( rpSourceContact, turn, Flags::Vertical );
AutoSegment::create( turn , htee, Flags::Horizontal );
} else {
AutoSegment::create( rpSourceContact, htee, Flags::Vertical );
}
setBothCornerContacts( htee );
cdebug_tabw(145,-1);
return true;
}
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
bool NetBuilderHV::_do_2G_1M1 ()
{
return _do_xG_1M1();
}
bool NetBuilderHV::_do_xG_1M1 ()
{
if ((int)getConnexity().fields.M1 != 1) return false;
cdebug_log(145,1) << getTypeName()
<< "::_do_xG_" << (int)getConnexity().fields.M1
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
<< "M1() [G:" << (int)getConnexity().fields.globals << " Managed Configuration]" << 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
cdebug_log(145,0) << "getConnexity(): " << getConnexity().connexity << endl;
cdebug_log(145,0) << "north: " << north() << endl;
cdebug_log(145,0) << "south: " << south() << endl;
cdebug_log(145,0) << "east: " << east () << endl;
cdebug_log(145,0) << "west: " << west () << endl;
const Layer* viaLayer1 = Session::getContactLayer(1);
if (getConnexity().fields.globals == 2) {
if (north() and south()) {
AutoContact* contact1 = doRp_Access( getGCell(), getRoutingPads()[0], HAccess );
In NetBuilderHV::_do_xG_1M1(), less rigid topology for straight vertical.
2021-06-10 08:29:00 -05:00
AutoContact* contact2 = AutoContactHTee::create( getGCell(), getNet(), viaLayer1 );
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
AutoSegment::create( contact1, contact2, Flags::Horizontal );
In NetBuilderHV::_do_xG_1M1(), less rigid topology for straight vertical.
2021-06-10 08:29:00 -05:00
contact1 = AutoContactTurn::create( getGCell(), getNet(), viaLayer1 );
AutoSegment::create( contact1, contact2, Flags::Horizontal );
setNorthEastContact( contact1 );
setSouthWestContact( contact2 );
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
} else if (east() and west()) {
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
// AutoContact* contact1 = doRp_Access( getGCell(), getRoutingPads()[0], NoFlags );
// AutoContact* contact2 = AutoContactHTee::create( getGCell(), getNet(), viaLayer1 );
// AutoSegment::create( contact1, contact2, Flags::Vertical );
// setBothCornerContacts( contact2 );
AutoContact* contact1 = NULL;
AutoContact* contact2 = NULL;
uint64_t flags = checkRoutingPadSize( getRoutingPads()[0] );
if (flags & VSmall) {
doRp_AutoContacts( getGCell(), getRoutingPads()[0], contact1, contact2, NoFlags );
contact2 = AutoContactHTee::create( getGCell(), getNet(), viaLayer1 );
AutoSegment::create( contact1, contact2, Flags::Vertical );
setBothCornerContacts( contact2 );
} else {
doRp_AutoContacts( getGCell(), getRoutingPads()[0], contact1, contact2, NoFlags );
setSouthWestContact( contact1 );
doRp_AutoContacts( getGCell(), getRoutingPads()[0], contact1, contact2, NoFlags );
setNorthEastContact( contact1 );
}
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
} else {
AutoContact* contact1 = doRp_Access( getGCell(), getRoutingPads()[0], HAccess );
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
AutoContact* contact2 = AutoContactTurn::create( getGCell(), getNet(), viaLayer1 );
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
AutoSegment::create( contact1, contact2, Flags::Horizontal );
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
contact1 = AutoContactVTee::create( getGCell(), getNet(), viaLayer1 );
AutoSegment::create( contact1, contact2, Flags::Vertical );
setBothCornerContacts( contact1 );
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
}
} else if (getConnexity().fields.globals == 3) {
if (not west() or not east()) {
AutoContact* contact1 = doRp_Access( getGCell(), getRoutingPads()[0], HAccess );
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
AutoContact* contact2 = AutoContactVTee::create( getGCell(), getNet(), viaLayer1 );
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
AutoSegment::create( contact1, contact2, Flags::Horizontal );
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
AutoContact* contact3 = AutoContactVTee::create( getGCell(), getNet(), viaLayer1 );
AutoSegment::create( contact2, contact3, Flags::Vertical );
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
setSouthWestContact( (east()) ? contact2 : contact3 );
setNorthEastContact( (east()) ? contact3 : contact2 );
} else {
AutoContact* contact1 = doRp_Access( getGCell(), getRoutingPads()[0], NoFlags );
AutoContact* contact2 = AutoContactVTee::create( getGCell(), getNet(), viaLayer1 );
AutoSegment::create( contact1, contact2, Flags::Vertical );
AutoContact* contact3 = AutoContactVTee::create( getGCell(), getNet(), viaLayer1 );
AutoSegment::create( contact2, contact3, Flags::Vertical );
setSouthWestContact( (north()) ? contact2 : contact3 );
setNorthEastContact( (north()) ? contact3 : contact2 );
}
} else {
AutoContact* contact1 = doRp_Access( getGCell(), getRoutingPads()[0], HAccess );
AutoContact* contact2 = AutoContactVTee::create( getGCell(), getNet(), viaLayer1 );
AutoSegment::create( contact1, contact2, Flags::Horizontal );
AutoContact* contact3 = AutoContactVTee::create( getGCell(), getNet(), viaLayer1 );
AutoSegment::create( contact2, contact3, Flags::Vertical );
AutoContact* contact4 = AutoContactVTee::create( getGCell(), getNet(), viaLayer1 );
AutoSegment::create( contact2, contact4, Flags::Vertical );
setSouthWestContact( contact3 );
setNorthEastContact( contact4 );
}
cdebug_tabw(145,-1);
return true;
}
Complete the modularization of the detail routing converter. * New: In Anabatic::NetBuilder, set all the attributes as private and create accessors and mutators. Finish virtualising all GCell build methods and transfer them into the NetBuilderHV class. Build methods now return a boolean to tell if the GCell was processed or not, to allow cascading in the "big switch". Reorganise the "big switch" in separate sections only partially cascading. Truly fuse the *big switch* for channel routing and over-the-cells. Create a new method "_do_globalSegment()" to delagate the drawing of global segments between two GCell to the derived classes.
2017-12-18 11:15:14 -06:00
bool NetBuilderHV::_do_xG_1M1_1M2 ()
{
cdebug_log(145,1) << getTypeName() << "::_do_xG_1M1_1M2() [Managed Configuration]" << endl;
Component* rpL1;
Component* rpL2;
if (getRoutingPads()[0]->getLayer() == Session::getRoutingLayer(0)) {
rpL1 = getRoutingPads()[0];
rpL2 = getRoutingPads()[1];
} else {
rpL1 = getRoutingPads()[1];
rpL2 = getRoutingPads()[0];
}
cdebug_log(145,0) << "rpL1 := " << rpL1 << endl;
cdebug_log(145,0) << "rpL2 := " << rpL2 << endl;
AutoContact* rpL1ContactSource = NULL;
AutoContact* rpL1ContactTarget = NULL;
AutoContact* rpL2ContactSource = NULL;
AutoContact* rpL2ContactTarget = NULL;
doRp_AutoContacts( getGCell(), rpL1, rpL1ContactSource, rpL1ContactTarget, NoFlags );
doRp_AutoContacts( getGCell(), rpL2, rpL2ContactSource, rpL2ContactTarget, NoFlags );
const Layer* viaLayer1 = Session::getContactLayer(1);
const Layer* viaLayer2 = Session::getContactLayer(2);
AutoContact* subContact = AutoContactTurn::create( getGCell(), getNet(), viaLayer1 );
AutoSegment::create( rpL1ContactSource, subContact, Flags::Horizontal );
AutoSegment::create( rpL2ContactSource, subContact, Flags::Vertical );
if (south() or west()) {
doRp_AutoContacts( getGCell(), rpL2, rpL2ContactSource, rpL2ContactTarget, DoSourceContact );
if (south() and west()) {
setSouthWestContact( AutoContactHTee::create( getGCell(), getNet(), viaLayer2 ) );
AutoSegment::create( rpL2ContactSource, getSouthWestContact(), Flags::Horizontal );
} else {
if (south()) {
setSouthWestContact( AutoContactTurn::create( getGCell(), getNet(), viaLayer2 ) );
AutoSegment::create( rpL2ContactSource, getSouthWestContact(), Flags::Horizontal );
} else {
setSouthWestContact( rpL2ContactSource );
}
}
}
if (north() or east()) {
doRp_AutoContacts( getGCell(), rpL2, rpL2ContactSource, rpL2ContactTarget, DoTargetContact );
if (north() and east()) {
setNorthEastContact( AutoContactHTee::create( getGCell(), getNet(), viaLayer2 ) );
AutoSegment::create( rpL2ContactTarget, getNorthEastContact(), Flags::Horizontal );
} else {
if (north()) {
setNorthEastContact( AutoContactTurn::create( getGCell(), getNet(), viaLayer2 ) );
AutoSegment::create( rpL2ContactTarget, getNorthEastContact(), Flags::Horizontal );
} else {
setNorthEastContact( rpL2ContactTarget );
}
}
}
cdebug_tabw(145,-1);
return true;
}
bool NetBuilderHV::_do_xG_xM1_xM3 ()
{
cdebug_log(145,1) << getTypeName()
<< "::_do_xG_" << (int)getConnexity().fields.M1
<< "M1_" << (int)getConnexity().fields.M3
<< "M3() [G:" << (int)getConnexity().fields.globals << " Managed Configuration]" << endl;
cdebug_log(145,0) << "getConnexity(): " << getConnexity().connexity << endl;
cdebug_log(145,0) << "north: " << north() << endl;
cdebug_log(145,0) << "south: " << south() << endl;
cdebug_log(145,0) << "east: " << east() << endl;
cdebug_log(145,0) << "west: " << west() << endl;
Component* rpM3 = NULL;
if (getRoutingPads()[0]->getLayer() == Session::getRoutingLayer(2))
rpM3 = getRoutingPads()[0];
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
sortRpByX( getRoutingPads(), NoFlags ); // increasing X.
Complete the modularization of the detail routing converter. * New: In Anabatic::NetBuilder, set all the attributes as private and create accessors and mutators. Finish virtualising all GCell build methods and transfer them into the NetBuilderHV class. Build methods now return a boolean to tell if the GCell was processed or not, to allow cascading in the "big switch". Reorganise the "big switch" in separate sections only partially cascading. Truly fuse the *big switch* for channel routing and over-the-cells. Create a new method "_do_globalSegment()" to delagate the drawing of global segments between two GCell to the derived classes.
2017-12-18 11:15:14 -06:00
for ( size_t i=1 ; i<getRoutingPads().size() ; ++i ) {
AutoContact* leftContact = doRp_Access( getGCell(), getRoutingPads()[i-1], HAccess );
AutoContact* rightContact = doRp_Access( getGCell(), getRoutingPads()[i ], HAccess );
AutoSegment::create( leftContact, rightContact, Flags::Horizontal );
if (not rpM3 and (getRoutingPads()[i]->getLayer() == Session::getRoutingLayer(2)))
rpM3 = getRoutingPads()[i];
}
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* viaLayer1 = Session::getContactLayer(1);
AutoContact* subContact1 = NULL;
Complete the modularization of the detail routing converter. * New: In Anabatic::NetBuilder, set all the attributes as private and create accessors and mutators. Finish virtualising all GCell build methods and transfer them into the NetBuilderHV class. Build methods now return a boolean to tell if the GCell was processed or not, to allow cascading in the "big switch". Reorganise the "big switch" in separate sections only partially cascading. Truly fuse the *big switch* for channel routing and over-the-cells. Create a new method "_do_globalSegment()" to delagate the drawing of global segments between two GCell to the derived classes.
2017-12-18 11:15:14 -06:00
if (rpM3) {
// At least one M3 RoutingPad is present: use it.
if (west() and not south()) {
setSouthWestContact( doRp_Access( getGCell(), getRoutingPads()[0], HAccess ) );
} else if (not west() and south()) {
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
doRp_AutoContacts( getGCell(), rpM3, getSouthWestContact(), subContact1, DoSourceContact );
Complete the modularization of the detail routing converter. * New: In Anabatic::NetBuilder, set all the attributes as private and create accessors and mutators. Finish virtualising all GCell build methods and transfer them into the NetBuilderHV class. Build methods now return a boolean to tell if the GCell was processed or not, to allow cascading in the "big switch". Reorganise the "big switch" in separate sections only partially cascading. Truly fuse the *big switch* for channel routing and over-the-cells. Create a new method "_do_globalSegment()" to delagate the drawing of global segments between two GCell to the derived classes.
2017-12-18 11:15:14 -06:00
} else if (west() and south()) {
AutoContact* rpContact = NULL;
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
doRp_AutoContacts( getGCell(), rpM3, rpContact, subContact1, DoSourceContact );
Complete the modularization of the detail routing converter. * New: In Anabatic::NetBuilder, set all the attributes as private and create accessors and mutators. Finish virtualising all GCell build methods and transfer them into the NetBuilderHV class. Build methods now return a boolean to tell if the GCell was processed or not, to allow cascading in the "big switch". Reorganise the "big switch" in separate sections only partially cascading. Truly fuse the *big switch* for channel routing and over-the-cells. Create a new method "_do_globalSegment()" to delagate the drawing of global segments between two GCell to the derived classes.
2017-12-18 11:15:14 -06:00
setSouthWestContact( AutoContactVTee::create( getGCell(), getNet(), viaLayer1 ) );
AutoSegment::create( rpContact, getSouthWestContact(), Flags::Vertical );
}
if (east() and not north()) {
setNorthEastContact( doRp_Access( getGCell(), getRoutingPads()[getRoutingPads().size()-1], HAccess ) );
} else if (not east() and north()) {
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
doRp_AutoContacts( getGCell(), rpM3, subContact1, getNorthEastContact(), DoTargetContact );
Complete the modularization of the detail routing converter. * New: In Anabatic::NetBuilder, set all the attributes as private and create accessors and mutators. Finish virtualising all GCell build methods and transfer them into the NetBuilderHV class. Build methods now return a boolean to tell if the GCell was processed or not, to allow cascading in the "big switch". Reorganise the "big switch" in separate sections only partially cascading. Truly fuse the *big switch* for channel routing and over-the-cells. Create a new method "_do_globalSegment()" to delagate the drawing of global segments between two GCell to the derived classes.
2017-12-18 11:15:14 -06:00
} else if (east() and north()) {
AutoContact* rpContact = NULL;
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
doRp_AutoContacts( getGCell(), rpM3, subContact1, rpContact, DoTargetContact );
Complete the modularization of the detail routing converter. * New: In Anabatic::NetBuilder, set all the attributes as private and create accessors and mutators. Finish virtualising all GCell build methods and transfer them into the NetBuilderHV class. Build methods now return a boolean to tell if the GCell was processed or not, to allow cascading in the "big switch". Reorganise the "big switch" in separate sections only partially cascading. Truly fuse the *big switch* for channel routing and over-the-cells. Create a new method "_do_globalSegment()" to delagate the drawing of global segments between two GCell to the derived classes.
2017-12-18 11:15:14 -06:00
setNorthEastContact( AutoContactVTee::create( getGCell(), getNet(), viaLayer1 ) );
AutoSegment::create( rpContact, getNorthEastContact(), Flags::Vertical );
}
} 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
cdebug_log(145,0) << "getRoutingPads().size():" << getRoutingPads().size()<< endl;
if (getRoutingPads().size() == 1) {
if ( ((east () != NULL) xor (west () != NULL))
and ((north() != NULL) xor (south() != NULL)) ) {
cdebug_log(145,0) << "case: One M1, with global turn." << endl;
AutoContact* rpContact = doRp_Access( getGCell(), getRoutingPads()[0], HAccess );
subContact1 = AutoContactTurn::create( getGCell(), getNet(), viaLayer1 );
AutoSegment::create( rpContact, subContact1, Flags::Horizontal );
rpContact = subContact1;
subContact1 = AutoContactVTee::create( getGCell(), getNet(), viaLayer1 );
AutoSegment::create( rpContact, subContact1, Flags::Vertical );
setBothCornerContacts( subContact1 );
cdebug_tabw(145,-1);
return true;
}
}
Complete the modularization of the detail routing converter. * New: In Anabatic::NetBuilder, set all the attributes as private and create accessors and mutators. Finish virtualising all GCell build methods and transfer them into the NetBuilderHV class. Build methods now return a boolean to tell if the GCell was processed or not, to allow cascading in the "big switch". Reorganise the "big switch" in separate sections only partially cascading. Truly fuse the *big switch* for channel routing and over-the-cells. Create a new method "_do_globalSegment()" to delagate the drawing of global segments between two GCell to the derived classes.
2017-12-18 11:15:14 -06:00
// All RoutingPad are M1.
Component* southWestRp = getRoutingPads()[0];
cdebug_log(145,0) << "| Initial S-W Global RP: " << southWestRp << 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 ( size_t i=1 ; i<getRoutingPads().size() ; ++i ) {
// if (southWestRp->getBoundingBox().getHeight() >= 4*Session::getPitch(1)) break;
// if (getRoutingPads()[i]->getBoundingBox().getHeight() > southWestRp->getBoundingBox().getHeight()) {
// southWestRp = getRoutingPads()[i];
// cdebug_log(145,0) << "| Better RP: " << southWestRp << endl;
// }
// }
Complete the modularization of the detail routing converter. * New: In Anabatic::NetBuilder, set all the attributes as private and create accessors and mutators. Finish virtualising all GCell build methods and transfer them into the NetBuilderHV class. Build methods now return a boolean to tell if the GCell was processed or not, to allow cascading in the "big switch". Reorganise the "big switch" in separate sections only partially cascading. Truly fuse the *big switch* for channel routing and over-the-cells. Create a new method "_do_globalSegment()" to delagate the drawing of global segments between two GCell to the derived classes.
2017-12-18 11:15:14 -06:00
if (west() and not south()) {
setSouthWestContact( doRp_Access( getGCell(), southWestRp, HAccess ) );
} else if (not west() and south()) {
AutoContact* rpContact = doRp_Access( getGCell(), southWestRp, HAccess );
setSouthWestContact( AutoContactTurn::create( getGCell(), getNet(), viaLayer1 ) );
AutoSegment::create( rpContact, getSouthWestContact(), Flags::Horizontal );
} else if (west() and south()) {
AutoContact* rpContact = doRp_Access( getGCell(), southWestRp, HAccess );
setSouthWestContact( AutoContactHTee::create( getGCell(), getNet(), viaLayer1 ) );
AutoSegment::create( rpContact, getSouthWestContact(), Flags::Horizontal );
}
Component* northEastRp = getRoutingPads()[getRoutingPads().size()-1];
cdebug_log(145,0) << "| Initial N-E Global RP: " << northEastRp << endl;
if (getRoutingPads().size() > 1) {
for ( size_t i=getRoutingPads().size()-1; i != 0 ; ) {
i -= 1;
if (northEastRp->getBoundingBox().getHeight() >= 4*Session::getPitch(1)) break;
if (getRoutingPads()[i]->getBoundingBox().getHeight() > northEastRp->getBoundingBox().getHeight()) {
cdebug_log(145,0) << "| Better RP: " << northEastRp << endl;
northEastRp = getRoutingPads()[i];
}
}
}
if (east() and not north()) {
setNorthEastContact( doRp_Access( getGCell(), northEastRp, HAccess ) );
} else if (not east() and north()) {
AutoContact* rpContact = doRp_Access( getGCell(), northEastRp, HAccess );
setNorthEastContact( AutoContactTurn::create( getGCell(), getNet(), viaLayer1 ) );
AutoSegment::create( rpContact, getNorthEastContact(), Flags::Horizontal );
} else if (east() and north()) {
AutoContact* rpContact = doRp_Access( getGCell(), northEastRp, HAccess );
setNorthEastContact( AutoContactHTee::create( getGCell(), getNet(), viaLayer1 ) );
AutoSegment::create( rpContact, getNorthEastContact(), Flags::Horizontal );
}
}
cdebug_tabw(145,-1);
return true;
}
bool NetBuilderHV::_do_4G_1M2 ()
{
cdebug_log(145,1) << getTypeName() << "::_do_4G_1M2() [Managed Configuration]" << endl;
Component* rpL2 = getRoutingPads()[0];
cdebug_log(145,0) << "rpL2 := " << rpL2 << endl;
AutoContact* rpL2ContactSource = NULL;
AutoContact* rpL2ContactTarget = NULL;
doRp_AutoContacts( getGCell(), rpL2, rpL2ContactSource, rpL2ContactTarget, DoSourceContact|DoTargetContact );
const Layer* viaLayer2 = Session::getContactLayer(2);
setSouthWestContact( AutoContactHTee::create( getGCell(), getNet(), viaLayer2 ) );
setNorthEastContact( AutoContactHTee::create( getGCell(), getNet(), viaLayer2 ) );
AutoSegment::create( getSouthWestContact(), rpL2ContactSource, Flags::Horizontal );
AutoSegment::create( rpL2ContactTarget, getNorthEastContact(), Flags::Horizontal );
cdebug_tabw(145,-1);
return true;
}
bool NetBuilderHV::_do_xG_xM2 ()
{
cdebug_log(145,1) << getTypeName() << "::_do_"
<< (int)getConnexity().fields.globals << "G_"
<< (int)getConnexity().fields.M2 << "M2() [Managed Configuration - x]" << endl;
Component* biggestRp = getRoutingPads()[0];
for ( size_t i=1 ; i<getRoutingPads().size() ; ++i ) {
doRp_StairCaseH( getGCell(), getRoutingPads()[i-1], getRoutingPads()[i] );
if (getRoutingPads()[i]->getBoundingBox().getWidth() > biggestRp->getBoundingBox().getWidth())
biggestRp = getRoutingPads()[i];
}
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* viaLayer1 = Session::getContactLayer(1);
if (east() and west() and not south() and not north()) {
AutoContact* rpContact = doRp_Access( getGCell(), biggestRp, HBothAccess );
setBothCornerContacts( rpContact );
cdebug_tabw(145,-1);
return true;
}
Complete the modularization of the detail routing converter. * New: In Anabatic::NetBuilder, set all the attributes as private and create accessors and mutators. Finish virtualising all GCell build methods and transfer them into the NetBuilderHV class. Build methods now return a boolean to tell if the GCell was processed or not, to allow cascading in the "big switch". Reorganise the "big switch" in separate sections only partially cascading. Truly fuse the *big switch* for channel routing and over-the-cells. Create a new method "_do_globalSegment()" to delagate the drawing of global segments between two GCell to the derived classes.
2017-12-18 11:15:14 -06:00
if (west() and not south()) {
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
setSouthWestContact( doRp_Access( getGCell(), getRoutingPads()[0], HAccess ) );
Complete the modularization of the detail routing converter. * New: In Anabatic::NetBuilder, set all the attributes as private and create accessors and mutators. Finish virtualising all GCell build methods and transfer them into the NetBuilderHV class. Build methods now return a boolean to tell if the GCell was processed or not, to allow cascading in the "big switch". Reorganise the "big switch" in separate sections only partially cascading. Truly fuse the *big switch* for channel routing and over-the-cells. Create a new method "_do_globalSegment()" to delagate the drawing of global segments between two GCell to the derived classes.
2017-12-18 11:15:14 -06:00
} else if (not west() and south()) {
setSouthWestContact( doRp_Access( getGCell(), biggestRp, NoFlags ) );
} else if (west() and south()) {
AutoContact* rpContact = doRp_Access( getGCell(), biggestRp, NoFlags );
setSouthWestContact( AutoContactVTee::create( getGCell(), getNet(), viaLayer1 ) );
AutoSegment::create( rpContact, getSouthWestContact(), Flags::Vertical );
}
if (east() and not north()) {
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
setNorthEastContact( doRp_Access( getGCell(), getRoutingPads()[getRoutingPads().size()-1], HAccess ) );
Complete the modularization of the detail routing converter. * New: In Anabatic::NetBuilder, set all the attributes as private and create accessors and mutators. Finish virtualising all GCell build methods and transfer them into the NetBuilderHV class. Build methods now return a boolean to tell if the GCell was processed or not, to allow cascading in the "big switch". Reorganise the "big switch" in separate sections only partially cascading. Truly fuse the *big switch* for channel routing and over-the-cells. Create a new method "_do_globalSegment()" to delagate the drawing of global segments between two GCell to the derived classes.
2017-12-18 11:15:14 -06:00
} else if (not east() and north()) {
setNorthEastContact( doRp_Access( getGCell(), biggestRp, NoFlags ) );
} else if (east() and north()) {
AutoContact* rpContact = doRp_Access( getGCell(), biggestRp, NoFlags );
setNorthEastContact( AutoContactVTee::create( getGCell(), getNet(), viaLayer1 ) );
AutoSegment::create( rpContact, getNorthEastContact(), Flags::Vertical );
}
cdebug_tabw(145,-1);
return true;
}
bool NetBuilderHV::_do_1G_1M3 ()
{
cdebug_log(145,1) << getTypeName() << "::_do_1G_1M3() [Optimised Configuration]" << endl;
uint64_t flags = (east() or west()) ? HAccess : NoFlags;
flags |= (north()) ? DoTargetContact : NoFlags;
flags |= (south()) ? DoSourceContact : NoFlags;
doRp_AutoContacts( getGCell()
, getRoutingPads()[0]
, getSouthWestContact()
, getNorthEastContact()
, flags
);
if (not getSouthWestContact()) setSouthWestContact( getNorthEastContact() );
if (not getNorthEastContact()) setNorthEastContact( getSouthWestContact() );
cdebug_log(145,0) << "_southWest: " << getSouthWestContact() << endl;
cdebug_log(145,0) << "_northEast: " << getNorthEastContact() << endl;
const Layer* viaLayer1 = Session::getContactLayer(1);
Partial support for offgrid METAL3 RoutingPad in Anabatic::NetBuilderHV. * New: In Anabatic::NetBuilder::_do_1G_1M3(), RoutingPad in METAL3 from blocks are most likely to be offgrid in real mode, we must account for that case. When an offgrid METAL3 is found, a strap of METAL2 is added, as it may be less than one pitch, it will be reduced into METAL3 often. * New: In Anabatic::NetBuilder::doRp_Access(), Support for offgrid METAL2 is added but not enabled yet. Seems to need more polishing.
2021-02-17 16:34:17 -06:00
Box cellAb = getAnabatic()->getCell()->getAbutmentBox();
RoutingLayerGauge* lgM3 = Session::getLayerGauge( 2 );
DbU::Unit trackAxis = lgM3->getTrackPosition( cellAb.getXMin()
, cellAb.getXMax()
, getRoutingPads()[0]->getX()
, Constant::Nearest );
bool offGrid = (trackAxis != getRoutingPads()[0]->getX());
if (offGrid) {
cdebug_log(145,0) << "Off grid, Misaligned M3, add horizontal strap" << endl;
AutoContactTurn* turn1 = AutoContactTurn::create( getGCell(), getNet(), viaLayer1 );
AutoSegment::create( getSouthWestContact(), turn1, Flags::Horizontal );
setBothCornerContacts( turn1 );
}
Complete the modularization of the detail routing converter. * New: In Anabatic::NetBuilder, set all the attributes as private and create accessors and mutators. Finish virtualising all GCell build methods and transfer them into the NetBuilderHV class. Build methods now return a boolean to tell if the GCell was processed or not, to allow cascading in the "big switch". Reorganise the "big switch" in separate sections only partially cascading. Truly fuse the *big switch* for channel routing and over-the-cells. Create a new method "_do_globalSegment()" to delagate the drawing of global segments between two GCell to the derived classes.
2017-12-18 11:15:14 -06:00
if (flags & HAccess) {
// HARDCODED VALUE.
if (getRoutingPads()[0]->getBoundingBox().getHeight() < 3*Session::getPitch(1)) {
AutoContact* subContact = AutoContactTurn::create( getGCell(), getNet(), viaLayer1 );
AutoSegment::create( getSouthWestContact(), subContact, Flags::Vertical );
setBothCornerContacts( subContact );
}
} else {
Partial support for offgrid METAL3 RoutingPad in Anabatic::NetBuilderHV. * New: In Anabatic::NetBuilder::_do_1G_1M3(), RoutingPad in METAL3 from blocks are most likely to be offgrid in real mode, we must account for that case. When an offgrid METAL3 is found, a strap of METAL2 is added, as it may be less than one pitch, it will be reduced into METAL3 often. * New: In Anabatic::NetBuilder::doRp_Access(), Support for offgrid METAL2 is added but not enabled yet. Seems to need more polishing.
2021-02-17 16:34:17 -06:00
if (getSourceContact() and (getSourceContact()->getX() != getSouthWestContact()->getX())) {
cdebug_log(145,0) << "On grid, Misaligned M3, add dogleg" << endl;
AutoContactTurn* turn1 = AutoContactTurn::create( getGCell(), getNet(), viaLayer1 );
AutoContactTurn* turn2 = AutoContactTurn::create( getGCell(), getNet(), viaLayer1 );
AutoSegment* v1 = AutoSegment::create( getSouthWestContact(), turn1, Flags::Vertical );
AutoSegment::create( turn1, turn2, Flags::Horizontal );
v1->setAxis( getSouthWestContact()->getX(), Flags::Force );
v1->setFlags( AutoSegment::SegFixed|AutoSegment::SegFixedAxis );
setBothCornerContacts( turn2 );
Complete the modularization of the detail routing converter. * New: In Anabatic::NetBuilder, set all the attributes as private and create accessors and mutators. Finish virtualising all GCell build methods and transfer them into the NetBuilderHV class. Build methods now return a boolean to tell if the GCell was processed or not, to allow cascading in the "big switch". Reorganise the "big switch" in separate sections only partially cascading. Truly fuse the *big switch* for channel routing and over-the-cells. Create a new method "_do_globalSegment()" to delagate the drawing of global segments between two GCell to the derived classes.
2017-12-18 11:15:14 -06:00
}
}
cdebug_tabw(145,-1);
return true;
}
bool NetBuilderHV::_do_xG_xM3 ()
{
cdebug_log(145,1) << getTypeName()
<< "::_do_xG_" << (int)getConnexity().fields.M3
<< "M3() [Managed Configuration]" << endl;
cdebug_log(145,0) << "west:" << west() << endl;
cdebug_log(145,0) << "east:" << east() << endl;
cdebug_log(145,0) << "south:" << south() << endl;
cdebug_log(145,0) << "north:" << north() << 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
sortRpByY( getRoutingPads(), NoFlags ); // increasing Y.
Complete the modularization of the detail routing converter. * New: In Anabatic::NetBuilder, set all the attributes as private and create accessors and mutators. Finish virtualising all GCell build methods and transfer them into the NetBuilderHV class. Build methods now return a boolean to tell if the GCell was processed or not, to allow cascading in the "big switch". Reorganise the "big switch" in separate sections only partially cascading. Truly fuse the *big switch* for channel routing and over-the-cells. Create a new method "_do_globalSegment()" to delagate the drawing of global segments between two GCell to the derived classes.
2017-12-18 11:15:14 -06:00
for ( size_t i=1 ; i<getRoutingPads().size() ; i++ ) {
doRp_StairCaseV( getGCell(), getRoutingPads()[i-1], getRoutingPads()[i] );
}
const Layer* viaLayer1 = Session::getContactLayer(1);
AutoContact* unusedContact = NULL;
Component* rp = getRoutingPads()[0];
if (west() and not south()) {
setSouthWestContact( doRp_Access( getGCell(), rp, HAccess ) );
} else if (not west() and south()) {
doRp_AutoContacts( getGCell(), rp, getSouthWestContact(), unusedContact, DoSourceContact );
if (getSourceContact()) {
if (getSourceContact()->getX() != getSouthWestContact()->getX()) {
cdebug_log(149,0) << "Misaligned South: _source:" << DbU::getValueString(getSourceContact()->getX())
<< "_southWest:" << DbU::getValueString(getSouthWestContact()->getX()) << endl;
AutoContactTurn* turn1 = AutoContactTurn::create( getGCell(), getNet(), viaLayer1 );
AutoContactTurn* turn2 = AutoContactTurn::create( getGCell(), getNet(), viaLayer1 );
AutoSegment::create( getSouthWestContact(), turn1, Flags::Vertical );
AutoSegment::create( turn1 , turn2, Flags::Horizontal );
setSouthWestContact( turn2 );
}
}
} else if (west() and south()) {
AutoContact* rpContact = NULL;
doRp_AutoContacts( getGCell(), rp, rpContact, unusedContact, DoSourceContact );
setSouthWestContact( AutoContactVTee::create( getGCell(), getNet(), viaLayer1 ) );
AutoSegment::create( rpContact, getSouthWestContact(), Flags::Vertical );
}
rp = getRoutingPads()[getRoutingPads().size()-1];
if (east() and not north()) {
setNorthEastContact( doRp_Access( getGCell(), rp, HAccess ) );
} else if (not east() and north()) {
doRp_AutoContacts( getGCell(), rp, unusedContact, getNorthEastContact(), DoTargetContact );
if (getSourceContact()) {
if (getSourceContact()->getX() != getNorthEastContact()->getX()) {
cdebug_log(149,0) << "Misaligned North: _source:" << DbU::getValueString(getSourceContact()->getX())
<< "_southWest:" << DbU::getValueString(getNorthEastContact()->getX()) << endl;
AutoContactTurn* turn1 = AutoContactTurn::create( getGCell(), getNet(), viaLayer1 );
AutoContactTurn* turn2 = AutoContactTurn::create( getGCell(), getNet(), viaLayer1 );
AutoSegment::create( getNorthEastContact(), turn1, Flags::Vertical );
AutoSegment::create( turn1 , turn2, Flags::Horizontal );
setNorthEastContact( turn2 );
}
}
} else if (east() and north()) {
AutoContact* rpContact = NULL;
doRp_AutoContacts( getGCell(), rp, unusedContact, rpContact, DoTargetContact );
setNorthEastContact( AutoContactVTee::create( getGCell(), getNet(), viaLayer1 ) );
AutoSegment::create( rpContact, getNorthEastContact(), Flags::Vertical );
}
cdebug_tabw(145,-1);
return true;
}
bool NetBuilderHV::_do_globalSegment ()
{
cdebug_log(145,1) << getTypeName() << "::_do_globalSegment()" << endl;
if (getSourceContact()) {
AutoContact* targetContact
= ( getSegmentHookType(getFromHook()) & (NorthBound|EastBound) )
? getNorthEastContact() : getSouthWestContact() ;
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 (not getFromHook()) cerr << "getFromHook() is NULL !" << endl;
Prevent wires part of a diode cluster to be moved up. Protections diodes may not be able to play their role if they are separated from their cluster by upper level layers (METAL4/METAL5). This seems not to diminish the total number of diodes. * New: In Anabatic::NetData, add a set of non move up segments in the object. To tag global wires that are part of a cluster. * New: In Anabatic::AutoSegment, add support for a SegNoMoveUp flag. This flag is propagated through _makeDogleg() . Used in ::canMoveUp() and ::canPivotUp(). * New: In Anabatic::NetBuilder, add NetData to the attributes so we can extract the NoMoveUp infomation given by the antenna protect stage. * New: In NetBuilderHV::_do_globalSegment(), lookup NoMoveUp information from NetData to put it in AutoSegment. * New: In AntennaProtect(Net*), flags the RoutingPad clusters wires as non movable up. * New: In AnabaticEngine::breatAt(GCell*), propagate the SegNoMoveUp flags. Based on NetData.
2021-06-18 12:17:24 -05:00
Segment* baseSegment = static_cast<Segment*>( getFromHook()->getComponent() );
Complete the modularization of the detail routing converter. * New: In Anabatic::NetBuilder, set all the attributes as private and create accessors and mutators. Finish virtualising all GCell build methods and transfer them into the NetBuilderHV class. Build methods now return a boolean to tell if the GCell was processed or not, to allow cascading in the "big switch". Reorganise the "big switch" in separate sections only partially cascading. Truly fuse the *big switch* for channel routing and over-the-cells. Create a new method "_do_globalSegment()" to delagate the drawing of global segments between two GCell to the derived classes.
2017-12-18 11:15:14 -06:00
AutoSegment* globalSegment = AutoSegment::create( getSourceContact()
, targetContact
Prevent wires part of a diode cluster to be moved up. Protections diodes may not be able to play their role if they are separated from their cluster by upper level layers (METAL4/METAL5). This seems not to diminish the total number of diodes. * New: In Anabatic::NetData, add a set of non move up segments in the object. To tag global wires that are part of a cluster. * New: In Anabatic::AutoSegment, add support for a SegNoMoveUp flag. This flag is propagated through _makeDogleg() . Used in ::canMoveUp() and ::canPivotUp(). * New: In Anabatic::NetBuilder, add NetData to the attributes so we can extract the NoMoveUp infomation given by the antenna protect stage. * New: In NetBuilderHV::_do_globalSegment(), lookup NoMoveUp information from NetData to put it in AutoSegment. * New: In AntennaProtect(Net*), flags the RoutingPad clusters wires as non movable up. * New: In AnabaticEngine::breatAt(GCell*), propagate the SegNoMoveUp flags. Based on NetData.
2021-06-18 12:17:24 -05:00
, baseSegment
Complete the modularization of the detail routing converter. * New: In Anabatic::NetBuilder, set all the attributes as private and create accessors and mutators. Finish virtualising all GCell build methods and transfer them into the NetBuilderHV class. Build methods now return a boolean to tell if the GCell was processed or not, to allow cascading in the "big switch". Reorganise the "big switch" in separate sections only partially cascading. Truly fuse the *big switch* for channel routing and over-the-cells. Create a new method "_do_globalSegment()" to delagate the drawing of global segments between two GCell to the derived classes.
2017-12-18 11:15:14 -06:00
);
globalSegment->setFlags( (getDegree() == 2) ? AutoSegment::SegBipoint : 0 );
Prevent wires part of a diode cluster to be moved up. Protections diodes may not be able to play their role if they are separated from their cluster by upper level layers (METAL4/METAL5). This seems not to diminish the total number of diodes. * New: In Anabatic::NetData, add a set of non move up segments in the object. To tag global wires that are part of a cluster. * New: In Anabatic::AutoSegment, add support for a SegNoMoveUp flag. This flag is propagated through _makeDogleg() . Used in ::canMoveUp() and ::canPivotUp(). * New: In Anabatic::NetBuilder, add NetData to the attributes so we can extract the NoMoveUp infomation given by the antenna protect stage. * New: In NetBuilderHV::_do_globalSegment(), lookup NoMoveUp information from NetData to put it in AutoSegment. * New: In AntennaProtect(Net*), flags the RoutingPad clusters wires as non movable up. * New: In AnabaticEngine::breatAt(GCell*), propagate the SegNoMoveUp flags. Based on NetData.
2021-06-18 12:17:24 -05:00
if (getNetData() and getNetData()->isNoMoveUp(baseSegment)) {
globalSegment->setFlags( AutoSegment::SegNoMoveUp );
}
Complete the modularization of the detail routing converter. * New: In Anabatic::NetBuilder, set all the attributes as private and create accessors and mutators. Finish virtualising all GCell build methods and transfer them into the NetBuilderHV class. Build methods now return a boolean to tell if the GCell was processed or not, to allow cascading in the "big switch". Reorganise the "big switch" in separate sections only partially cascading. Truly fuse the *big switch* for channel routing and over-the-cells. Create a new method "_do_globalSegment()" to delagate the drawing of global segments between two GCell to the derived classes.
2017-12-18 11:15:14 -06:00
cdebug_log(145,0) << "Create global segment: " << globalSegment << endl;
// HARDCODED VALUE.
Rewrite support for minimum area metal wires (stacked VIAs). We use the segments extensions (dxTarget & dyTarget) to enlarge if needed the segments. This new implementation is completely at Anabatic level and should not be seen (i.e. managed) at Katana level. * Change: In AutoHorizontal & AutoVertical, change the semantic of getSourceU() and getTargetU(). formerly they where the end position of the segment (with extension included), now they gives the position of the anchor contacts, that is the axis of the perpandiculars. * New: AutoSegment::getLength() is still a proxy toward the Segment::getLength() which returns the length of segment with dxSource & dxTarget. We introduce a getAnchoredlength() which returns the length between the centers of the S/T anchors. That is axis to axis. This is this length which is now used througout Anabatic. * New: In AutoSegment::_extentionCaps, add a fourth item to hold the segment minimal length (to respect minimal area given the wire width). * New: In AutoSegment::getExtensionCap(), if the segment has a non-zero S/T extension, returns it instead of the S/T contact extension *if it is greated*. The check of the extension can be disabled by the Flag::NoSegExt flag. * Change: In AutoSegment::isMiddleStack(), security check on the presence of source and targets. More accurate detection of perpandicular in "same layer" with a non-zero length, So the area is OK, even with a short segment. * New: AutoSegment::expandToMinLength(), check if a segment is under the minimal length and expand it if need be by playing with the dxTarget & dxSource. Tag minimal segments with the AutoSegment::SegAtMinArea flag. Also try to keep the segment *inside* it's former (supposedly wider) interval. * New: AutoSegment::unexpandToMinLength(), to be called on a formerly minimal sized segment which as grown up. Reset it's S/T extensions to zero and unset the flag SegAtMinArea. * Change: In AutoHorizontal::updateOrient(), when the extension are non-zero, also swap them if needed, to keep the exact footprint of the segment. * New: In AutoSegment::revalidate(), check that the segment respect the minimal length (area), and expand it if needed. Conversely, if the segment has grown up from a minimal length state, reset it's extensions to zero. * Change: In Anabatic::Session::revalidate(), invalidateds segments are now sorted in such a way that the "middle stack" ones are revalidateds last. Not recall 100% why... * New: In Katana::TrackCost, add a computation of the free interval length we are into (if any). Not used yet, keep it for future use. * Change: In KatanaEngine::finalizeLayout(), remove the call to segments minimum area protection. It is now obsoleted by the new implementation in Anabatic. * Change: In Track::check(), call the minimum size/area checker Track::checkMinArea().
2021-04-04 17:01:54 -05:00
if ( (getTopology() & Global_Fixed) and (globalSegment->getAnchoredLength() < 2*Session::getSliceHeight()) )
Complete the modularization of the detail routing converter. * New: In Anabatic::NetBuilder, set all the attributes as private and create accessors and mutators. Finish virtualising all GCell build methods and transfer them into the NetBuilderHV class. Build methods now return a boolean to tell if the GCell was processed or not, to allow cascading in the "big switch". Reorganise the "big switch" in separate sections only partially cascading. Truly fuse the *big switch* for channel routing and over-the-cells. Create a new method "_do_globalSegment()" to delagate the drawing of global segments between two GCell to the derived classes.
2017-12-18 11:15:14 -06:00
addToFixSegments( globalSegment );
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 (getConnexity().fields.globals < 2) {
cdebug_tabw(145,-1);
return false;
}
Complete the modularization of the detail routing converter. * New: In Anabatic::NetBuilder, set all the attributes as private and create accessors and mutators. Finish virtualising all GCell build methods and transfer them into the NetBuilderHV class. Build methods now return a boolean to tell if the GCell was processed or not, to allow cascading in the "big switch". Reorganise the "big switch" in separate sections only partially cascading. Truly fuse the *big switch* for channel routing and over-the-cells. Create a new method "_do_globalSegment()" to delagate the drawing of global segments between two GCell to the derived classes.
2017-12-18 11:15:14 -06:00
} else
setFromHook( NULL );
push( east (), getNorthEastContact() );
push( west (), getSouthWestContact() );
push( north(), getNorthEastContact() );
push( south(), getSouthWestContact() );
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
cdebug_tabw(145,-1);
Complete the modularization of the detail routing converter. * New: In Anabatic::NetBuilder, set all the attributes as private and create accessors and mutators. Finish virtualising all GCell build methods and transfer them into the NetBuilderHV class. Build methods now return a boolean to tell if the GCell was processed or not, to allow cascading in the "big switch". Reorganise the "big switch" in separate sections only partially cascading. Truly fuse the *big switch* for channel routing and over-the-cells. Create a new method "_do_globalSegment()" to delagate the drawing of global segments between two GCell to the derived classes.
2017-12-18 11:15:14 -06:00
return true;
}
First working recursive place & route (Libre-SOC ALU16 benchmark). * New: In Hurricane::Cell::isLeaf(), a leaf cell was defined as one without instances. Now it can be software defined with the "FlattenLeaf" state. If it is flagged "flatten leaf" is will be considered as a leaf by the trans-hierarchical walkthrough. This flag is also set for Cells in the Alliance CATAL. * New: In Hurricane::Cell_LeafInstanceOccurrences::progress(), add a test to prune leaf cells. * Bug: In Hurricane::getComponentOccurrences(), when calling the locator constructor, the "withLeafCells" argument was forgotten and was getting the value of "doExtraction" (wich, fortunately, was the same). * Change: In Hurricane::HypreNet_LeafPlugOccurrences::progress(), prune the leaf cells instead of the terminal ones. * Change: In Hurricane::HypreNet_ComponentOccurrences::progress(), prune the leaf cells instead of the terminal ones. * New: In Hurricane::Cell, add a "useFlattenLeaf" mode to take into account or ignore the "FlattenLeaf" flag when doing trans- hierarchical walkthrough. Not sure if it shouldn't be systematic. * New: In Etesian::toColoquinte() and other, activate the "FlattenLeaf" mode for all walkthough. So we can stop at already placed sub-blocks. * Bug: In Etesian::toColoquinte(), when placed sub-block where present, the number of instances was under estimated, so the instance array was oversized and the end went unitialized. So we were getting all kind of strange behavior from Coloquinte... Now throw an exception if the number of instances differs from the computed size. * Bug: In Etesian::toColoquinte(), as for the instance array, now correctly compute the size with the ecluded nets (supplies, clocks). * Bug: In Etesian::resetPlacement(), event if the Cell is unplaced, go through the instance list to remove the potential feed cells. The netlist is written back after placement *with* the feed cells added. So, when loading again, they where cluttering the netlist and giving feed redefintions warning. They also generated erratic behaviors. * New: In Anabatic::NetBuilderHV, added configurations for: * 1G_1M1_1PinM3 * 1G_xM1_1PinM2 * 2G_xM1_1PinM2 * doRp_AccessNorthPin * New: In Anabatic::NetBuilderHV, added Pin management in SingleGCell. * Bug: In Anabatic::NetBuilderHV::_do_xG_1PinM3(), the North/South global routing configuration was forgotten (only East/West). * Bug: In Katana::TrackFixedSegment, the blockage net was a static variable initialized at the first run. So we were getting stuck with the blockage net of the first cell to be routed. Of course, it did show only when Katana was run multiple times in the same process.
2020-03-03 17:50:18 -06:00
void NetBuilderHV::singleGCell ( AnabaticEngine* anbt, Net* net )
{
cdebug_log(145,1) << "NetBuilderHV::singleGCell() " << net << endl;
vector<RoutingPad*> rpM1s;
Component* rpM2 = NULL;
RoutingPad* rpPin = NULL;
for ( RoutingPad* rp : net->getRoutingPads() ) {
if (dynamic_cast<Pin*>(rp->getOccurrence().getEntity())) {
rpPin = rp;
continue;
}
if (Session::getRoutingGauge()->getLayerDepth(rp->getLayer()) == 1)
rpM2 = rp;
else
rpM1s.push_back( rp );
}
if ( (rpM1s.size() < 2) and not (rpM2 or rpPin) ) {
cerr << Error( "For %s, less than two Plugs/Pins (%d)."
, getString(net).c_str()
, rpM1s.size() ) << endl;
cdebug_tabw(145,-1);
return;
}
if (rpM1s.empty()) {
cerr << Error( "No METAL1 in Single GCell for Net \"%s\".", getString(net->getName()).c_str() ) << endl;
cdebug_tabw(145,-1);
return;
}
sortRpByX( rpM1s, NetBuilder::NoFlags ); // increasing X.
GCell* gcell1 = anbt->getGCellUnder( (*rpM1s.begin ())->getCenter() );
GCell* gcell2 = anbt->getGCellUnder( (*rpM1s.rbegin())->getCenter() );
if (not gcell1) {
cerr << Error( "No GCell under %s.", getString(*(rpM1s.begin())).c_str() ) << endl;
cdebug_tabw(145,-1);
return;
}
if (gcell1 != gcell2) {
cerr << Error( "Not under a single GCell %s.", getString(*(rpM1s.rbegin())).c_str() ) << endl;
cdebug_tabw(145,-1);
return;
}
cdebug_log(145,0) << "singleGCell " << gcell1 << endl;
AutoContact* turn1 = NULL;
AutoContact* turn2 = NULL;
AutoContact* source = NULL;
AutoContact* target = NULL;
for ( size_t irp=1 ; irp<rpM1s.size() ; ++irp ) {
doRp_AutoContacts( gcell1, rpM1s[irp-1], source, turn1, DoSourceContact );
doRp_AutoContacts( gcell1, rpM1s[irp ], target, turn1, DoSourceContact );
if (source->getUConstraints(Flags::Vertical).intersect(target->getUConstraints(Flags::Vertical))) {
uint64_t flags = checkRoutingPadSize( rpM1s[irp-1] );
if ((flags & VSmall) or Session::getConfiguration()->isVH()) {
if (Session::getConfiguration()->isHV()) {
turn1 = AutoContactTurn::create( gcell1, rpM1s[irp]->getNet(), Session::getDContactLayer() );
AutoSegment::create( source, turn1, Flags::Horizontal );
source = turn1;
}
turn1 = AutoContactTurn::create( gcell1, rpM1s[irp]->getNet(), Session::getDContactLayer() );
AutoSegment::create( source, turn1 , Flags::Vertical );
source = turn1;
}
flags = checkRoutingPadSize( rpM1s[irp] );
if ((flags & VSmall) or Session::getConfiguration()->isVH()) {
if (Session::getConfiguration()->isHV()) {
turn1 = AutoContactTurn::create( gcell1, rpM1s[irp]->getNet(), Session::getDContactLayer() );
AutoSegment::create( target, turn1, Flags::Horizontal );
target = turn1;
}
turn1 = AutoContactTurn::create( gcell1, rpM1s[irp]->getNet(), Session::getDContactLayer() );
AutoSegment::create( target, turn1 , Flags::Vertical );
target = turn1;
}
AutoSegment::create( source, target, Flags::Horizontal );
} else {
turn1 = AutoContactTurn::create( gcell1, rpM1s[irp]->getNet(), Session::getDContactLayer() );
turn2 = AutoContactTurn::create( gcell1, rpM1s[irp]->getNet(), Session::getDContactLayer() );
AutoSegment::create( source, turn1 , Flags::Horizontal );
AutoSegment::create( turn1 , turn2 , Flags::Vertical );
AutoSegment::create( turn2 , target, Flags::Horizontal );
}
}
if (rpM2) {
doRp_AutoContacts( gcell1, rpM1s[0], source, turn1, DoSourceContact );
doRp_AutoContacts( gcell1, rpM2 , target, turn1, DoSourceContact );
turn1 = AutoContactTurn::create( gcell1, rpM2->getNet(), Session::getContactLayer(1) );
AutoSegment::create( source, turn1 , Flags::Horizontal );
AutoSegment::create( turn1 , target, Flags::Vertical );
}
if (rpPin) {
Added new configuration in NetBuilderHV (for M3 Pins). * New: In Anabatic::NetBuilderHV(), added configurations to manage one Pin M3 + one M1 terminal plus 2 & 3 globals: * _do_2G_1M1_1PinM3() * _do_3G_1M1_1PinM3() They were occuring for the first time in soclayout/experiment7 in the "flat" approach. * New: In Katana::runNegociate(), mark the newly routed netlist as "NetlistTerminal" so it is not placed and routed *again* when reused as an instance (mostly interract with Etesian).
2020-04-20 05:53:07 -05:00
Pin* pin = dynamic_cast<Pin*>( rpPin->getOccurrence().getEntity() );
Pin::AccessDirection pinDir = pin->getAccessDirection();
if ( (pinDir == Pin::AccessDirection::NORTH)
or (pinDir == Pin::AccessDirection::SOUTH) ) {
doRp_AutoContacts( gcell1, rpM1s[0], source, turn1, DoSourceContact );
target = doRp_AccessNorthPin( gcell1, rpPin );
turn1 = AutoContactTurn::create( gcell1, rpPin->getNet(), Session::getContactLayer(1) );
AutoSegment::create( source, turn1 , Flags::Horizontal );
AutoSegment::create( turn1 , target, Flags::Vertical );
} else {
RoutingPad* closest = NULL;
if (pinDir == Pin::AccessDirection::EAST) closest = rpM1s.back();
else closest = rpM1s.front();
doRp_AutoContacts( gcell1, closest, source, turn1, DoSourceContact );
target = doRp_AccessEastWestPin( gcell1, rpPin );
AutoSegment::create( source, target , Flags::Horizontal );
}
First working recursive place & route (Libre-SOC ALU16 benchmark). * New: In Hurricane::Cell::isLeaf(), a leaf cell was defined as one without instances. Now it can be software defined with the "FlattenLeaf" state. If it is flagged "flatten leaf" is will be considered as a leaf by the trans-hierarchical walkthrough. This flag is also set for Cells in the Alliance CATAL. * New: In Hurricane::Cell_LeafInstanceOccurrences::progress(), add a test to prune leaf cells. * Bug: In Hurricane::getComponentOccurrences(), when calling the locator constructor, the "withLeafCells" argument was forgotten and was getting the value of "doExtraction" (wich, fortunately, was the same). * Change: In Hurricane::HypreNet_LeafPlugOccurrences::progress(), prune the leaf cells instead of the terminal ones. * Change: In Hurricane::HypreNet_ComponentOccurrences::progress(), prune the leaf cells instead of the terminal ones. * New: In Hurricane::Cell, add a "useFlattenLeaf" mode to take into account or ignore the "FlattenLeaf" flag when doing trans- hierarchical walkthrough. Not sure if it shouldn't be systematic. * New: In Etesian::toColoquinte() and other, activate the "FlattenLeaf" mode for all walkthough. So we can stop at already placed sub-blocks. * Bug: In Etesian::toColoquinte(), when placed sub-block where present, the number of instances was under estimated, so the instance array was oversized and the end went unitialized. So we were getting all kind of strange behavior from Coloquinte... Now throw an exception if the number of instances differs from the computed size. * Bug: In Etesian::toColoquinte(), as for the instance array, now correctly compute the size with the ecluded nets (supplies, clocks). * Bug: In Etesian::resetPlacement(), event if the Cell is unplaced, go through the instance list to remove the potential feed cells. The netlist is written back after placement *with* the feed cells added. So, when loading again, they where cluttering the netlist and giving feed redefintions warning. They also generated erratic behaviors. * New: In Anabatic::NetBuilderHV, added configurations for: * 1G_1M1_1PinM3 * 1G_xM1_1PinM2 * 2G_xM1_1PinM2 * doRp_AccessNorthPin * New: In Anabatic::NetBuilderHV, added Pin management in SingleGCell. * Bug: In Anabatic::NetBuilderHV::_do_xG_1PinM3(), the North/South global routing configuration was forgotten (only East/West). * Bug: In Katana::TrackFixedSegment, the blockage net was a static variable initialized at the first run. So we were getting stuck with the blockage net of the first cell to be routed. Of course, it did show only when Katana was run multiple times in the same process.
2020-03-03 17:50:18 -06:00
}
cdebug_tabw(145,-1);
}
Complete the modularization of the detail routing converter. * New: In Anabatic::NetBuilder, set all the attributes as private and create accessors and mutators. Finish virtualising all GCell build methods and transfer them into the NetBuilderHV class. Build methods now return a boolean to tell if the GCell was processed or not, to allow cascading in the "big switch". Reorganise the "big switch" in separate sections only partially cascading. Truly fuse the *big switch* for channel routing and over-the-cells. Create a new method "_do_globalSegment()" to delagate the drawing of global segments between two GCell to the derived classes.
2017-12-18 11:15:14 -06:00
string NetBuilderHV::getTypeName () const
{ return "NetBuilderHV"; }
Modularize the global to detail routing converter. * New: In Anabatic, rename GCellTopology to NetBuilder, expose the class (no longer in a local namespace). NetBuilder become a base class performing the walkthrough the Net tree. Derived class are tasked to build routing for specific gauge. NetBuilderHV is created to manage all gauge with metal2 horizontal and metal3 vertical. In our terminolgy we consider that the first routing metal is metal2. Metal1 is used inside the standard cells.
2017-12-16 17:13:19 -06:00
} // Anabatic namespace.