2010-03-09 09:24:29 -06:00
|
|
|
// -*- C++ -*-
|
|
|
|
//
|
|
|
|
// This file is part of the Coriolis Software.
|
2018-01-06 10:55:44 -06:00
|
|
|
// Copyright (c) UPMC 2008-2018, All Rights Reserved
|
2010-03-09 09:24:29 -06:00
|
|
|
//
|
2013-12-03 18:58:58 -06:00
|
|
|
// +-----------------------------------------------------------------+
|
2010-03-09 09:24:29 -06:00
|
|
|
// | C O R I O L I S |
|
|
|
|
// | K a t a b a t i c - Routing Toolbox |
|
|
|
|
// | |
|
|
|
|
// | Author : Jean-Paul CHAPUT |
|
2013-12-03 18:58:58 -06:00
|
|
|
// | E-mail : Jean-Paul.Chaput@lip6.fr |
|
2010-03-09 09:24:29 -06:00
|
|
|
// | =============================================================== |
|
|
|
|
// | C++ Module : "./LayerAssign.cpp" |
|
2013-12-03 18:58:58 -06:00
|
|
|
// +-----------------------------------------------------------------+
|
|
|
|
|
|
|
|
|
|
|
|
#include <cstdlib>
|
|
|
|
#include <sstream>
|
|
|
|
#include "hurricane/Bug.h"
|
|
|
|
#include "hurricane/Warning.h"
|
|
|
|
#include "hurricane/DebugSession.h"
|
|
|
|
#include "hurricane/Net.h"
|
|
|
|
#include "hurricane/NetExternalComponents.h"
|
|
|
|
#include "hurricane/Layer.h"
|
|
|
|
#include "hurricane/RoutingPad.h"
|
|
|
|
#include "hurricane/Pad.h"
|
|
|
|
#include "hurricane/Plug.h"
|
|
|
|
#include "hurricane/Instance.h"
|
|
|
|
#include "hurricane/Vertical.h"
|
|
|
|
#include "hurricane/Horizontal.h"
|
Add NetRoutingState (Property) on Nets to tell Kite what to do.
* New: In Katabatic, add the ability to decorate some (i.e. few) nets
with a state that indicate to Katabatic & Kite what to do with it.
The possible states are:
1. Fixed : all the wire are in fixed positions. The router cannot
move them and account them as obstacles.
2. ManualGlobalRoute : a user-defined topology is supplied. The
wires still have to be detailed route in "Detailed Pre-Route"
but will be skipped for the global routing and fixed for the
general Detailed route.
3. AutomaticGlobalRoute : ordinary nets, to be global routed then
detail routed.
4. Excluded : do not try to global or detail route thoses nets.
5. MixedPreRoute : mask combining Fixed and ManualGlobalRoute.
Not all nets have this property, only those that needs a special
processing.
To ease the access to the state, it is nested inside a
PrivateProperty in the net (NetRoutingProperty), with an extension
access class (NetRoutingExtension).
* New: In Kite, take account of NetRoutingState. Pointers to the
net's states are strored inside an internal map for faster access.
The property is *not* deleted when Kite is destroyed. The property
will remains until the Net itself is destroyed.
As a consequence, the lists that where passed to high level
function are removed as the information can now be accessed directly
through the net NetRoutingProperty.
* New: In Unicorn, in CgtMain, comply with the update interface.
* New: In documentation, update the User's Guide to explain the Pre-routed
step of Kite.
2014-07-02 07:36:58 -05:00
|
|
|
#include "hurricane/Cell.h"
|
2013-12-03 18:58:58 -06:00
|
|
|
#include "crlcore/RoutingGauge.h"
|
|
|
|
#include "katabatic/AutoContact.h"
|
|
|
|
#include "katabatic/AutoSegment.h"
|
|
|
|
#include "katabatic/GCellGrid.h"
|
|
|
|
#include "katabatic/KatabaticEngine.h"
|
2010-03-09 09:24:29 -06:00
|
|
|
|
|
|
|
|
|
|
|
namespace Katabatic {
|
|
|
|
|
|
|
|
using Hurricane::DebugSession;
|
|
|
|
using Hurricane::ForEachIterator;
|
|
|
|
using Hurricane::Warning;
|
|
|
|
|
|
|
|
|
2010-03-23 04:25:15 -05:00
|
|
|
void KatabaticEngine::_desaturate ( unsigned int depth
|
|
|
|
, set<Net*>& globalNets
|
|
|
|
, unsigned long& total
|
|
|
|
, unsigned long& globals )
|
2010-03-09 09:24:29 -06:00
|
|
|
{
|
Improved UpdateSession & exception catching. Start of RoutingGauge implem.
Miscellaneous:
* Change: In <crlcore>, in display.conf use the same display threshold
for both METAL2 & METAL3.
In alliance.conf, the side of VIAs in the gauge is 2l (not 3l).
In kite.conf, separate edge densities for H/V.
* Change: In <Cell>, in flattenNets() use flag as argument, not a
boolean. Do not create rings for clock or supply nets.
* Change: In <DeepNet>, in _createRoutingPads() do not create rings
for clock or supply net (duplicated policy as in Cell::flattenNets()).
* Bug: In <ControllerWidget>, at last find the bad signal disconnect
that was causing ungraceful messages.
* Change: In <knik>, in Edge display occupancy/capacity in the string
name. Improved display progress and debugging capabilities.
Improved exception catch & breakpoint managment:
* Bug: In <PaletteWidget>, in updateExtensions() replace the calls to
deleteLayer() by delete. This cause the widget to be immediatly
erased instead of waiting for the event queue to be completly
processed. This was causing the widget to be left in a incoherent
state when stoping at a breakpoint.
* Bug: In <BreakpointWidget>, in execNoModal(), flush the main event
loop (QApplication::flush()) *before* lauching the *local* event
loop. This is to ensure all widgets are in their final state when
waiting (especially <PaletteWidget>).
* Change: In <ExceptionWidget>, new method catchAllWrapper() to
execute any std::function< void() > function/method with a "try"/
"catch" wraparound and lauch the widget in case something is catch.
* New: In <hurricane>, support for a oberver pattern, backported from
<katabatic> with an Obervable capable of being linked to any
number of Obervers.
* New: In <Cell>, made it observable to detect Cell change, currently
emit two kind of signals:
- Cell::CellAboutToChange : *before* any change.
- Cell::CellChanged : *after* the change has been completed.
* New: In <UpdateSession>, in Go::invalidate() add the Cell owning the
Go to the UPDATOR_STACK (of course the cell is added only once).
In addition, when the Cell is added, send a notification of
Cell::CellAboutToChange to all it's observers. The slave instances
are also invalidated.
Conversely in UpdateSession::_preDestroy() for each invalidated
Cell send a Cell::CellChanged notification to all observer.
The UPDATOR_STACK has been slightly amended to accept Cell which
are not Gos. Prior to this, the Cell where completly excluded from
the UpdateSession mechanism, so it's instances where never actualised
of anything referring to the Cell for that matter.
Note: we use two different mechanisms to transmit a Cell change,
observers and the slave instance map. I think at some point it
should be unificated.
* Change: In <CellViewer>, make it a Cell observer to redraw when the
cell is modificated (also update the palette).
Uses the catchAllWrapper() to protect all critical actions.
* Change: In <GraphicTool>, no longer need of cellPreModificated and
cellPostModificated signals. Now done through the Cell obersvers.
* Change: In <mauka>, <etesian> & <kite> now uses the catchAllWrapper
method for protection (need to split methods in two, to be able
to pass it as argument). No longer emit cellPreModificated and
cellPostModificated.
Support for RoutingGauge in P&R:
* Bug: In <placeandroute.py>, the connection from the internal power
ring to the connectors was not done correctly. Wrong contact layers
leading to a gap.
* Change: In <BuildPowerRails>, detection of the corona signals based
on how the "pck_px" pad is connected. No longer based on name
matching.
* Change: In <placeandroute.py>, support for 2 routing metal only
(3 metal in the technology).
* Change: In <katabatic> & <kite> support for a "top layer" limitation
on the routing gauge, this allows to use only two routing metals
(METAL2 & METAL3). Work in progress.
2014-04-20 12:25:08 -05:00
|
|
|
if (depth+2 > Session::getConfiguration()->getAllowedDepth()) {
|
2010-03-09 09:24:29 -06:00
|
|
|
cerr << Warning("Katabatic::_desaturate(): %s, no remaining upper layers."
|
|
|
|
,getString(Session::getRoutingGauge()->getRoutingLayer(depth)->getName()).c_str()
|
|
|
|
) << endl;
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
cmess1 << " o Desaturate layer "
|
|
|
|
<< Session::getRoutingGauge()->getRoutingLayer(depth)->getName() << endl;
|
|
|
|
|
2013-12-03 18:58:58 -06:00
|
|
|
GCellDensitySet queue ( depth, *(_gcellGrid->getGCellVector()) );
|
|
|
|
GCell::SetIndex invalidateds;
|
2010-03-09 09:24:29 -06:00
|
|
|
|
|
|
|
bool optimized = true;
|
|
|
|
while ( optimized ) {
|
2010-12-12 15:43:25 -06:00
|
|
|
Session::revalidate ();
|
2010-03-09 09:24:29 -06:00
|
|
|
optimized = false;
|
2013-12-03 18:58:58 -06:00
|
|
|
queue.requeue ();
|
2010-03-23 04:25:15 -05:00
|
|
|
|
|
|
|
std::set<GCell*,GCell::CompareByKey>::const_iterator igcell = queue.getGCells().begin();
|
|
|
|
size_t i = 0;
|
|
|
|
for ( ; igcell!=queue.getGCells().end() ; ++igcell, ++i ) {
|
2016-06-11 14:56:12 -05:00
|
|
|
cdebug_log(149,0) << "_desaturate: [" << depth << "]:"
|
2011-01-09 12:08:31 -06:00
|
|
|
<< (*igcell)->getDensity(depth) << " " << *igcell << endl;
|
|
|
|
|
2013-12-03 18:58:58 -06:00
|
|
|
if (not (*igcell)->isSaturated(depth)) {
|
2016-06-11 14:56:12 -05:00
|
|
|
cdebug_log(149,0) << "STOP desaturated: @" << i << " " << *igcell << endl;
|
2010-03-23 04:25:15 -05:00
|
|
|
for ( ; igcell!=queue.getGCells().end() ; ++igcell ) {
|
2013-12-03 18:58:58 -06:00
|
|
|
if ( (*igcell)->isSaturated( depth ) ) {
|
|
|
|
cparanoid << "[ERROR] Still saturated: @" << i << " " << *igcell << endl;
|
2010-03-23 04:25:15 -05:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
2013-12-03 18:58:58 -06:00
|
|
|
optimized = (*igcell)->stepNetDesaturate( depth, globalNets, invalidateds );
|
* ./katabatic:
- New: In AutoSegment, adds a "_parent" attribute to keep track of the fragmentation
processus. Currently used only for strap segments, points to the original
segment in the appropriate direction (before the split).
- New: In GCell & LayerAssign, new method of layer assignment. Move up
the whole net trunk if only one of it's segment is inside an over-saturated
GCell. AutoSegment are moved up only if there is at least 2 free tracks
remaining on the upper level.
- Change: In Session::_canonize(), uses the lowest segment Id as canonical.
More reliable than geometricals criterions in the end. Assuming that the
segments are being created in deterministic order, which *should* be the
case consediring the way we are walking through the global routing.
- Change: In AutoSegment, completly suppress the CompareCanonical(), replace
it by the much simpler CompareId().
- Change: In GCell::rpDesaturate(), stops desaturation when bottom density
is under 0.5, otherwise we are causing a severe imbalance in M2/M4
densities. All wires pushed up to M4...
- Change: In ChipTools, for the Pad's RoutingPad, reslect the best component
using the one in the lowest layer. To avoid problem when splitting
AutoContact as we expect the base Contact to be on the lower layer.
- Bug: In GCellConfiguration::_GCell_xG_xL1_xL3(), add H/V alignement constraints
in fork case. This allow NE/SW contact to be splitted correctly later.
- Bug: In AutoContact::split(), the connexity on the splitted contacts was
not correctly restored, leading to canonization and parentage looping
errors. This was concealed by the Kite Track::_check() bug (incomplete
individual TrackSegment checking).
2010-12-30 12:41:19 -06:00
|
|
|
if ( optimized ) {
|
2013-12-03 18:58:58 -06:00
|
|
|
for ( GCell::SetIndex::iterator igcell=invalidateds.begin() ; igcell!=invalidateds.end() ; ++igcell ) {
|
|
|
|
queue.unqueue( *igcell );
|
2010-03-23 04:25:15 -05:00
|
|
|
}
|
* ./katabatic:
- New: In AutoSegment, adds a "_parent" attribute to keep track of the fragmentation
processus. Currently used only for strap segments, points to the original
segment in the appropriate direction (before the split).
- New: In GCell & LayerAssign, new method of layer assignment. Move up
the whole net trunk if only one of it's segment is inside an over-saturated
GCell. AutoSegment are moved up only if there is at least 2 free tracks
remaining on the upper level.
- Change: In Session::_canonize(), uses the lowest segment Id as canonical.
More reliable than geometricals criterions in the end. Assuming that the
segments are being created in deterministic order, which *should* be the
case consediring the way we are walking through the global routing.
- Change: In AutoSegment, completly suppress the CompareCanonical(), replace
it by the much simpler CompareId().
- Change: In GCell::rpDesaturate(), stops desaturation when bottom density
is under 0.5, otherwise we are causing a severe imbalance in M2/M4
densities. All wires pushed up to M4...
- Change: In ChipTools, for the Pad's RoutingPad, reslect the best component
using the one in the lowest layer. To avoid problem when splitting
AutoContact as we expect the base Contact to be on the lower layer.
- Bug: In GCellConfiguration::_GCell_xG_xL1_xL3(), add H/V alignement constraints
in fork case. This allow NE/SW contact to be splitted correctly later.
- Bug: In AutoContact::split(), the connexity on the splitted contacts was
not correctly restored, leading to canonization and parentage looping
errors. This was concealed by the Kite Track::_check() bug (incomplete
individual TrackSegment checking).
2010-12-30 12:41:19 -06:00
|
|
|
break;
|
2010-03-23 04:25:15 -05:00
|
|
|
}
|
2010-03-09 09:24:29 -06:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
void KatabaticEngine::_layerAssignByLength ( Net* net, unsigned long& total, unsigned long& global, set<Net*>& globalNets )
|
|
|
|
{
|
2016-05-17 16:00:06 -05:00
|
|
|
DebugSession::open ( net, 140, 150 );
|
2010-03-09 09:24:29 -06:00
|
|
|
|
2016-06-11 14:56:12 -05:00
|
|
|
cdebug_log(149,0) << "Katabatic::_layerAssignByLength( " << net << " )" << endl;
|
|
|
|
cdebug_tabw(145,1);
|
2010-03-09 09:24:29 -06:00
|
|
|
|
|
|
|
bool isGlobal = false;
|
|
|
|
set<AutoContact*> globalContacts;
|
|
|
|
|
|
|
|
forEach ( Segment*, isegment, net->getSegments() ) {
|
|
|
|
total++;
|
|
|
|
if ( (*isegment)->getLength() > getGlobalThreshold() ) {
|
2013-12-03 18:58:58 -06:00
|
|
|
if (not isGlobal) {
|
2010-03-09 09:24:29 -06:00
|
|
|
isGlobal = true;
|
2013-12-03 18:58:58 -06:00
|
|
|
globalNets.insert( net );
|
2010-03-09 09:24:29 -06:00
|
|
|
}
|
|
|
|
|
|
|
|
global++;
|
2013-12-03 18:58:58 -06:00
|
|
|
if ((*isegment)->getLayer() == Session::getRoutingLayer(1)) (*isegment)->setLayer( Session::getRoutingLayer(3) );
|
|
|
|
if ((*isegment)->getLayer() == Session::getRoutingLayer(2)) (*isegment)->setLayer( Session::getRoutingLayer(4) );
|
2010-03-09 09:24:29 -06:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2016-06-11 14:56:12 -05:00
|
|
|
cdebug_tabw(145,-1);
|
2010-03-09 09:24:29 -06:00
|
|
|
|
2013-12-03 18:58:58 -06:00
|
|
|
DebugSession::close();
|
2010-03-09 09:24:29 -06:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
void KatabaticEngine::_layerAssignByLength ( unsigned long& total, unsigned long& global, set<Net*>& globalNets )
|
|
|
|
{
|
|
|
|
cmess1 << " o Assign Layer (simple wirelength)." << endl;
|
|
|
|
|
Add NetRoutingState (Property) on Nets to tell Kite what to do.
* New: In Katabatic, add the ability to decorate some (i.e. few) nets
with a state that indicate to Katabatic & Kite what to do with it.
The possible states are:
1. Fixed : all the wire are in fixed positions. The router cannot
move them and account them as obstacles.
2. ManualGlobalRoute : a user-defined topology is supplied. The
wires still have to be detailed route in "Detailed Pre-Route"
but will be skipped for the global routing and fixed for the
general Detailed route.
3. AutomaticGlobalRoute : ordinary nets, to be global routed then
detail routed.
4. Excluded : do not try to global or detail route thoses nets.
5. MixedPreRoute : mask combining Fixed and ManualGlobalRoute.
Not all nets have this property, only those that needs a special
processing.
To ease the access to the state, it is nested inside a
PrivateProperty in the net (NetRoutingProperty), with an extension
access class (NetRoutingExtension).
* New: In Kite, take account of NetRoutingState. Pointers to the
net's states are strored inside an internal map for faster access.
The property is *not* deleted when Kite is destroyed. The property
will remains until the Net itself is destroyed.
As a consequence, the lists that where passed to high level
function are removed as the information can now be accessed directly
through the net NetRoutingProperty.
* New: In Unicorn, in CgtMain, comply with the update interface.
* New: In documentation, update the User's Guide to explain the Pre-routed
step of Kite.
2014-07-02 07:36:58 -05:00
|
|
|
forEach ( Net* , inet , getCell()->getNets() ) {
|
|
|
|
if (NetRoutingExtension::get(*inet)->isAutomaticGlobalRoute()) {
|
|
|
|
_layerAssignByLength ( *inet, total, global, globalNets );
|
|
|
|
}
|
|
|
|
} // forEach(Net*)
|
2010-03-09 09:24:29 -06:00
|
|
|
}
|
|
|
|
|
|
|
|
|
* ./katabatic:
- New: In AutoSegment, adds a "_parent" attribute to keep track of the fragmentation
processus. Currently used only for strap segments, points to the original
segment in the appropriate direction (before the split).
- New: In GCell & LayerAssign, new method of layer assignment. Move up
the whole net trunk if only one of it's segment is inside an over-saturated
GCell. AutoSegment are moved up only if there is at least 2 free tracks
remaining on the upper level.
- Change: In Session::_canonize(), uses the lowest segment Id as canonical.
More reliable than geometricals criterions in the end. Assuming that the
segments are being created in deterministic order, which *should* be the
case consediring the way we are walking through the global routing.
- Change: In AutoSegment, completly suppress the CompareCanonical(), replace
it by the much simpler CompareId().
- Change: In GCell::rpDesaturate(), stops desaturation when bottom density
is under 0.5, otherwise we are causing a severe imbalance in M2/M4
densities. All wires pushed up to M4...
- Change: In ChipTools, for the Pad's RoutingPad, reslect the best component
using the one in the lowest layer. To avoid problem when splitting
AutoContact as we expect the base Contact to be on the lower layer.
- Bug: In GCellConfiguration::_GCell_xG_xL1_xL3(), add H/V alignement constraints
in fork case. This allow NE/SW contact to be splitted correctly later.
- Bug: In AutoContact::split(), the connexity on the splitted contacts was
not correctly restored, leading to canonization and parentage looping
errors. This was concealed by the Kite Track::_check() bug (incomplete
individual TrackSegment checking).
2010-12-30 12:41:19 -06:00
|
|
|
void KatabaticEngine::_layerAssignByTrunk ( Net* net, set<Net*>& globalNets, unsigned long& total, unsigned long& global )
|
2010-03-09 09:24:29 -06:00
|
|
|
{
|
2016-05-17 16:00:06 -05:00
|
|
|
DebugSession::open( net, 140, 150 );
|
2010-03-09 09:24:29 -06:00
|
|
|
|
2016-06-11 14:56:12 -05:00
|
|
|
cdebug_log(149,0) << "Katabatic::_layerAssignByTrunk ( " << net << " )" << endl;
|
|
|
|
cdebug_tabw(145,1);
|
2010-03-09 09:24:29 -06:00
|
|
|
|
|
|
|
bool isGlobal = false;
|
|
|
|
unsigned long netGlobal = 0;
|
|
|
|
unsigned long netTotal = 0;
|
|
|
|
set<AutoContact*> globalContacts;
|
|
|
|
|
|
|
|
forEach ( Segment*, isegment, net->getSegments() ) {
|
|
|
|
netTotal++;
|
|
|
|
|
2013-12-03 18:58:58 -06:00
|
|
|
if ((*isegment)->getLength() > getGlobalThreshold()) {
|
2010-03-09 09:24:29 -06:00
|
|
|
isGlobal = true;
|
|
|
|
netTotal = 0;
|
2013-12-03 18:58:58 -06:00
|
|
|
globalNets.insert( net );
|
2010-03-09 09:24:29 -06:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2013-12-03 18:58:58 -06:00
|
|
|
if (isGlobal) {
|
2010-03-09 09:24:29 -06:00
|
|
|
forEach ( Segment*, isegment, net->getSegments() ) {
|
|
|
|
netTotal++;
|
|
|
|
|
2013-12-03 18:58:58 -06:00
|
|
|
AutoSegment* autoSegment = Session::lookup( *isegment );
|
|
|
|
if ( autoSegment and not autoSegment->isStrongTerminal() ) {
|
2010-03-09 09:24:29 -06:00
|
|
|
netGlobal++;
|
|
|
|
|
2016-06-11 14:56:12 -05:00
|
|
|
cdebug_log(145,0) << "Migrate to M4/M5: " << autoSegment << endl;
|
2013-12-03 18:58:58 -06:00
|
|
|
if (autoSegment->isHorizontal()) autoSegment->setLayer( Session::getRoutingLayer(3) );
|
|
|
|
if (autoSegment->isVertical ()) autoSegment->setLayer( Session::getRoutingLayer(4) );
|
2010-03-09 09:24:29 -06:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
total += netTotal;
|
|
|
|
global += netGlobal;
|
|
|
|
|
2016-06-11 14:56:12 -05:00
|
|
|
cdebug_tabw(145,-1);
|
2010-03-09 09:24:29 -06:00
|
|
|
|
2013-12-03 18:58:58 -06:00
|
|
|
DebugSession::close();
|
2010-03-09 09:24:29 -06:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
void KatabaticEngine::_layerAssignByTrunk ( unsigned long& total, unsigned long& global, set<Net*>& globalNets )
|
|
|
|
{
|
|
|
|
cmess1 << " o Assign Layer (whole net trunk)." << endl;
|
|
|
|
|
Add NetRoutingState (Property) on Nets to tell Kite what to do.
* New: In Katabatic, add the ability to decorate some (i.e. few) nets
with a state that indicate to Katabatic & Kite what to do with it.
The possible states are:
1. Fixed : all the wire are in fixed positions. The router cannot
move them and account them as obstacles.
2. ManualGlobalRoute : a user-defined topology is supplied. The
wires still have to be detailed route in "Detailed Pre-Route"
but will be skipped for the global routing and fixed for the
general Detailed route.
3. AutomaticGlobalRoute : ordinary nets, to be global routed then
detail routed.
4. Excluded : do not try to global or detail route thoses nets.
5. MixedPreRoute : mask combining Fixed and ManualGlobalRoute.
Not all nets have this property, only those that needs a special
processing.
To ease the access to the state, it is nested inside a
PrivateProperty in the net (NetRoutingProperty), with an extension
access class (NetRoutingExtension).
* New: In Kite, take account of NetRoutingState. Pointers to the
net's states are strored inside an internal map for faster access.
The property is *not* deleted when Kite is destroyed. The property
will remains until the Net itself is destroyed.
As a consequence, the lists that where passed to high level
function are removed as the information can now be accessed directly
through the net NetRoutingProperty.
* New: In Unicorn, in CgtMain, comply with the update interface.
* New: In documentation, update the User's Guide to explain the Pre-routed
step of Kite.
2014-07-02 07:36:58 -05:00
|
|
|
forEach ( Net* , inet , getCell()->getNets() ) {
|
|
|
|
if (NetRoutingExtension::get(*inet)->isAutomaticGlobalRoute()) {
|
|
|
|
_layerAssignByTrunk( *inet, globalNets, total, global );
|
|
|
|
}
|
|
|
|
}
|
* ./katabatic:
- New: In AutoSegment, adds a "_parent" attribute to keep track of the fragmentation
processus. Currently used only for strap segments, points to the original
segment in the appropriate direction (before the split).
- New: In GCell & LayerAssign, new method of layer assignment. Move up
the whole net trunk if only one of it's segment is inside an over-saturated
GCell. AutoSegment are moved up only if there is at least 2 free tracks
remaining on the upper level.
- Change: In Session::_canonize(), uses the lowest segment Id as canonical.
More reliable than geometricals criterions in the end. Assuming that the
segments are being created in deterministic order, which *should* be the
case consediring the way we are walking through the global routing.
- Change: In AutoSegment, completly suppress the CompareCanonical(), replace
it by the much simpler CompareId().
- Change: In GCell::rpDesaturate(), stops desaturation when bottom density
is under 0.5, otherwise we are causing a severe imbalance in M2/M4
densities. All wires pushed up to M4...
- Change: In ChipTools, for the Pad's RoutingPad, reslect the best component
using the one in the lowest layer. To avoid problem when splitting
AutoContact as we expect the base Contact to be on the lower layer.
- Bug: In GCellConfiguration::_GCell_xG_xL1_xL3(), add H/V alignement constraints
in fork case. This allow NE/SW contact to be splitted correctly later.
- Bug: In AutoContact::split(), the connexity on the splitted contacts was
not correctly restored, leading to canonization and parentage looping
errors. This was concealed by the Kite Track::_check() bug (incomplete
individual TrackSegment checking).
2010-12-30 12:41:19 -06:00
|
|
|
}
|
|
|
|
|
|
|
|
|
2013-12-03 18:58:58 -06:00
|
|
|
void KatabaticEngine::moveULeft ( AutoSegment* seed, set<Net*>& globalNets, GCell::SetIndex& invalidateds )
|
* ./katabatic:
- New: In AutoSegment, adds a "_parent" attribute to keep track of the fragmentation
processus. Currently used only for strap segments, points to the original
segment in the appropriate direction (before the split).
- New: In GCell & LayerAssign, new method of layer assignment. Move up
the whole net trunk if only one of it's segment is inside an over-saturated
GCell. AutoSegment are moved up only if there is at least 2 free tracks
remaining on the upper level.
- Change: In Session::_canonize(), uses the lowest segment Id as canonical.
More reliable than geometricals criterions in the end. Assuming that the
segments are being created in deterministic order, which *should* be the
case consediring the way we are walking through the global routing.
- Change: In AutoSegment, completly suppress the CompareCanonical(), replace
it by the much simpler CompareId().
- Change: In GCell::rpDesaturate(), stops desaturation when bottom density
is under 0.5, otherwise we are causing a severe imbalance in M2/M4
densities. All wires pushed up to M4...
- Change: In ChipTools, for the Pad's RoutingPad, reslect the best component
using the one in the lowest layer. To avoid problem when splitting
AutoContact as we expect the base Contact to be on the lower layer.
- Bug: In GCellConfiguration::_GCell_xG_xL1_xL3(), add H/V alignement constraints
in fork case. This allow NE/SW contact to be splitted correctly later.
- Bug: In AutoContact::split(), the connexity on the splitted contacts was
not correctly restored, leading to canonization and parentage looping
errors. This was concealed by the Kite Track::_check() bug (incomplete
individual TrackSegment checking).
2010-12-30 12:41:19 -06:00
|
|
|
{
|
|
|
|
Net* net = seed->getNet();
|
2016-05-17 16:00:06 -05:00
|
|
|
DebugSession::open( net, 140, 150 );
|
2013-12-03 18:58:58 -06:00
|
|
|
|
2016-06-11 14:56:12 -05:00
|
|
|
cdebug_log(9000,0) << "Deter| Move left: " << seed << endl;
|
2013-12-03 18:58:58 -06:00
|
|
|
|
|
|
|
seed->moveULeft();
|
|
|
|
globalNets.insert( net );
|
|
|
|
|
|
|
|
GCell* begin = seed->getAutoSource()->getGCell();
|
|
|
|
GCell* end = seed->getAutoTarget()->getGCell();
|
|
|
|
|
|
|
|
if (seed->isHorizontal()) {
|
|
|
|
for ( GCell* gcell=begin ; gcell and gcell!=end ; gcell=gcell->getRight() )
|
|
|
|
invalidateds.insert( gcell );
|
|
|
|
|
|
|
|
begin = begin->getDown();
|
|
|
|
end = end ->getDown();
|
|
|
|
|
|
|
|
for ( GCell* gcell=begin ; gcell and gcell!=end ; gcell=gcell->getRight() )
|
|
|
|
invalidateds.insert( gcell );
|
|
|
|
} else {
|
|
|
|
for ( GCell* gcell=begin ; gcell and gcell!=end ; gcell=gcell->getUp() )
|
|
|
|
invalidateds.insert( gcell );
|
|
|
|
|
|
|
|
begin = begin->getLeft();
|
|
|
|
end = end ->getLeft();
|
|
|
|
|
|
|
|
for ( GCell* gcell=begin ; gcell and gcell!=end ; gcell=gcell->getUp() )
|
|
|
|
invalidateds.insert( gcell );
|
|
|
|
}
|
|
|
|
|
|
|
|
DebugSession::close();
|
|
|
|
}
|
* ./katabatic:
- New: In AutoSegment, adds a "_parent" attribute to keep track of the fragmentation
processus. Currently used only for strap segments, points to the original
segment in the appropriate direction (before the split).
- New: In GCell & LayerAssign, new method of layer assignment. Move up
the whole net trunk if only one of it's segment is inside an over-saturated
GCell. AutoSegment are moved up only if there is at least 2 free tracks
remaining on the upper level.
- Change: In Session::_canonize(), uses the lowest segment Id as canonical.
More reliable than geometricals criterions in the end. Assuming that the
segments are being created in deterministic order, which *should* be the
case consediring the way we are walking through the global routing.
- Change: In AutoSegment, completly suppress the CompareCanonical(), replace
it by the much simpler CompareId().
- Change: In GCell::rpDesaturate(), stops desaturation when bottom density
is under 0.5, otherwise we are causing a severe imbalance in M2/M4
densities. All wires pushed up to M4...
- Change: In ChipTools, for the Pad's RoutingPad, reslect the best component
using the one in the lowest layer. To avoid problem when splitting
AutoContact as we expect the base Contact to be on the lower layer.
- Bug: In GCellConfiguration::_GCell_xG_xL1_xL3(), add H/V alignement constraints
in fork case. This allow NE/SW contact to be splitted correctly later.
- Bug: In AutoContact::split(), the connexity on the splitted contacts was
not correctly restored, leading to canonization and parentage looping
errors. This was concealed by the Kite Track::_check() bug (incomplete
individual TrackSegment checking).
2010-12-30 12:41:19 -06:00
|
|
|
|
|
|
|
|
2013-12-03 18:58:58 -06:00
|
|
|
void KatabaticEngine::moveURight ( AutoSegment* seed, set<Net*>& globalNets, GCell::SetIndex& invalidateds )
|
|
|
|
{
|
|
|
|
Net* net = seed->getNet();
|
2016-05-17 16:00:06 -05:00
|
|
|
DebugSession::open( net, 140, 150 );
|
2013-12-03 18:58:58 -06:00
|
|
|
|
2016-06-11 14:56:12 -05:00
|
|
|
cdebug_log(9000,0) << "Deter| Move right: " << seed << endl;
|
2013-12-03 18:58:58 -06:00
|
|
|
|
|
|
|
seed->moveURight();
|
|
|
|
globalNets.insert( net );
|
|
|
|
|
|
|
|
GCell* begin = seed->getAutoSource()->getGCell();
|
|
|
|
GCell* end = seed->getAutoTarget()->getGCell();
|
|
|
|
|
|
|
|
if (seed->isHorizontal()) {
|
|
|
|
for ( GCell* gcell=begin ; gcell and gcell!=end ; gcell=gcell->getRight() )
|
|
|
|
invalidateds.insert( gcell );
|
|
|
|
|
|
|
|
begin = begin->getUp();
|
|
|
|
end = end ->getUp();
|
|
|
|
|
|
|
|
for ( GCell* gcell=begin ; gcell and gcell!=end ; gcell=gcell->getRight() )
|
|
|
|
invalidateds.insert( gcell );
|
|
|
|
} else {
|
|
|
|
for ( GCell* gcell=begin ; gcell and gcell!=end ; gcell=gcell->getUp() )
|
|
|
|
invalidateds.insert( gcell );
|
|
|
|
|
|
|
|
begin = begin->getRight();
|
|
|
|
end = end ->getRight();
|
|
|
|
|
|
|
|
for ( GCell* gcell=begin ; gcell and gcell!=end ; gcell=gcell->getUp() )
|
|
|
|
invalidateds.insert( gcell );
|
|
|
|
}
|
|
|
|
|
|
|
|
DebugSession::close();
|
|
|
|
}
|
|
|
|
|
|
|
|
|
In Katabatic, the layer assignment move up part of Net trunks.
* In Katabatic, in AutoSegment::canMoveUp(), adds the ability to reject
the move up if the density of the depth we comes from is below 0.6
in all the GCells, activated with the KbChecklowDensity flag.
* In Katabatic, In LayerAssign, new method moveUpNetTrunk2() which,
unlike the previous one, move up only the part of the Net trunk
which is connex to the seed segment. Use the canMoveUp with
low density checking to avoid unbalancing higher metal layers.
* In Kite, in SegmentFsm, use the canMoveUp() with low density checking.
2016-03-15 12:06:43 -05:00
|
|
|
bool KatabaticEngine::moveUpNetTrunk2 ( AutoSegment* seed, set<Net*>& globalNets, GCell::SetIndex& invalidateds )
|
|
|
|
{
|
|
|
|
Net* net = seed->getNet();
|
|
|
|
unsigned int seedDepth = Session::getRoutingGauge()->getLayerDepth(seed->getLayer());
|
|
|
|
|
2016-05-17 16:00:06 -05:00
|
|
|
DebugSession::open( net, 140, 150 );
|
2016-06-11 14:56:12 -05:00
|
|
|
cdebug_log(9000,0) << "Deter| moveUpNetTrunk() depth:" << seedDepth << " " << seed << endl;
|
In Katabatic, the layer assignment move up part of Net trunks.
* In Katabatic, in AutoSegment::canMoveUp(), adds the ability to reject
the move up if the density of the depth we comes from is below 0.6
in all the GCells, activated with the KbChecklowDensity flag.
* In Katabatic, In LayerAssign, new method moveUpNetTrunk2() which,
unlike the previous one, move up only the part of the Net trunk
which is connex to the seed segment. Use the canMoveUp with
low density checking to avoid unbalancing higher metal layers.
* In Kite, in SegmentFsm, use the canMoveUp() with low density checking.
2016-03-15 12:06:43 -05:00
|
|
|
|
|
|
|
if (not seed->canMoveUp( 1.0, KbPropagate|KbAllowTerminal|KbNoCheckLayer) ) {
|
2016-06-11 14:56:12 -05:00
|
|
|
cdebug_log(9000,0) << "Deter| Reject seed move up, cannot move up." << endl;
|
In Katabatic, the layer assignment move up part of Net trunks.
* In Katabatic, in AutoSegment::canMoveUp(), adds the ability to reject
the move up if the density of the depth we comes from is below 0.6
in all the GCells, activated with the KbChecklowDensity flag.
* In Katabatic, In LayerAssign, new method moveUpNetTrunk2() which,
unlike the previous one, move up only the part of the Net trunk
which is connex to the seed segment. Use the canMoveUp with
low density checking to avoid unbalancing higher metal layers.
* In Kite, in SegmentFsm, use the canMoveUp() with low density checking.
2016-03-15 12:06:43 -05:00
|
|
|
DebugSession::close();
|
|
|
|
return false;
|
|
|
|
}
|
2016-06-11 14:56:12 -05:00
|
|
|
cdebug_tabw(149,1);
|
In Katabatic, the layer assignment move up part of Net trunks.
* In Katabatic, in AutoSegment::canMoveUp(), adds the ability to reject
the move up if the density of the depth we comes from is below 0.6
in all the GCells, activated with the KbChecklowDensity flag.
* In Katabatic, In LayerAssign, new method moveUpNetTrunk2() which,
unlike the previous one, move up only the part of the Net trunk
which is connex to the seed segment. Use the canMoveUp with
low density checking to avoid unbalancing higher metal layers.
* In Kite, in SegmentFsm, use the canMoveUp() with low density checking.
2016-03-15 12:06:43 -05:00
|
|
|
|
|
|
|
globalNets.insert( net );
|
|
|
|
|
|
|
|
vector< pair<AutoContact*,AutoSegment*> > stack;
|
|
|
|
vector<AutoSegment*> globals;
|
|
|
|
vector<AutoSegment*> locals;
|
|
|
|
|
|
|
|
stack.push_back( pair<AutoContact*,AutoSegment*>(NULL,seed) );
|
|
|
|
while ( not stack.empty() ) {
|
|
|
|
AutoContact* from = stack.back().first;
|
|
|
|
AutoSegment* segment = stack.back().second;
|
|
|
|
stack.pop_back();
|
|
|
|
|
|
|
|
if (segment->isLocal()) {
|
|
|
|
if (not segment->isStrongTerminal()) locals.push_back( segment );
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
if ( (segment->getLength() < 3*Session::getSliceHeight()) and (segment != seed) ) {
|
|
|
|
locals.push_back( segment );
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
// Do something here.
|
|
|
|
if (not segment->canMoveUp(1.0,KbPropagate|KbAllowTerminal|KbNoCheckLayer|KbCheckLowDensity) )
|
|
|
|
continue;
|
|
|
|
|
|
|
|
globals.push_back( segment );
|
|
|
|
|
|
|
|
AutoContact* source = segment->getAutoSource();
|
|
|
|
if (source != from) {
|
|
|
|
for ( AutoSegment* connected : source->getAutoSegments() ) {
|
|
|
|
if (connected != segment) { stack.push_back( make_pair(source,connected) ); }
|
|
|
|
}
|
|
|
|
}
|
|
|
|
AutoContact* target = segment->getAutoTarget();
|
|
|
|
if (target != from) {
|
|
|
|
for ( AutoSegment* connected : target->getAutoSegments() ) {
|
|
|
|
if (connected != segment) { stack.push_back( make_pair(target,connected) ); }
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
for ( size_t i=0 ; i<globals.size() ; ++i ) {
|
2016-06-11 14:56:12 -05:00
|
|
|
//cdebug_log(9000,0) << "Deter| Looking up G:" << globals[i] << endl;
|
In Katabatic, the layer assignment move up part of Net trunks.
* In Katabatic, in AutoSegment::canMoveUp(), adds the ability to reject
the move up if the density of the depth we comes from is below 0.6
in all the GCells, activated with the KbChecklowDensity flag.
* In Katabatic, In LayerAssign, new method moveUpNetTrunk2() which,
unlike the previous one, move up only the part of the Net trunk
which is connex to the seed segment. Use the canMoveUp with
low density checking to avoid unbalancing higher metal layers.
* In Kite, in SegmentFsm, use the canMoveUp() with low density checking.
2016-03-15 12:06:43 -05:00
|
|
|
unsigned int depth = Session::getRoutingGauge()->getLayerDepth( globals[i]->getLayer() );
|
|
|
|
globals[i]->changeDepth( depth+2, KbWithNeighbors );
|
|
|
|
|
|
|
|
vector<GCell*> gcells;
|
|
|
|
globals[i]->getGCells( gcells );
|
|
|
|
for ( size_t j=0 ; j<gcells.size() ; j++ ) {
|
|
|
|
invalidateds.insert( gcells[j] );
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
for ( size_t i=0 ; i<locals.size() ; ++i ) {
|
2016-06-11 14:56:12 -05:00
|
|
|
//cdebug_log(9000,0) << "Deter| Looking up L:" << locals[i] << endl;
|
In Katabatic, the layer assignment move up part of Net trunks.
* In Katabatic, in AutoSegment::canMoveUp(), adds the ability to reject
the move up if the density of the depth we comes from is below 0.6
in all the GCells, activated with the KbChecklowDensity flag.
* In Katabatic, In LayerAssign, new method moveUpNetTrunk2() which,
unlike the previous one, move up only the part of the Net trunk
which is connex to the seed segment. Use the canMoveUp with
low density checking to avoid unbalancing higher metal layers.
* In Kite, in SegmentFsm, use the canMoveUp() with low density checking.
2016-03-15 12:06:43 -05:00
|
|
|
|
|
|
|
unsigned int depth = Session::getRoutingGauge()->getLayerDepth(locals[i]->getLayer());
|
|
|
|
if (depth > seedDepth+1) continue;
|
|
|
|
|
|
|
|
if (locals[i]->canPivotUp(2.0,KbPropagate|KbNoCheckLayer)) {
|
|
|
|
locals[i]->changeDepth( depth+2, KbWithNeighbors );
|
|
|
|
|
2016-06-11 14:56:12 -05:00
|
|
|
//cdebug_log(9000,0) << "Deter| Trunk move up L:" << locals[i] << endl;
|
In Katabatic, the layer assignment move up part of Net trunks.
* In Katabatic, in AutoSegment::canMoveUp(), adds the ability to reject
the move up if the density of the depth we comes from is below 0.6
in all the GCells, activated with the KbChecklowDensity flag.
* In Katabatic, In LayerAssign, new method moveUpNetTrunk2() which,
unlike the previous one, move up only the part of the Net trunk
which is connex to the seed segment. Use the canMoveUp with
low density checking to avoid unbalancing higher metal layers.
* In Kite, in SegmentFsm, use the canMoveUp() with low density checking.
2016-03-15 12:06:43 -05:00
|
|
|
|
|
|
|
vector<GCell*> gcells;
|
|
|
|
locals[i]->getGCells( gcells );
|
|
|
|
for ( size_t j=0 ; j<gcells.size() ; j++ ) {
|
|
|
|
invalidateds.insert( gcells[j] );
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2016-06-11 14:56:12 -05:00
|
|
|
cdebug_tabw(149,-1);
|
In Katabatic, the layer assignment move up part of Net trunks.
* In Katabatic, in AutoSegment::canMoveUp(), adds the ability to reject
the move up if the density of the depth we comes from is below 0.6
in all the GCells, activated with the KbChecklowDensity flag.
* In Katabatic, In LayerAssign, new method moveUpNetTrunk2() which,
unlike the previous one, move up only the part of the Net trunk
which is connex to the seed segment. Use the canMoveUp with
low density checking to avoid unbalancing higher metal layers.
* In Kite, in SegmentFsm, use the canMoveUp() with low density checking.
2016-03-15 12:06:43 -05:00
|
|
|
DebugSession::close();
|
|
|
|
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
2013-12-03 18:58:58 -06:00
|
|
|
bool KatabaticEngine::moveUpNetTrunk ( AutoSegment* seed, set<Net*>& globalNets, GCell::SetIndex& invalidateds )
|
|
|
|
{
|
|
|
|
Net* net = seed->getNet();
|
|
|
|
unsigned int seedDepth = Session::getRoutingGauge()->getLayerDepth(seed->getLayer());
|
|
|
|
|
2016-05-17 16:00:06 -05:00
|
|
|
DebugSession::open( net, 140, 150 );
|
2016-06-11 14:56:12 -05:00
|
|
|
cdebug_log(9000,0) << "Deter| moveUpNetTrunk() depth:" << seedDepth << " " << seed << endl;
|
2013-12-03 18:58:58 -06:00
|
|
|
|
|
|
|
if (not seed->canMoveUp( 1.0, KbPropagate|KbAllowTerminal|KbNoCheckLayer) ) {
|
2016-06-11 14:56:12 -05:00
|
|
|
cdebug_log(9000,0) << "Deter| Reject seed move up, cannot move up." << endl;
|
2013-12-03 18:58:58 -06:00
|
|
|
DebugSession::close();
|
* ./katabatic:
- New: In AutoSegment, adds a "_parent" attribute to keep track of the fragmentation
processus. Currently used only for strap segments, points to the original
segment in the appropriate direction (before the split).
- New: In GCell & LayerAssign, new method of layer assignment. Move up
the whole net trunk if only one of it's segment is inside an over-saturated
GCell. AutoSegment are moved up only if there is at least 2 free tracks
remaining on the upper level.
- Change: In Session::_canonize(), uses the lowest segment Id as canonical.
More reliable than geometricals criterions in the end. Assuming that the
segments are being created in deterministic order, which *should* be the
case consediring the way we are walking through the global routing.
- Change: In AutoSegment, completly suppress the CompareCanonical(), replace
it by the much simpler CompareId().
- Change: In GCell::rpDesaturate(), stops desaturation when bottom density
is under 0.5, otherwise we are causing a severe imbalance in M2/M4
densities. All wires pushed up to M4...
- Change: In ChipTools, for the Pad's RoutingPad, reslect the best component
using the one in the lowest layer. To avoid problem when splitting
AutoContact as we expect the base Contact to be on the lower layer.
- Bug: In GCellConfiguration::_GCell_xG_xL1_xL3(), add H/V alignement constraints
in fork case. This allow NE/SW contact to be splitted correctly later.
- Bug: In AutoContact::split(), the connexity on the splitted contacts was
not correctly restored, leading to canonization and parentage looping
errors. This was concealed by the Kite Track::_check() bug (incomplete
individual TrackSegment checking).
2010-12-30 12:41:19 -06:00
|
|
|
return false;
|
|
|
|
}
|
2016-06-11 14:56:12 -05:00
|
|
|
cdebug_tabw(149,1);
|
* ./katabatic:
- New: In AutoSegment, adds a "_parent" attribute to keep track of the fragmentation
processus. Currently used only for strap segments, points to the original
segment in the appropriate direction (before the split).
- New: In GCell & LayerAssign, new method of layer assignment. Move up
the whole net trunk if only one of it's segment is inside an over-saturated
GCell. AutoSegment are moved up only if there is at least 2 free tracks
remaining on the upper level.
- Change: In Session::_canonize(), uses the lowest segment Id as canonical.
More reliable than geometricals criterions in the end. Assuming that the
segments are being created in deterministic order, which *should* be the
case consediring the way we are walking through the global routing.
- Change: In AutoSegment, completly suppress the CompareCanonical(), replace
it by the much simpler CompareId().
- Change: In GCell::rpDesaturate(), stops desaturation when bottom density
is under 0.5, otherwise we are causing a severe imbalance in M2/M4
densities. All wires pushed up to M4...
- Change: In ChipTools, for the Pad's RoutingPad, reslect the best component
using the one in the lowest layer. To avoid problem when splitting
AutoContact as we expect the base Contact to be on the lower layer.
- Bug: In GCellConfiguration::_GCell_xG_xL1_xL3(), add H/V alignement constraints
in fork case. This allow NE/SW contact to be splitted correctly later.
- Bug: In AutoContact::split(), the connexity on the splitted contacts was
not correctly restored, leading to canonization and parentage looping
errors. This was concealed by the Kite Track::_check() bug (incomplete
individual TrackSegment checking).
2010-12-30 12:41:19 -06:00
|
|
|
|
2013-12-03 18:58:58 -06:00
|
|
|
globalNets.insert( net );
|
* ./katabatic:
- New: In AutoSegment, adds a "_parent" attribute to keep track of the fragmentation
processus. Currently used only for strap segments, points to the original
segment in the appropriate direction (before the split).
- New: In GCell & LayerAssign, new method of layer assignment. Move up
the whole net trunk if only one of it's segment is inside an over-saturated
GCell. AutoSegment are moved up only if there is at least 2 free tracks
remaining on the upper level.
- Change: In Session::_canonize(), uses the lowest segment Id as canonical.
More reliable than geometricals criterions in the end. Assuming that the
segments are being created in deterministic order, which *should* be the
case consediring the way we are walking through the global routing.
- Change: In AutoSegment, completly suppress the CompareCanonical(), replace
it by the much simpler CompareId().
- Change: In GCell::rpDesaturate(), stops desaturation when bottom density
is under 0.5, otherwise we are causing a severe imbalance in M2/M4
densities. All wires pushed up to M4...
- Change: In ChipTools, for the Pad's RoutingPad, reslect the best component
using the one in the lowest layer. To avoid problem when splitting
AutoContact as we expect the base Contact to be on the lower layer.
- Bug: In GCellConfiguration::_GCell_xG_xL1_xL3(), add H/V alignement constraints
in fork case. This allow NE/SW contact to be splitted correctly later.
- Bug: In AutoContact::split(), the connexity on the splitted contacts was
not correctly restored, leading to canonization and parentage looping
errors. This was concealed by the Kite Track::_check() bug (incomplete
individual TrackSegment checking).
2010-12-30 12:41:19 -06:00
|
|
|
|
|
|
|
vector<AutoSegment*> globals;
|
|
|
|
vector<AutoSegment*> locals;
|
|
|
|
|
|
|
|
forEach ( Segment*, isegment, net->getSegments() ) {
|
2013-12-03 18:58:58 -06:00
|
|
|
AutoSegment* autoSegment = Session::lookup( *isegment );
|
* ./katabatic:
- New: In AutoSegment, adds a "_parent" attribute to keep track of the fragmentation
processus. Currently used only for strap segments, points to the original
segment in the appropriate direction (before the split).
- New: In GCell & LayerAssign, new method of layer assignment. Move up
the whole net trunk if only one of it's segment is inside an over-saturated
GCell. AutoSegment are moved up only if there is at least 2 free tracks
remaining on the upper level.
- Change: In Session::_canonize(), uses the lowest segment Id as canonical.
More reliable than geometricals criterions in the end. Assuming that the
segments are being created in deterministic order, which *should* be the
case consediring the way we are walking through the global routing.
- Change: In AutoSegment, completly suppress the CompareCanonical(), replace
it by the much simpler CompareId().
- Change: In GCell::rpDesaturate(), stops desaturation when bottom density
is under 0.5, otherwise we are causing a severe imbalance in M2/M4
densities. All wires pushed up to M4...
- Change: In ChipTools, for the Pad's RoutingPad, reslect the best component
using the one in the lowest layer. To avoid problem when splitting
AutoContact as we expect the base Contact to be on the lower layer.
- Bug: In GCellConfiguration::_GCell_xG_xL1_xL3(), add H/V alignement constraints
in fork case. This allow NE/SW contact to be splitted correctly later.
- Bug: In AutoContact::split(), the connexity on the splitted contacts was
not correctly restored, leading to canonization and parentage looping
errors. This was concealed by the Kite Track::_check() bug (incomplete
individual TrackSegment checking).
2010-12-30 12:41:19 -06:00
|
|
|
|
2016-06-11 14:56:12 -05:00
|
|
|
//cdebug_log(9000,0) << "Deter| Loop " << autoSegment << endl;
|
2013-12-03 18:58:58 -06:00
|
|
|
|
|
|
|
if (not autoSegment) continue;
|
|
|
|
if (autoSegment->isLocal()) {
|
|
|
|
if (autoSegment->isStrongTerminal()) continue;
|
|
|
|
locals.push_back( autoSegment );
|
* ./katabatic:
- New: In AutoSegment, adds a "_parent" attribute to keep track of the fragmentation
processus. Currently used only for strap segments, points to the original
segment in the appropriate direction (before the split).
- New: In GCell & LayerAssign, new method of layer assignment. Move up
the whole net trunk if only one of it's segment is inside an over-saturated
GCell. AutoSegment are moved up only if there is at least 2 free tracks
remaining on the upper level.
- Change: In Session::_canonize(), uses the lowest segment Id as canonical.
More reliable than geometricals criterions in the end. Assuming that the
segments are being created in deterministic order, which *should* be the
case consediring the way we are walking through the global routing.
- Change: In AutoSegment, completly suppress the CompareCanonical(), replace
it by the much simpler CompareId().
- Change: In GCell::rpDesaturate(), stops desaturation when bottom density
is under 0.5, otherwise we are causing a severe imbalance in M2/M4
densities. All wires pushed up to M4...
- Change: In ChipTools, for the Pad's RoutingPad, reslect the best component
using the one in the lowest layer. To avoid problem when splitting
AutoContact as we expect the base Contact to be on the lower layer.
- Bug: In GCellConfiguration::_GCell_xG_xL1_xL3(), add H/V alignement constraints
in fork case. This allow NE/SW contact to be splitted correctly later.
- Bug: In AutoContact::split(), the connexity on the splitted contacts was
not correctly restored, leading to canonization and parentage looping
errors. This was concealed by the Kite Track::_check() bug (incomplete
individual TrackSegment checking).
2010-12-30 12:41:19 -06:00
|
|
|
} else {
|
|
|
|
// Ugly: Hard-coded GCell side.
|
Support of RoutingGauge, part 2.
In Katabatic & Kite, remove all hard-coded values related to track pitches.
* New: In <Session>, add more convenience function to access RoutingGauge
characteristics.
* New: In <AutoSegment>, <AutoContact>, add support for the "depth spin",
that is, if the source/target contacts are going "top" or "down".
Used to compute the perpandicular pitch. Need a small modification
of the revalidation mechanism. The observers of <AutoSegment> are
notified when the spin changes.
* New: In <AutoSegment>, the getPPitch() method allow to compute the
"perpandicular pitch". For now it is simply the greatest from the
source perpandicular pitch and the target perpandicular pitch.
Make uses of the "depth spin".
* New: In <TrackElement>, <TrackSegment>, cache the perpandicular pitch.
Updated through the notification from the observable.
2014-05-19 10:58:38 -05:00
|
|
|
if ( (autoSegment->getLength() < 3*Session::getSliceHeight()) and (autoSegment != seed) )
|
2013-12-03 18:58:58 -06:00
|
|
|
locals.push_back( autoSegment );
|
* ./katabatic:
- New: In AutoSegment, adds a "_parent" attribute to keep track of the fragmentation
processus. Currently used only for strap segments, points to the original
segment in the appropriate direction (before the split).
- New: In GCell & LayerAssign, new method of layer assignment. Move up
the whole net trunk if only one of it's segment is inside an over-saturated
GCell. AutoSegment are moved up only if there is at least 2 free tracks
remaining on the upper level.
- Change: In Session::_canonize(), uses the lowest segment Id as canonical.
More reliable than geometricals criterions in the end. Assuming that the
segments are being created in deterministic order, which *should* be the
case consediring the way we are walking through the global routing.
- Change: In AutoSegment, completly suppress the CompareCanonical(), replace
it by the much simpler CompareId().
- Change: In GCell::rpDesaturate(), stops desaturation when bottom density
is under 0.5, otherwise we are causing a severe imbalance in M2/M4
densities. All wires pushed up to M4...
- Change: In ChipTools, for the Pad's RoutingPad, reslect the best component
using the one in the lowest layer. To avoid problem when splitting
AutoContact as we expect the base Contact to be on the lower layer.
- Bug: In GCellConfiguration::_GCell_xG_xL1_xL3(), add H/V alignement constraints
in fork case. This allow NE/SW contact to be splitted correctly later.
- Bug: In AutoContact::split(), the connexity on the splitted contacts was
not correctly restored, leading to canonization and parentage looping
errors. This was concealed by the Kite Track::_check() bug (incomplete
individual TrackSegment checking).
2010-12-30 12:41:19 -06:00
|
|
|
else
|
2013-12-03 18:58:58 -06:00
|
|
|
globals.push_back( autoSegment );
|
* ./katabatic:
- New: In AutoSegment, adds a "_parent" attribute to keep track of the fragmentation
processus. Currently used only for strap segments, points to the original
segment in the appropriate direction (before the split).
- New: In GCell & LayerAssign, new method of layer assignment. Move up
the whole net trunk if only one of it's segment is inside an over-saturated
GCell. AutoSegment are moved up only if there is at least 2 free tracks
remaining on the upper level.
- Change: In Session::_canonize(), uses the lowest segment Id as canonical.
More reliable than geometricals criterions in the end. Assuming that the
segments are being created in deterministic order, which *should* be the
case consediring the way we are walking through the global routing.
- Change: In AutoSegment, completly suppress the CompareCanonical(), replace
it by the much simpler CompareId().
- Change: In GCell::rpDesaturate(), stops desaturation when bottom density
is under 0.5, otherwise we are causing a severe imbalance in M2/M4
densities. All wires pushed up to M4...
- Change: In ChipTools, for the Pad's RoutingPad, reslect the best component
using the one in the lowest layer. To avoid problem when splitting
AutoContact as we expect the base Contact to be on the lower layer.
- Bug: In GCellConfiguration::_GCell_xG_xL1_xL3(), add H/V alignement constraints
in fork case. This allow NE/SW contact to be splitted correctly later.
- Bug: In AutoContact::split(), the connexity on the splitted contacts was
not correctly restored, leading to canonization and parentage looping
errors. This was concealed by the Kite Track::_check() bug (incomplete
individual TrackSegment checking).
2010-12-30 12:41:19 -06:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2013-12-03 18:58:58 -06:00
|
|
|
sort( globals.begin(), globals.end(), AutoSegment::CompareId() );
|
|
|
|
sort( locals.begin(), locals.end(), AutoSegment::CompareId() );
|
* ./katabatic:
- New: In AutoSegment, adds a "_parent" attribute to keep track of the fragmentation
processus. Currently used only for strap segments, points to the original
segment in the appropriate direction (before the split).
- New: In GCell & LayerAssign, new method of layer assignment. Move up
the whole net trunk if only one of it's segment is inside an over-saturated
GCell. AutoSegment are moved up only if there is at least 2 free tracks
remaining on the upper level.
- Change: In Session::_canonize(), uses the lowest segment Id as canonical.
More reliable than geometricals criterions in the end. Assuming that the
segments are being created in deterministic order, which *should* be the
case consediring the way we are walking through the global routing.
- Change: In AutoSegment, completly suppress the CompareCanonical(), replace
it by the much simpler CompareId().
- Change: In GCell::rpDesaturate(), stops desaturation when bottom density
is under 0.5, otherwise we are causing a severe imbalance in M2/M4
densities. All wires pushed up to M4...
- Change: In ChipTools, for the Pad's RoutingPad, reslect the best component
using the one in the lowest layer. To avoid problem when splitting
AutoContact as we expect the base Contact to be on the lower layer.
- Bug: In GCellConfiguration::_GCell_xG_xL1_xL3(), add H/V alignement constraints
in fork case. This allow NE/SW contact to be splitted correctly later.
- Bug: In AutoContact::split(), the connexity on the splitted contacts was
not correctly restored, leading to canonization and parentage looping
errors. This was concealed by the Kite Track::_check() bug (incomplete
individual TrackSegment checking).
2010-12-30 12:41:19 -06:00
|
|
|
|
|
|
|
for ( size_t i=0 ; i<globals.size() ; ++i ) {
|
2016-06-11 14:56:12 -05:00
|
|
|
//cdebug_log(9000,0) << "Deter| Looking up G:" << globals[i] << endl;
|
2013-12-03 18:58:58 -06:00
|
|
|
|
|
|
|
unsigned int depth = Session::getRoutingGauge()->getLayerDepth( globals[i]->getLayer() );
|
|
|
|
if (depth > seedDepth+1) continue;
|
* ./katabatic:
- New: In AutoSegment, adds a "_parent" attribute to keep track of the fragmentation
processus. Currently used only for strap segments, points to the original
segment in the appropriate direction (before the split).
- New: In GCell & LayerAssign, new method of layer assignment. Move up
the whole net trunk if only one of it's segment is inside an over-saturated
GCell. AutoSegment are moved up only if there is at least 2 free tracks
remaining on the upper level.
- Change: In Session::_canonize(), uses the lowest segment Id as canonical.
More reliable than geometricals criterions in the end. Assuming that the
segments are being created in deterministic order, which *should* be the
case consediring the way we are walking through the global routing.
- Change: In AutoSegment, completly suppress the CompareCanonical(), replace
it by the much simpler CompareId().
- Change: In GCell::rpDesaturate(), stops desaturation when bottom density
is under 0.5, otherwise we are causing a severe imbalance in M2/M4
densities. All wires pushed up to M4...
- Change: In ChipTools, for the Pad's RoutingPad, reslect the best component
using the one in the lowest layer. To avoid problem when splitting
AutoContact as we expect the base Contact to be on the lower layer.
- Bug: In GCellConfiguration::_GCell_xG_xL1_xL3(), add H/V alignement constraints
in fork case. This allow NE/SW contact to be splitted correctly later.
- Bug: In AutoContact::split(), the connexity on the splitted contacts was
not correctly restored, leading to canonization and parentage looping
errors. This was concealed by the Kite Track::_check() bug (incomplete
individual TrackSegment checking).
2010-12-30 12:41:19 -06:00
|
|
|
|
2011-01-25 11:16:27 -06:00
|
|
|
if ( globals[i]->isVertical() and (globals[i] != seed) ) continue;
|
|
|
|
|
2013-12-03 18:58:58 -06:00
|
|
|
if ( globals[i]->canMoveUp(1.0,KbPropagate|KbAllowTerminal|KbNoCheckLayer) ) {
|
|
|
|
globals[i]->changeDepth( depth+2, KbWithNeighbors );
|
* ./katabatic:
- New: In AutoSegment, adds a "_parent" attribute to keep track of the fragmentation
processus. Currently used only for strap segments, points to the original
segment in the appropriate direction (before the split).
- New: In GCell & LayerAssign, new method of layer assignment. Move up
the whole net trunk if only one of it's segment is inside an over-saturated
GCell. AutoSegment are moved up only if there is at least 2 free tracks
remaining on the upper level.
- Change: In Session::_canonize(), uses the lowest segment Id as canonical.
More reliable than geometricals criterions in the end. Assuming that the
segments are being created in deterministic order, which *should* be the
case consediring the way we are walking through the global routing.
- Change: In AutoSegment, completly suppress the CompareCanonical(), replace
it by the much simpler CompareId().
- Change: In GCell::rpDesaturate(), stops desaturation when bottom density
is under 0.5, otherwise we are causing a severe imbalance in M2/M4
densities. All wires pushed up to M4...
- Change: In ChipTools, for the Pad's RoutingPad, reslect the best component
using the one in the lowest layer. To avoid problem when splitting
AutoContact as we expect the base Contact to be on the lower layer.
- Bug: In GCellConfiguration::_GCell_xG_xL1_xL3(), add H/V alignement constraints
in fork case. This allow NE/SW contact to be splitted correctly later.
- Bug: In AutoContact::split(), the connexity on the splitted contacts was
not correctly restored, leading to canonization and parentage looping
errors. This was concealed by the Kite Track::_check() bug (incomplete
individual TrackSegment checking).
2010-12-30 12:41:19 -06:00
|
|
|
|
2016-06-11 14:56:12 -05:00
|
|
|
//cdebug_log(9000,0) << "Deter| Trunk move up G:" << globals[i] << endl;
|
* ./katabatic:
- New: In AutoSegment, adds a "_parent" attribute to keep track of the fragmentation
processus. Currently used only for strap segments, points to the original
segment in the appropriate direction (before the split).
- New: In GCell & LayerAssign, new method of layer assignment. Move up
the whole net trunk if only one of it's segment is inside an over-saturated
GCell. AutoSegment are moved up only if there is at least 2 free tracks
remaining on the upper level.
- Change: In Session::_canonize(), uses the lowest segment Id as canonical.
More reliable than geometricals criterions in the end. Assuming that the
segments are being created in deterministic order, which *should* be the
case consediring the way we are walking through the global routing.
- Change: In AutoSegment, completly suppress the CompareCanonical(), replace
it by the much simpler CompareId().
- Change: In GCell::rpDesaturate(), stops desaturation when bottom density
is under 0.5, otherwise we are causing a severe imbalance in M2/M4
densities. All wires pushed up to M4...
- Change: In ChipTools, for the Pad's RoutingPad, reslect the best component
using the one in the lowest layer. To avoid problem when splitting
AutoContact as we expect the base Contact to be on the lower layer.
- Bug: In GCellConfiguration::_GCell_xG_xL1_xL3(), add H/V alignement constraints
in fork case. This allow NE/SW contact to be splitted correctly later.
- Bug: In AutoContact::split(), the connexity on the splitted contacts was
not correctly restored, leading to canonization and parentage looping
errors. This was concealed by the Kite Track::_check() bug (incomplete
individual TrackSegment checking).
2010-12-30 12:41:19 -06:00
|
|
|
|
|
|
|
vector<GCell*> gcells;
|
2013-12-03 18:58:58 -06:00
|
|
|
globals[i]->getGCells( gcells );
|
* ./katabatic:
- New: In AutoSegment, adds a "_parent" attribute to keep track of the fragmentation
processus. Currently used only for strap segments, points to the original
segment in the appropriate direction (before the split).
- New: In GCell & LayerAssign, new method of layer assignment. Move up
the whole net trunk if only one of it's segment is inside an over-saturated
GCell. AutoSegment are moved up only if there is at least 2 free tracks
remaining on the upper level.
- Change: In Session::_canonize(), uses the lowest segment Id as canonical.
More reliable than geometricals criterions in the end. Assuming that the
segments are being created in deterministic order, which *should* be the
case consediring the way we are walking through the global routing.
- Change: In AutoSegment, completly suppress the CompareCanonical(), replace
it by the much simpler CompareId().
- Change: In GCell::rpDesaturate(), stops desaturation when bottom density
is under 0.5, otherwise we are causing a severe imbalance in M2/M4
densities. All wires pushed up to M4...
- Change: In ChipTools, for the Pad's RoutingPad, reslect the best component
using the one in the lowest layer. To avoid problem when splitting
AutoContact as we expect the base Contact to be on the lower layer.
- Bug: In GCellConfiguration::_GCell_xG_xL1_xL3(), add H/V alignement constraints
in fork case. This allow NE/SW contact to be splitted correctly later.
- Bug: In AutoContact::split(), the connexity on the splitted contacts was
not correctly restored, leading to canonization and parentage looping
errors. This was concealed by the Kite Track::_check() bug (incomplete
individual TrackSegment checking).
2010-12-30 12:41:19 -06:00
|
|
|
for ( size_t j=0 ; j<gcells.size() ; j++ ) {
|
2013-12-03 18:58:58 -06:00
|
|
|
invalidateds.insert( gcells[j] );
|
* ./katabatic:
- New: In AutoSegment, adds a "_parent" attribute to keep track of the fragmentation
processus. Currently used only for strap segments, points to the original
segment in the appropriate direction (before the split).
- New: In GCell & LayerAssign, new method of layer assignment. Move up
the whole net trunk if only one of it's segment is inside an over-saturated
GCell. AutoSegment are moved up only if there is at least 2 free tracks
remaining on the upper level.
- Change: In Session::_canonize(), uses the lowest segment Id as canonical.
More reliable than geometricals criterions in the end. Assuming that the
segments are being created in deterministic order, which *should* be the
case consediring the way we are walking through the global routing.
- Change: In AutoSegment, completly suppress the CompareCanonical(), replace
it by the much simpler CompareId().
- Change: In GCell::rpDesaturate(), stops desaturation when bottom density
is under 0.5, otherwise we are causing a severe imbalance in M2/M4
densities. All wires pushed up to M4...
- Change: In ChipTools, for the Pad's RoutingPad, reslect the best component
using the one in the lowest layer. To avoid problem when splitting
AutoContact as we expect the base Contact to be on the lower layer.
- Bug: In GCellConfiguration::_GCell_xG_xL1_xL3(), add H/V alignement constraints
in fork case. This allow NE/SW contact to be splitted correctly later.
- Bug: In AutoContact::split(), the connexity on the splitted contacts was
not correctly restored, leading to canonization and parentage looping
errors. This was concealed by the Kite Track::_check() bug (incomplete
individual TrackSegment checking).
2010-12-30 12:41:19 -06:00
|
|
|
}
|
|
|
|
} else {
|
2016-06-11 14:56:12 -05:00
|
|
|
//cdebug_log(9000,0) << "Deter| Reject Trunk move up G:" << globals[i] << endl;
|
* ./katabatic:
- New: In AutoSegment, adds a "_parent" attribute to keep track of the fragmentation
processus. Currently used only for strap segments, points to the original
segment in the appropriate direction (before the split).
- New: In GCell & LayerAssign, new method of layer assignment. Move up
the whole net trunk if only one of it's segment is inside an over-saturated
GCell. AutoSegment are moved up only if there is at least 2 free tracks
remaining on the upper level.
- Change: In Session::_canonize(), uses the lowest segment Id as canonical.
More reliable than geometricals criterions in the end. Assuming that the
segments are being created in deterministic order, which *should* be the
case consediring the way we are walking through the global routing.
- Change: In AutoSegment, completly suppress the CompareCanonical(), replace
it by the much simpler CompareId().
- Change: In GCell::rpDesaturate(), stops desaturation when bottom density
is under 0.5, otherwise we are causing a severe imbalance in M2/M4
densities. All wires pushed up to M4...
- Change: In ChipTools, for the Pad's RoutingPad, reslect the best component
using the one in the lowest layer. To avoid problem when splitting
AutoContact as we expect the base Contact to be on the lower layer.
- Bug: In GCellConfiguration::_GCell_xG_xL1_xL3(), add H/V alignement constraints
in fork case. This allow NE/SW contact to be splitted correctly later.
- Bug: In AutoContact::split(), the connexity on the splitted contacts was
not correctly restored, leading to canonization and parentage looping
errors. This was concealed by the Kite Track::_check() bug (incomplete
individual TrackSegment checking).
2010-12-30 12:41:19 -06:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
for ( size_t i=0 ; i<locals.size() ; ++i ) {
|
2016-06-11 14:56:12 -05:00
|
|
|
//cdebug_log(9000,0) << "Deter| Looking up L:" << locals[i] << endl;
|
* ./katabatic:
- New: In AutoSegment, adds a "_parent" attribute to keep track of the fragmentation
processus. Currently used only for strap segments, points to the original
segment in the appropriate direction (before the split).
- New: In GCell & LayerAssign, new method of layer assignment. Move up
the whole net trunk if only one of it's segment is inside an over-saturated
GCell. AutoSegment are moved up only if there is at least 2 free tracks
remaining on the upper level.
- Change: In Session::_canonize(), uses the lowest segment Id as canonical.
More reliable than geometricals criterions in the end. Assuming that the
segments are being created in deterministic order, which *should* be the
case consediring the way we are walking through the global routing.
- Change: In AutoSegment, completly suppress the CompareCanonical(), replace
it by the much simpler CompareId().
- Change: In GCell::rpDesaturate(), stops desaturation when bottom density
is under 0.5, otherwise we are causing a severe imbalance in M2/M4
densities. All wires pushed up to M4...
- Change: In ChipTools, for the Pad's RoutingPad, reslect the best component
using the one in the lowest layer. To avoid problem when splitting
AutoContact as we expect the base Contact to be on the lower layer.
- Bug: In GCellConfiguration::_GCell_xG_xL1_xL3(), add H/V alignement constraints
in fork case. This allow NE/SW contact to be splitted correctly later.
- Bug: In AutoContact::split(), the connexity on the splitted contacts was
not correctly restored, leading to canonization and parentage looping
errors. This was concealed by the Kite Track::_check() bug (incomplete
individual TrackSegment checking).
2010-12-30 12:41:19 -06:00
|
|
|
|
2013-12-03 18:58:58 -06:00
|
|
|
unsigned int depth = Session::getRoutingGauge()->getLayerDepth(locals[i]->getLayer());
|
|
|
|
if (depth > seedDepth+1) continue;
|
2011-01-25 11:16:27 -06:00
|
|
|
|
2013-12-03 18:58:58 -06:00
|
|
|
if (locals[i]->canPivotUp(2.0,KbPropagate|KbNoCheckLayer)) {
|
|
|
|
locals[i]->changeDepth( depth+2, KbWithNeighbors );
|
* ./katabatic:
- New: In AutoSegment, adds a "_parent" attribute to keep track of the fragmentation
processus. Currently used only for strap segments, points to the original
segment in the appropriate direction (before the split).
- New: In GCell & LayerAssign, new method of layer assignment. Move up
the whole net trunk if only one of it's segment is inside an over-saturated
GCell. AutoSegment are moved up only if there is at least 2 free tracks
remaining on the upper level.
- Change: In Session::_canonize(), uses the lowest segment Id as canonical.
More reliable than geometricals criterions in the end. Assuming that the
segments are being created in deterministic order, which *should* be the
case consediring the way we are walking through the global routing.
- Change: In AutoSegment, completly suppress the CompareCanonical(), replace
it by the much simpler CompareId().
- Change: In GCell::rpDesaturate(), stops desaturation when bottom density
is under 0.5, otherwise we are causing a severe imbalance in M2/M4
densities. All wires pushed up to M4...
- Change: In ChipTools, for the Pad's RoutingPad, reslect the best component
using the one in the lowest layer. To avoid problem when splitting
AutoContact as we expect the base Contact to be on the lower layer.
- Bug: In GCellConfiguration::_GCell_xG_xL1_xL3(), add H/V alignement constraints
in fork case. This allow NE/SW contact to be splitted correctly later.
- Bug: In AutoContact::split(), the connexity on the splitted contacts was
not correctly restored, leading to canonization and parentage looping
errors. This was concealed by the Kite Track::_check() bug (incomplete
individual TrackSegment checking).
2010-12-30 12:41:19 -06:00
|
|
|
|
2016-06-11 14:56:12 -05:00
|
|
|
//cdebug_log(9000,0) << "Deter| Trunk move up L:" << locals[i] << endl;
|
* ./katabatic:
- New: In AutoSegment, adds a "_parent" attribute to keep track of the fragmentation
processus. Currently used only for strap segments, points to the original
segment in the appropriate direction (before the split).
- New: In GCell & LayerAssign, new method of layer assignment. Move up
the whole net trunk if only one of it's segment is inside an over-saturated
GCell. AutoSegment are moved up only if there is at least 2 free tracks
remaining on the upper level.
- Change: In Session::_canonize(), uses the lowest segment Id as canonical.
More reliable than geometricals criterions in the end. Assuming that the
segments are being created in deterministic order, which *should* be the
case consediring the way we are walking through the global routing.
- Change: In AutoSegment, completly suppress the CompareCanonical(), replace
it by the much simpler CompareId().
- Change: In GCell::rpDesaturate(), stops desaturation when bottom density
is under 0.5, otherwise we are causing a severe imbalance in M2/M4
densities. All wires pushed up to M4...
- Change: In ChipTools, for the Pad's RoutingPad, reslect the best component
using the one in the lowest layer. To avoid problem when splitting
AutoContact as we expect the base Contact to be on the lower layer.
- Bug: In GCellConfiguration::_GCell_xG_xL1_xL3(), add H/V alignement constraints
in fork case. This allow NE/SW contact to be splitted correctly later.
- Bug: In AutoContact::split(), the connexity on the splitted contacts was
not correctly restored, leading to canonization and parentage looping
errors. This was concealed by the Kite Track::_check() bug (incomplete
individual TrackSegment checking).
2010-12-30 12:41:19 -06:00
|
|
|
|
|
|
|
vector<GCell*> gcells;
|
2013-12-03 18:58:58 -06:00
|
|
|
locals[i]->getGCells( gcells );
|
* ./katabatic:
- New: In AutoSegment, adds a "_parent" attribute to keep track of the fragmentation
processus. Currently used only for strap segments, points to the original
segment in the appropriate direction (before the split).
- New: In GCell & LayerAssign, new method of layer assignment. Move up
the whole net trunk if only one of it's segment is inside an over-saturated
GCell. AutoSegment are moved up only if there is at least 2 free tracks
remaining on the upper level.
- Change: In Session::_canonize(), uses the lowest segment Id as canonical.
More reliable than geometricals criterions in the end. Assuming that the
segments are being created in deterministic order, which *should* be the
case consediring the way we are walking through the global routing.
- Change: In AutoSegment, completly suppress the CompareCanonical(), replace
it by the much simpler CompareId().
- Change: In GCell::rpDesaturate(), stops desaturation when bottom density
is under 0.5, otherwise we are causing a severe imbalance in M2/M4
densities. All wires pushed up to M4...
- Change: In ChipTools, for the Pad's RoutingPad, reslect the best component
using the one in the lowest layer. To avoid problem when splitting
AutoContact as we expect the base Contact to be on the lower layer.
- Bug: In GCellConfiguration::_GCell_xG_xL1_xL3(), add H/V alignement constraints
in fork case. This allow NE/SW contact to be splitted correctly later.
- Bug: In AutoContact::split(), the connexity on the splitted contacts was
not correctly restored, leading to canonization and parentage looping
errors. This was concealed by the Kite Track::_check() bug (incomplete
individual TrackSegment checking).
2010-12-30 12:41:19 -06:00
|
|
|
for ( size_t j=0 ; j<gcells.size() ; j++ ) {
|
2013-12-03 18:58:58 -06:00
|
|
|
invalidateds.insert( gcells[j] );
|
* ./katabatic:
- New: In AutoSegment, adds a "_parent" attribute to keep track of the fragmentation
processus. Currently used only for strap segments, points to the original
segment in the appropriate direction (before the split).
- New: In GCell & LayerAssign, new method of layer assignment. Move up
the whole net trunk if only one of it's segment is inside an over-saturated
GCell. AutoSegment are moved up only if there is at least 2 free tracks
remaining on the upper level.
- Change: In Session::_canonize(), uses the lowest segment Id as canonical.
More reliable than geometricals criterions in the end. Assuming that the
segments are being created in deterministic order, which *should* be the
case consediring the way we are walking through the global routing.
- Change: In AutoSegment, completly suppress the CompareCanonical(), replace
it by the much simpler CompareId().
- Change: In GCell::rpDesaturate(), stops desaturation when bottom density
is under 0.5, otherwise we are causing a severe imbalance in M2/M4
densities. All wires pushed up to M4...
- Change: In ChipTools, for the Pad's RoutingPad, reslect the best component
using the one in the lowest layer. To avoid problem when splitting
AutoContact as we expect the base Contact to be on the lower layer.
- Bug: In GCellConfiguration::_GCell_xG_xL1_xL3(), add H/V alignement constraints
in fork case. This allow NE/SW contact to be splitted correctly later.
- Bug: In AutoContact::split(), the connexity on the splitted contacts was
not correctly restored, leading to canonization and parentage looping
errors. This was concealed by the Kite Track::_check() bug (incomplete
individual TrackSegment checking).
2010-12-30 12:41:19 -06:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2016-06-11 14:56:12 -05:00
|
|
|
cdebug_tabw(149,-1);
|
2013-12-03 18:58:58 -06:00
|
|
|
DebugSession::close();
|
* ./katabatic:
- New: In AutoSegment, adds a "_parent" attribute to keep track of the fragmentation
processus. Currently used only for strap segments, points to the original
segment in the appropriate direction (before the split).
- New: In GCell & LayerAssign, new method of layer assignment. Move up
the whole net trunk if only one of it's segment is inside an over-saturated
GCell. AutoSegment are moved up only if there is at least 2 free tracks
remaining on the upper level.
- Change: In Session::_canonize(), uses the lowest segment Id as canonical.
More reliable than geometricals criterions in the end. Assuming that the
segments are being created in deterministic order, which *should* be the
case consediring the way we are walking through the global routing.
- Change: In AutoSegment, completly suppress the CompareCanonical(), replace
it by the much simpler CompareId().
- Change: In GCell::rpDesaturate(), stops desaturation when bottom density
is under 0.5, otherwise we are causing a severe imbalance in M2/M4
densities. All wires pushed up to M4...
- Change: In ChipTools, for the Pad's RoutingPad, reslect the best component
using the one in the lowest layer. To avoid problem when splitting
AutoContact as we expect the base Contact to be on the lower layer.
- Bug: In GCellConfiguration::_GCell_xG_xL1_xL3(), add H/V alignement constraints
in fork case. This allow NE/SW contact to be splitted correctly later.
- Bug: In AutoContact::split(), the connexity on the splitted contacts was
not correctly restored, leading to canonization and parentage looping
errors. This was concealed by the Kite Track::_check() bug (incomplete
individual TrackSegment checking).
2010-12-30 12:41:19 -06:00
|
|
|
|
|
|
|
return true;
|
2010-03-09 09:24:29 -06:00
|
|
|
}
|
|
|
|
|
|
|
|
|
2013-12-03 18:58:58 -06:00
|
|
|
void KatabaticEngine::_balanceGlobalDensity ( unsigned int depth )
|
|
|
|
{
|
|
|
|
startMeasures();
|
|
|
|
Session::open( this );
|
|
|
|
|
|
|
|
cmess1 << " o Balance Global Density "
|
|
|
|
<< Session::getRoutingGauge()->getRoutingLayer(depth)->getName() << endl;
|
|
|
|
|
|
|
|
GCellDensitySet queue ( depth, *(_gcellGrid->getGCellVector()) );
|
|
|
|
GCell::SetIndex invalidateds;
|
|
|
|
|
|
|
|
bool optimized = true;
|
|
|
|
while ( optimized ) {
|
|
|
|
Session::revalidate();
|
|
|
|
optimized = false;
|
|
|
|
queue.requeue();
|
|
|
|
|
|
|
|
std::set<GCell*,GCell::CompareByKey>::const_iterator igcell = queue.getGCells().begin();
|
|
|
|
size_t i = 0;
|
|
|
|
for ( ; igcell!=queue.getGCells().end() ; ++igcell, ++i ) {
|
2016-06-11 14:56:12 -05:00
|
|
|
cdebug_log(149,0) << "_balance: [" << depth << "]:"
|
2013-12-03 18:58:58 -06:00
|
|
|
<< (*igcell)->getDensity(depth) << " " << *igcell << endl;
|
|
|
|
|
|
|
|
if (not (*igcell)->isSaturated(depth)) {
|
2016-06-11 14:56:12 -05:00
|
|
|
cdebug_log(149,0) << "STOP desaturated: @" << i << " " << *igcell << endl;
|
2013-12-03 18:58:58 -06:00
|
|
|
for ( ; igcell!=queue.getGCells().end() ; ++igcell ) {
|
|
|
|
if ((*igcell)->isSaturated(depth)) {
|
|
|
|
cparanoid << Error( "Still saturated: @%d %s", i, getString(*igcell).c_str() ) << endl;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
optimized = (*igcell)->stepBalance( depth, invalidateds );
|
|
|
|
if (optimized) {
|
|
|
|
for ( GCell::SetIndex::iterator igcell=invalidateds.begin() ; igcell!=invalidateds.end() ; ++igcell ) {
|
|
|
|
queue.unqueue( *igcell );
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
Session::close();
|
|
|
|
stopMeasures();
|
|
|
|
printMeasures( "balance" );
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
void KatabaticEngine::balanceGlobalDensity ()
|
|
|
|
{
|
2016-06-11 14:56:12 -05:00
|
|
|
cdebug_log(9000,0) << "Deter| Balance Global Density" << endl;
|
2013-12-03 18:58:58 -06:00
|
|
|
|
|
|
|
//_balanceGlobalDensity( 1 ); // metal2
|
|
|
|
//_balanceGlobalDensity( 2 ); // metal3
|
|
|
|
|
|
|
|
set<Net*> globalNets;
|
|
|
|
GCell::SetIndex invalidateds;
|
|
|
|
|
|
|
|
Session::open( this );
|
|
|
|
|
|
|
|
vector<AutoSegment*> segments;
|
|
|
|
|
|
|
|
AutoSegmentLut::iterator ilut = _autoSegmentLut.begin();
|
|
|
|
for ( ; ilut!=_autoSegmentLut.end() ; ++ilut ) {
|
|
|
|
AutoSegment* segment = (*ilut).second;
|
|
|
|
|
Implementation of pre-routing support (for clock-tree compliance).
* New: In Katabatic, in <AutoContact>, this class is no longer derived
from ExtentionGo. With the simplificated AutoContacts, there is no
reason to do so, and it will save some QuadTree insertions/deletions.
New factory function AutoContact::createFrom(Contact*) which try to
build an AutoContact on top of a Hurricane::Contact. Of course that
base contact *must fit* into one of the predefined Contact
configurations (Terminal, Turn, HTee or VTee).
NOTE: This implies that the pre-routed segments & contacts *are*
correctly articulated, which is not the case when a Cell is read
from disk in "ap" format. The pre-routing feature must be used for
now without any re-read from disk. We will implement a re-articulating
pre-process in the future.
* Change: In Katabatic, in <AutoContact> derived classes, the ::updateCache()
method now display an accurate error message if a segment is connected
but has no AutoSegment conterpart (i.e. the lookup fails).
* New: In Katabatic, in <AutoSegment>, the ::computeOptimal() method is
short-circuited for pre-routed segments, the optimal axis position is
considered to be the one it is currently on (i.e. we trust the designer).
* New: In Katabatic, in <KatabaticEngine>, the ::loadGlobalRouting()
method now accept a map of excluded nets (same as Knik). This map is
the one of pre-routed nets.
* New: In Katabatic, in layer assignment, do not try to displace fixed
segments...
* New: In Katabatic, in <AutoSegment>, new flag SegUserDefined and related
methods to know if a segment comes from the global router (Knik) or
is pre-routed (supplied by the user).
* New: In Kite, In <BuildPowerRails>, support (exclusion) for pre-routed
nets.
* New: In Kite, In <GraphicKiteEngine> new menu entry for running the
router on pre-routed nets ("Detailed Pre-Route"), also integrated
in the all-on-one route command.
* New: In Kite, In KiteEngine, new method ::_initDataBase() that group
all the initialisation steps. It is a mix of calls between Knik and
Kite initializations which are intertwinneds (may have to devellop
a shared common base at a later point). It creates the Knik grid,
then the Katabatic grid, then load pre-routed wires and power rails
and protect isolated RoutingPads.
Add support for a map of pre-routed nets (to be excluded for
Knik calls).
The method "::run()" now uses function flags, firstly to know if
it is managing pre-routed wires or general purposes ones.
* New: In Kite, in <NegociateWindow>, the "::run()" methods has now two
modes. The normal one and the 'KtPreRoutedStage' that is for routing
pre-routed nets. When in pre-route stage, the wires are fixed at the
end of this step.
* New: In Kite, in <TrackElement> add decorator for AutoSegment
isUsedDefined().
* New: In Kite, in <TrackSegment>, the various ::canDogleg() methods
returns false for a pre-routed (user-defined segment).
* New: In Kite, in PyKiteEngine, added new method runNegociatePreRouted().
2014-06-21 13:16:47 -05:00
|
|
|
if (segment->isLocal() or segment->isFixed()) continue;
|
2013-12-03 18:58:58 -06:00
|
|
|
if (not segment->isCanonical()) continue;
|
|
|
|
|
|
|
|
segments.push_back( segment );
|
|
|
|
}
|
|
|
|
|
|
|
|
// Sort on id before moving to ensure determinism.
|
|
|
|
sort( segments.begin(), segments.end(), AutoSegment::CompareId() );
|
|
|
|
|
|
|
|
for ( size_t i=0 ; i<segments.size() ; ++i ) {
|
|
|
|
// Hard-coded: reserve 3 tracks (1/20 * 3).
|
|
|
|
if (segments[i]->canMoveULeft(0.10)) {
|
|
|
|
getGCellGrid()->getKatabatic()->moveULeft(segments[i],globalNets,invalidateds);
|
|
|
|
} else if (segments[i]->canMoveURight(0.10)) {
|
|
|
|
getGCellGrid()->getKatabatic()->moveURight(segments[i],globalNets,invalidateds);
|
|
|
|
}
|
|
|
|
|
|
|
|
for ( GCell::SetIndex::iterator igcell=invalidateds.begin() ; igcell!=invalidateds.end() ; ++igcell ) {
|
|
|
|
(*igcell)->updateDensity();
|
|
|
|
}
|
|
|
|
invalidateds.clear();
|
|
|
|
Session::revalidate();
|
|
|
|
}
|
|
|
|
|
|
|
|
Session::close();
|
|
|
|
}
|
|
|
|
|
|
|
|
|
2010-03-09 09:24:29 -06:00
|
|
|
void KatabaticEngine::layerAssign ( unsigned int method )
|
|
|
|
{
|
2016-06-11 14:56:12 -05:00
|
|
|
cdebug_log(9000,0) << "Deter| Layer Assignment" << endl;
|
2013-12-03 18:58:58 -06:00
|
|
|
|
2010-03-09 09:24:29 -06:00
|
|
|
set<Net*> globalNets;
|
|
|
|
|
|
|
|
unsigned long total = 0;
|
|
|
|
unsigned long global = 0;
|
|
|
|
|
2013-12-03 18:58:58 -06:00
|
|
|
startMeasures();
|
|
|
|
Session::open( this );
|
2010-03-09 09:24:29 -06:00
|
|
|
|
Improved UpdateSession & exception catching. Start of RoutingGauge implem.
Miscellaneous:
* Change: In <crlcore>, in display.conf use the same display threshold
for both METAL2 & METAL3.
In alliance.conf, the side of VIAs in the gauge is 2l (not 3l).
In kite.conf, separate edge densities for H/V.
* Change: In <Cell>, in flattenNets() use flag as argument, not a
boolean. Do not create rings for clock or supply nets.
* Change: In <DeepNet>, in _createRoutingPads() do not create rings
for clock or supply net (duplicated policy as in Cell::flattenNets()).
* Bug: In <ControllerWidget>, at last find the bad signal disconnect
that was causing ungraceful messages.
* Change: In <knik>, in Edge display occupancy/capacity in the string
name. Improved display progress and debugging capabilities.
Improved exception catch & breakpoint managment:
* Bug: In <PaletteWidget>, in updateExtensions() replace the calls to
deleteLayer() by delete. This cause the widget to be immediatly
erased instead of waiting for the event queue to be completly
processed. This was causing the widget to be left in a incoherent
state when stoping at a breakpoint.
* Bug: In <BreakpointWidget>, in execNoModal(), flush the main event
loop (QApplication::flush()) *before* lauching the *local* event
loop. This is to ensure all widgets are in their final state when
waiting (especially <PaletteWidget>).
* Change: In <ExceptionWidget>, new method catchAllWrapper() to
execute any std::function< void() > function/method with a "try"/
"catch" wraparound and lauch the widget in case something is catch.
* New: In <hurricane>, support for a oberver pattern, backported from
<katabatic> with an Obervable capable of being linked to any
number of Obervers.
* New: In <Cell>, made it observable to detect Cell change, currently
emit two kind of signals:
- Cell::CellAboutToChange : *before* any change.
- Cell::CellChanged : *after* the change has been completed.
* New: In <UpdateSession>, in Go::invalidate() add the Cell owning the
Go to the UPDATOR_STACK (of course the cell is added only once).
In addition, when the Cell is added, send a notification of
Cell::CellAboutToChange to all it's observers. The slave instances
are also invalidated.
Conversely in UpdateSession::_preDestroy() for each invalidated
Cell send a Cell::CellChanged notification to all observer.
The UPDATOR_STACK has been slightly amended to accept Cell which
are not Gos. Prior to this, the Cell where completly excluded from
the UpdateSession mechanism, so it's instances where never actualised
of anything referring to the Cell for that matter.
Note: we use two different mechanisms to transmit a Cell change,
observers and the slave instance map. I think at some point it
should be unificated.
* Change: In <CellViewer>, make it a Cell observer to redraw when the
cell is modificated (also update the palette).
Uses the catchAllWrapper() to protect all critical actions.
* Change: In <GraphicTool>, no longer need of cellPreModificated and
cellPostModificated signals. Now done through the Cell obersvers.
* Change: In <mauka>, <etesian> & <kite> now uses the catchAllWrapper
method for protection (need to split methods in two, to be able
to pass it as argument). No longer emit cellPreModificated and
cellPostModificated.
Support for RoutingGauge in P&R:
* Bug: In <placeandroute.py>, the connection from the internal power
ring to the connectors was not done correctly. Wrong contact layers
leading to a gap.
* Change: In <BuildPowerRails>, detection of the corona signals based
on how the "pck_px" pad is connected. No longer based on name
matching.
* Change: In <placeandroute.py>, support for 2 routing metal only
(3 metal in the technology).
* Change: In <katabatic> & <kite> support for a "top layer" limitation
on the routing gauge, this allows to use only two routing metals
(METAL2 & METAL3). Work in progress.
2014-04-20 12:25:08 -05:00
|
|
|
if (Session::getAllowedDepth() >= 3) {
|
|
|
|
switch ( method ) {
|
|
|
|
case EngineLayerAssignByLength: _layerAssignByLength( total, global, globalNets ); break;
|
|
|
|
case EngineLayerAssignByTrunk: _layerAssignByTrunk ( total, global, globalNets ); break;
|
|
|
|
case EngineNoNetLayerAssign: break;
|
|
|
|
default:
|
|
|
|
stopMeasures();
|
|
|
|
Session::close();
|
|
|
|
throw Error( badMethod
|
|
|
|
, "Katabatic::layerAssign()"
|
|
|
|
, method
|
|
|
|
, getString(_cell).c_str()
|
|
|
|
);
|
|
|
|
}
|
|
|
|
|
|
|
|
globalNets.clear();
|
2013-12-03 18:58:58 -06:00
|
|
|
Session::revalidate();
|
Improved UpdateSession & exception catching. Start of RoutingGauge implem.
Miscellaneous:
* Change: In <crlcore>, in display.conf use the same display threshold
for both METAL2 & METAL3.
In alliance.conf, the side of VIAs in the gauge is 2l (not 3l).
In kite.conf, separate edge densities for H/V.
* Change: In <Cell>, in flattenNets() use flag as argument, not a
boolean. Do not create rings for clock or supply nets.
* Change: In <DeepNet>, in _createRoutingPads() do not create rings
for clock or supply net (duplicated policy as in Cell::flattenNets()).
* Bug: In <ControllerWidget>, at last find the bad signal disconnect
that was causing ungraceful messages.
* Change: In <knik>, in Edge display occupancy/capacity in the string
name. Improved display progress and debugging capabilities.
Improved exception catch & breakpoint managment:
* Bug: In <PaletteWidget>, in updateExtensions() replace the calls to
deleteLayer() by delete. This cause the widget to be immediatly
erased instead of waiting for the event queue to be completly
processed. This was causing the widget to be left in a incoherent
state when stoping at a breakpoint.
* Bug: In <BreakpointWidget>, in execNoModal(), flush the main event
loop (QApplication::flush()) *before* lauching the *local* event
loop. This is to ensure all widgets are in their final state when
waiting (especially <PaletteWidget>).
* Change: In <ExceptionWidget>, new method catchAllWrapper() to
execute any std::function< void() > function/method with a "try"/
"catch" wraparound and lauch the widget in case something is catch.
* New: In <hurricane>, support for a oberver pattern, backported from
<katabatic> with an Obervable capable of being linked to any
number of Obervers.
* New: In <Cell>, made it observable to detect Cell change, currently
emit two kind of signals:
- Cell::CellAboutToChange : *before* any change.
- Cell::CellChanged : *after* the change has been completed.
* New: In <UpdateSession>, in Go::invalidate() add the Cell owning the
Go to the UPDATOR_STACK (of course the cell is added only once).
In addition, when the Cell is added, send a notification of
Cell::CellAboutToChange to all it's observers. The slave instances
are also invalidated.
Conversely in UpdateSession::_preDestroy() for each invalidated
Cell send a Cell::CellChanged notification to all observer.
The UPDATOR_STACK has been slightly amended to accept Cell which
are not Gos. Prior to this, the Cell where completly excluded from
the UpdateSession mechanism, so it's instances where never actualised
of anything referring to the Cell for that matter.
Note: we use two different mechanisms to transmit a Cell change,
observers and the slave instance map. I think at some point it
should be unificated.
* Change: In <CellViewer>, make it a Cell observer to redraw when the
cell is modificated (also update the palette).
Uses the catchAllWrapper() to protect all critical actions.
* Change: In <GraphicTool>, no longer need of cellPreModificated and
cellPostModificated signals. Now done through the Cell obersvers.
* Change: In <mauka>, <etesian> & <kite> now uses the catchAllWrapper
method for protection (need to split methods in two, to be able
to pass it as argument). No longer emit cellPreModificated and
cellPostModificated.
Support for RoutingGauge in P&R:
* Bug: In <placeandroute.py>, the connection from the internal power
ring to the connectors was not done correctly. Wrong contact layers
leading to a gap.
* Change: In <BuildPowerRails>, detection of the corona signals based
on how the "pck_px" pad is connected. No longer based on name
matching.
* Change: In <placeandroute.py>, support for 2 routing metal only
(3 metal in the technology).
* Change: In <katabatic> & <kite> support for a "top layer" limitation
on the routing gauge, this allows to use only two routing metals
(METAL2 & METAL3). Work in progress.
2014-04-20 12:25:08 -05:00
|
|
|
|
|
|
|
if (getConfiguration()->getAllowedDepth() > 2) {
|
|
|
|
//for ( int i=0 ; i < 3 ; i++ ) {
|
|
|
|
for ( size_t depth=1 ; depth <= getConfiguration()->getAllowedDepth()-2; ++depth ) {
|
|
|
|
_desaturate( depth, globalNets, total, global );
|
|
|
|
if ( (depth > 1) and ((depth-1)%2 == 1) ) Session::revalidate();
|
|
|
|
}
|
|
|
|
|
|
|
|
globalNets.clear ();
|
|
|
|
|
|
|
|
// if (not _gcellGrid->updateDensity()) break;
|
|
|
|
//}
|
|
|
|
Session::revalidate();
|
|
|
|
}
|
|
|
|
|
2010-03-09 09:24:29 -06:00
|
|
|
#if defined(CHECK_DATABASE)
|
Improved UpdateSession & exception catching. Start of RoutingGauge implem.
Miscellaneous:
* Change: In <crlcore>, in display.conf use the same display threshold
for both METAL2 & METAL3.
In alliance.conf, the side of VIAs in the gauge is 2l (not 3l).
In kite.conf, separate edge densities for H/V.
* Change: In <Cell>, in flattenNets() use flag as argument, not a
boolean. Do not create rings for clock or supply nets.
* Change: In <DeepNet>, in _createRoutingPads() do not create rings
for clock or supply net (duplicated policy as in Cell::flattenNets()).
* Bug: In <ControllerWidget>, at last find the bad signal disconnect
that was causing ungraceful messages.
* Change: In <knik>, in Edge display occupancy/capacity in the string
name. Improved display progress and debugging capabilities.
Improved exception catch & breakpoint managment:
* Bug: In <PaletteWidget>, in updateExtensions() replace the calls to
deleteLayer() by delete. This cause the widget to be immediatly
erased instead of waiting for the event queue to be completly
processed. This was causing the widget to be left in a incoherent
state when stoping at a breakpoint.
* Bug: In <BreakpointWidget>, in execNoModal(), flush the main event
loop (QApplication::flush()) *before* lauching the *local* event
loop. This is to ensure all widgets are in their final state when
waiting (especially <PaletteWidget>).
* Change: In <ExceptionWidget>, new method catchAllWrapper() to
execute any std::function< void() > function/method with a "try"/
"catch" wraparound and lauch the widget in case something is catch.
* New: In <hurricane>, support for a oberver pattern, backported from
<katabatic> with an Obervable capable of being linked to any
number of Obervers.
* New: In <Cell>, made it observable to detect Cell change, currently
emit two kind of signals:
- Cell::CellAboutToChange : *before* any change.
- Cell::CellChanged : *after* the change has been completed.
* New: In <UpdateSession>, in Go::invalidate() add the Cell owning the
Go to the UPDATOR_STACK (of course the cell is added only once).
In addition, when the Cell is added, send a notification of
Cell::CellAboutToChange to all it's observers. The slave instances
are also invalidated.
Conversely in UpdateSession::_preDestroy() for each invalidated
Cell send a Cell::CellChanged notification to all observer.
The UPDATOR_STACK has been slightly amended to accept Cell which
are not Gos. Prior to this, the Cell where completly excluded from
the UpdateSession mechanism, so it's instances where never actualised
of anything referring to the Cell for that matter.
Note: we use two different mechanisms to transmit a Cell change,
observers and the slave instance map. I think at some point it
should be unificated.
* Change: In <CellViewer>, make it a Cell observer to redraw when the
cell is modificated (also update the palette).
Uses the catchAllWrapper() to protect all critical actions.
* Change: In <GraphicTool>, no longer need of cellPreModificated and
cellPostModificated signals. Now done through the Cell obersvers.
* Change: In <mauka>, <etesian> & <kite> now uses the catchAllWrapper
method for protection (need to split methods in two, to be able
to pass it as argument). No longer emit cellPreModificated and
cellPostModificated.
Support for RoutingGauge in P&R:
* Bug: In <placeandroute.py>, the connection from the internal power
ring to the connectors was not done correctly. Wrong contact layers
leading to a gap.
* Change: In <BuildPowerRails>, detection of the corona signals based
on how the "pck_px" pad is connected. No longer based on name
matching.
* Change: In <placeandroute.py>, support for 2 routing metal only
(3 metal in the technology).
* Change: In <katabatic> & <kite> support for a "top layer" limitation
on the routing gauge, this allows to use only two routing metals
(METAL2 & METAL3). Work in progress.
2014-04-20 12:25:08 -05:00
|
|
|
_check( "after layer assignment" );
|
2010-03-09 09:24:29 -06:00
|
|
|
#endif
|
Improved UpdateSession & exception catching. Start of RoutingGauge implem.
Miscellaneous:
* Change: In <crlcore>, in display.conf use the same display threshold
for both METAL2 & METAL3.
In alliance.conf, the side of VIAs in the gauge is 2l (not 3l).
In kite.conf, separate edge densities for H/V.
* Change: In <Cell>, in flattenNets() use flag as argument, not a
boolean. Do not create rings for clock or supply nets.
* Change: In <DeepNet>, in _createRoutingPads() do not create rings
for clock or supply net (duplicated policy as in Cell::flattenNets()).
* Bug: In <ControllerWidget>, at last find the bad signal disconnect
that was causing ungraceful messages.
* Change: In <knik>, in Edge display occupancy/capacity in the string
name. Improved display progress and debugging capabilities.
Improved exception catch & breakpoint managment:
* Bug: In <PaletteWidget>, in updateExtensions() replace the calls to
deleteLayer() by delete. This cause the widget to be immediatly
erased instead of waiting for the event queue to be completly
processed. This was causing the widget to be left in a incoherent
state when stoping at a breakpoint.
* Bug: In <BreakpointWidget>, in execNoModal(), flush the main event
loop (QApplication::flush()) *before* lauching the *local* event
loop. This is to ensure all widgets are in their final state when
waiting (especially <PaletteWidget>).
* Change: In <ExceptionWidget>, new method catchAllWrapper() to
execute any std::function< void() > function/method with a "try"/
"catch" wraparound and lauch the widget in case something is catch.
* New: In <hurricane>, support for a oberver pattern, backported from
<katabatic> with an Obervable capable of being linked to any
number of Obervers.
* New: In <Cell>, made it observable to detect Cell change, currently
emit two kind of signals:
- Cell::CellAboutToChange : *before* any change.
- Cell::CellChanged : *after* the change has been completed.
* New: In <UpdateSession>, in Go::invalidate() add the Cell owning the
Go to the UPDATOR_STACK (of course the cell is added only once).
In addition, when the Cell is added, send a notification of
Cell::CellAboutToChange to all it's observers. The slave instances
are also invalidated.
Conversely in UpdateSession::_preDestroy() for each invalidated
Cell send a Cell::CellChanged notification to all observer.
The UPDATOR_STACK has been slightly amended to accept Cell which
are not Gos. Prior to this, the Cell where completly excluded from
the UpdateSession mechanism, so it's instances where never actualised
of anything referring to the Cell for that matter.
Note: we use two different mechanisms to transmit a Cell change,
observers and the slave instance map. I think at some point it
should be unificated.
* Change: In <CellViewer>, make it a Cell observer to redraw when the
cell is modificated (also update the palette).
Uses the catchAllWrapper() to protect all critical actions.
* Change: In <GraphicTool>, no longer need of cellPreModificated and
cellPostModificated signals. Now done through the Cell obersvers.
* Change: In <mauka>, <etesian> & <kite> now uses the catchAllWrapper
method for protection (need to split methods in two, to be able
to pass it as argument). No longer emit cellPreModificated and
cellPostModificated.
Support for RoutingGauge in P&R:
* Bug: In <placeandroute.py>, the connection from the internal power
ring to the connectors was not done correctly. Wrong contact layers
leading to a gap.
* Change: In <BuildPowerRails>, detection of the corona signals based
on how the "pck_px" pad is connected. No longer based on name
matching.
* Change: In <placeandroute.py>, support for 2 routing metal only
(3 metal in the technology).
* Change: In <katabatic> & <kite> support for a "top layer" limitation
on the routing gauge, this allows to use only two routing metals
(METAL2 & METAL3). Work in progress.
2014-04-20 12:25:08 -05:00
|
|
|
|
|
|
|
Session::setKatabaticFlags( EngineWarnOnGCellOverload );
|
|
|
|
}
|
|
|
|
|
2015-06-06 11:22:48 -05:00
|
|
|
_gcellGrid->checkDensity();
|
2013-12-03 18:58:58 -06:00
|
|
|
Session::close();
|
2010-03-09 09:24:29 -06:00
|
|
|
|
2013-12-03 18:58:58 -06:00
|
|
|
stopMeasures();
|
|
|
|
printMeasures( "assign" );
|
2010-03-09 09:24:29 -06:00
|
|
|
|
2013-12-03 18:58:58 -06:00
|
|
|
// cmess2 << " - Total segments : " << total << endl;
|
|
|
|
// cmess2 << " - Global segments : " << global << endl;
|
|
|
|
// cmess2 << " - Ratio : "
|
|
|
|
// << ((float)global/(float)total)*100.0 << "%." << endl;
|
2010-03-09 09:24:29 -06:00
|
|
|
}
|
|
|
|
|
|
|
|
|
2013-12-03 18:58:58 -06:00
|
|
|
} // Katabatic namespace.
|