coriolis/bootstrap/builder/Builder.py

491 lines
20 KiB
Python
Raw Normal View History

# -*- mode:Python -*-
#
# This file is part of the Coriolis Software.
# Copyright (c) UPMC/LIP6 2008-2018, All Rights Reserved
#
# +-----------------------------------------------------------------+
# | C O R I O L I S |
# | C o r i o l i s / C h a m s B u i l d e r |
# | |
# | Author : Jean-Paul Chaput |
# | E-mail : Jean-Paul.Chaput@asim.lip6.fr |
# | =============================================================== |
# | Python : "./builder/Builder.py" |
# +-----------------------------------------------------------------+
import sys
import re
import os
import os.path
import datetime
import subprocess
from . import ErrorMessage
from Project import Project
from Configuration import Configuration
class Builder:
def __init__ ( self ):
self._conf = Configuration()
self._quiet = False
self._rmBuild = False
self._doBuild = True
self._noCache = False
self._ninja = False
self._clang = False
self._noSystemBoost = False
self._macports = False
self._devtoolset = 0
Groudwork for routing density driven placement. Compliance with clang 5.0.1. This commit contains two set of features that should have been commited separately. 1. Compliance with clang 5.0.1, tested with the RedHat collection llvm-toolset-7. This allow Coriolis to be compiled under Darwin (MacOS) with Xcode & macports. The bootstrap install system has been modificated accordingly. 2. The basic support for routing density driven placement. Related features are: * Bloat property. Each Occurrence of an Instance can be individually bloated. This property not attached to any tool to allow the placer and router to share it as wanted. Nevertheless, it is defined in Etesian. * BloatProfile in Katana, add individual Bloat properties to Instances occurrences based on the East & North overflowed edges of each GCell. * Support in ToolEngine for a "pass number" of a tool. This pass number is mainly used to make "per pass" measurements. The MeasureSet system is improved accordingly to support multiple values of a same measure. * Embryo of "P&R Conductor" to perform the place & route loop until the design is successfully placed. May be the first brick of a Silicon Compiler. * Change: In boostrap/FindBoostrap.cmake, in setup_boost(), added tag to the python component for macport (ex: python27). * Change: In boostrap/build.conf, put etesian before anabatic for instance occurrence BloatProperty dependency. Added option support for the "llvm-toolset-7" collection to build against clang 5.0.1. * Bug: In Hurricane::getRecord( const pair<T,U>& ), the getSlot<> templates for first & second arguments must be called with <const T> and <const U> as the pair itself is const (and not simply <T> & <U>). * Change: In Hurricane::getSlot() temlate, only use "string" arguments and not const string&, simpler for template argument deduction. * Bug: In Hurricane::AnalogCellExtension, the StandardPrivateProperty<> template has a static member "_name". Clang did show that the template for this static number has to be put inside the namespace where the template *is defined* (i.e. Hurricane) instead of the namespace where it is instanciated (i.e. Analog). * Bug: In Isobar, Matrix_FromListOfList(), PyInt_AsPlacementStatus() must be put outside the C linkage back in the Isobar C++ namespace (clang). * Bug: In Hurricane::DBo::~DBo, and derived add a throw() specification (clang). * Bug: In Hurricane::RegularLayer::getEnclosure() & setEnclosure(), change signature so it matches the one of the base class (clang). * Bug: In Hurricane::CellPrinter, use double brackets for initializer list (clang). * Change: In Hurricane::Breakpoint, reverse the meaning of the error level. Only error level *lesser or equal* than the stop level will be enabled. * Bug: In CRL/python/helpers/__init__.loadUserSettings(), must put the current working directory in the sys.path as in certain configuration it may not be included. * Bug: In CRL::ApDriver, DumpSegments(), no longer generate segments when encountering a RoutingPad on a top-level Pin Occurrence. The segment was generated in the wrong direction, creating DRC violations on the "mips_core_flat" example. * Change: In CRL::Measures, partial re-design of the measurements management. Now, each kind of measure can accept multiple values put in a vector. The index is intented to match a tool run number. * Change: In CRL::Histogram, add support for multiple sets of datas, indexeds with tool run number. * Change: In CRL::ToolEngine, add support for multiple pass number, add addMeasure<> templates for the various data-types. * Change: In CRL::gdsDriver & CRL::gdsParser(), comment out unused GDS record name constants. * New: Etesian::BloatProperty, property to attach to Instance occurrences that contains the extra number of pitch to add to the cell width. * Bug: In AutoSegment::CompareByDepthLength, the segment length comparison was wrong, it was always returning true, which broke the "strick weak ordering" of the comparison. This was producing a core-dump in GCell::updateDensity() when sorting a vector<>. The end() iterator was being dereferenced, leading to the problem. * Bug: In Katana::DataSymmetric::checkPairing(), the test for segments whose axis is perpandicular to the symmetry axis was wrong ("!=" instead of "-"). * New: In Katana/GlobalRoute, new ::selectSegments(), selectOverloadedgcells() and selectBloatedInstances() to automatically select segments from overloaded edges, overloaded GCells and bloated cells. * Change: In KatanaEngine, return a more detailed success state, distinguish between global and detailed. Add support for multiple routing iterations. * New: In cumulus/python/plugins/ConductorPlugin.py, embryo of routing driven placement.
2019-12-08 18:57:44 -06:00
self._llvmtoolset = 0
self._qt5 = False
self._openmp = False
self._enableShared = "ON"
self._enableDoc = "OFF"
self._checkDatabase = "OFF"
self._checkDeterminism = "OFF"
self._verboseMakefile = "OFF"
self._makeArguments = []
self._environment = os.environ
return
def __setattr__ ( self, attribute, value ):
if attribute[0] == "_":
self.__dict__[attribute] = value
return
if attribute in self._conf.getAllIds(): setattr( self._conf, attribute, value )
2013-11-05 07:27:24 -06:00
if attribute == "quiet": self._quiet = value
elif attribute == "rmBuild": self._rmBuild = value
elif attribute == "doBuild": self._doBuild = value
elif attribute == "noCache": self._noCache = value
elif attribute == "ninja": self._ninja = value
elif attribute == "clang": self._clang = value
elif attribute == "macports":
self._macports = value
if value: self._noSystemBoost = True
elif attribute == "devtoolset":
self._devtoolset = value
if value: self._noSystemBoost = True
Groudwork for routing density driven placement. Compliance with clang 5.0.1. This commit contains two set of features that should have been commited separately. 1. Compliance with clang 5.0.1, tested with the RedHat collection llvm-toolset-7. This allow Coriolis to be compiled under Darwin (MacOS) with Xcode & macports. The bootstrap install system has been modificated accordingly. 2. The basic support for routing density driven placement. Related features are: * Bloat property. Each Occurrence of an Instance can be individually bloated. This property not attached to any tool to allow the placer and router to share it as wanted. Nevertheless, it is defined in Etesian. * BloatProfile in Katana, add individual Bloat properties to Instances occurrences based on the East & North overflowed edges of each GCell. * Support in ToolEngine for a "pass number" of a tool. This pass number is mainly used to make "per pass" measurements. The MeasureSet system is improved accordingly to support multiple values of a same measure. * Embryo of "P&R Conductor" to perform the place & route loop until the design is successfully placed. May be the first brick of a Silicon Compiler. * Change: In boostrap/FindBoostrap.cmake, in setup_boost(), added tag to the python component for macport (ex: python27). * Change: In boostrap/build.conf, put etesian before anabatic for instance occurrence BloatProperty dependency. Added option support for the "llvm-toolset-7" collection to build against clang 5.0.1. * Bug: In Hurricane::getRecord( const pair<T,U>& ), the getSlot<> templates for first & second arguments must be called with <const T> and <const U> as the pair itself is const (and not simply <T> & <U>). * Change: In Hurricane::getSlot() temlate, only use "string" arguments and not const string&, simpler for template argument deduction. * Bug: In Hurricane::AnalogCellExtension, the StandardPrivateProperty<> template has a static member "_name". Clang did show that the template for this static number has to be put inside the namespace where the template *is defined* (i.e. Hurricane) instead of the namespace where it is instanciated (i.e. Analog). * Bug: In Isobar, Matrix_FromListOfList(), PyInt_AsPlacementStatus() must be put outside the C linkage back in the Isobar C++ namespace (clang). * Bug: In Hurricane::DBo::~DBo, and derived add a throw() specification (clang). * Bug: In Hurricane::RegularLayer::getEnclosure() & setEnclosure(), change signature so it matches the one of the base class (clang). * Bug: In Hurricane::CellPrinter, use double brackets for initializer list (clang). * Change: In Hurricane::Breakpoint, reverse the meaning of the error level. Only error level *lesser or equal* than the stop level will be enabled. * Bug: In CRL/python/helpers/__init__.loadUserSettings(), must put the current working directory in the sys.path as in certain configuration it may not be included. * Bug: In CRL::ApDriver, DumpSegments(), no longer generate segments when encountering a RoutingPad on a top-level Pin Occurrence. The segment was generated in the wrong direction, creating DRC violations on the "mips_core_flat" example. * Change: In CRL::Measures, partial re-design of the measurements management. Now, each kind of measure can accept multiple values put in a vector. The index is intented to match a tool run number. * Change: In CRL::Histogram, add support for multiple sets of datas, indexeds with tool run number. * Change: In CRL::ToolEngine, add support for multiple pass number, add addMeasure<> templates for the various data-types. * Change: In CRL::gdsDriver & CRL::gdsParser(), comment out unused GDS record name constants. * New: Etesian::BloatProperty, property to attach to Instance occurrences that contains the extra number of pitch to add to the cell width. * Bug: In AutoSegment::CompareByDepthLength, the segment length comparison was wrong, it was always returning true, which broke the "strick weak ordering" of the comparison. This was producing a core-dump in GCell::updateDensity() when sorting a vector<>. The end() iterator was being dereferenced, leading to the problem. * Bug: In Katana::DataSymmetric::checkPairing(), the test for segments whose axis is perpandicular to the symmetry axis was wrong ("!=" instead of "-"). * New: In Katana/GlobalRoute, new ::selectSegments(), selectOverloadedgcells() and selectBloatedInstances() to automatically select segments from overloaded edges, overloaded GCells and bloated cells. * Change: In KatanaEngine, return a more detailed success state, distinguish between global and detailed. Add support for multiple routing iterations. * New: In cumulus/python/plugins/ConductorPlugin.py, embryo of routing driven placement.
2019-12-08 18:57:44 -06:00
elif attribute == "llvmtoolset":
self._llvmtoolset = value
elif attribute == "qt5": self._qt5 = value
elif attribute == "openmp": self._openmp = value
2013-11-05 07:27:24 -06:00
elif attribute == "enableDoc": self._enableDoc = value
elif attribute == "enableShared": self._enableShared = value
elif attribute == "checkDatabase": self._checkDatabase = value
elif attribute == "checkDeterminism": self._checkDeterminism = value
elif attribute == "verboseMakefile": self._verboseMakefile = value
elif attribute == "makeArguments": self._makeArguments = value.split ()
return
def __getattr__ ( self, attribute ):
if attribute[0] != "_":
if attribute == 'conf': return self._conf
if attribute in self._conf.getAllIds():
return getattr( self._conf, attribute )
if not self.__dict__.has_key(attribute):
raise ErrorMessage( 1, 'Builder has no attribute <%s>.'%attribute )
return self.__dict__[attribute]
def _guessGitHash ( self, project ):
self.gitHash = 'x'
os.chdir ( self.sourceDir+'/'+project.getName() )
command = [ 'git', 'log', '--pretty=format:%h', '-n', '1']
self.gitHash = subprocess.Popen ( command, stdout=subprocess.PIPE ).stdout.readlines()[0]
return
def _configure ( self, fileIn, fileOut ):
fdFileIn = open ( fileIn , "r" )
fdFileOut = open ( fileOut, "w" )
for line in fdFileIn.readlines():
stable = False
substituted0 = line
while not stable:
substituted1 = re.sub ( r"@revdate@" , self.revDate, substituted0 )
substituted1 = re.sub ( r"@githash@" , self.gitHash, substituted1 )
substituted1 = re.sub ( r"@coriolisTop@", "/usr" , substituted1 )
if substituted0 == substituted1: stable = True
else: substituted0 = substituted1
fdFileOut.write ( substituted0 )
fdFileIn.close ()
fdFileOut.close ()
return
def _doSpec ( self ):
self._configure ( self.specFileIn, self.specFile )
return
def _doDebChangelog ( self ):
self._configure ( self.debChangelogIn, self.debChangelog )
return
def _execute ( self, command, error ):
Groudwork for routing density driven placement. Compliance with clang 5.0.1. This commit contains two set of features that should have been commited separately. 1. Compliance with clang 5.0.1, tested with the RedHat collection llvm-toolset-7. This allow Coriolis to be compiled under Darwin (MacOS) with Xcode & macports. The bootstrap install system has been modificated accordingly. 2. The basic support for routing density driven placement. Related features are: * Bloat property. Each Occurrence of an Instance can be individually bloated. This property not attached to any tool to allow the placer and router to share it as wanted. Nevertheless, it is defined in Etesian. * BloatProfile in Katana, add individual Bloat properties to Instances occurrences based on the East & North overflowed edges of each GCell. * Support in ToolEngine for a "pass number" of a tool. This pass number is mainly used to make "per pass" measurements. The MeasureSet system is improved accordingly to support multiple values of a same measure. * Embryo of "P&R Conductor" to perform the place & route loop until the design is successfully placed. May be the first brick of a Silicon Compiler. * Change: In boostrap/FindBoostrap.cmake, in setup_boost(), added tag to the python component for macport (ex: python27). * Change: In boostrap/build.conf, put etesian before anabatic for instance occurrence BloatProperty dependency. Added option support for the "llvm-toolset-7" collection to build against clang 5.0.1. * Bug: In Hurricane::getRecord( const pair<T,U>& ), the getSlot<> templates for first & second arguments must be called with <const T> and <const U> as the pair itself is const (and not simply <T> & <U>). * Change: In Hurricane::getSlot() temlate, only use "string" arguments and not const string&, simpler for template argument deduction. * Bug: In Hurricane::AnalogCellExtension, the StandardPrivateProperty<> template has a static member "_name". Clang did show that the template for this static number has to be put inside the namespace where the template *is defined* (i.e. Hurricane) instead of the namespace where it is instanciated (i.e. Analog). * Bug: In Isobar, Matrix_FromListOfList(), PyInt_AsPlacementStatus() must be put outside the C linkage back in the Isobar C++ namespace (clang). * Bug: In Hurricane::DBo::~DBo, and derived add a throw() specification (clang). * Bug: In Hurricane::RegularLayer::getEnclosure() & setEnclosure(), change signature so it matches the one of the base class (clang). * Bug: In Hurricane::CellPrinter, use double brackets for initializer list (clang). * Change: In Hurricane::Breakpoint, reverse the meaning of the error level. Only error level *lesser or equal* than the stop level will be enabled. * Bug: In CRL/python/helpers/__init__.loadUserSettings(), must put the current working directory in the sys.path as in certain configuration it may not be included. * Bug: In CRL::ApDriver, DumpSegments(), no longer generate segments when encountering a RoutingPad on a top-level Pin Occurrence. The segment was generated in the wrong direction, creating DRC violations on the "mips_core_flat" example. * Change: In CRL::Measures, partial re-design of the measurements management. Now, each kind of measure can accept multiple values put in a vector. The index is intented to match a tool run number. * Change: In CRL::Histogram, add support for multiple sets of datas, indexeds with tool run number. * Change: In CRL::ToolEngine, add support for multiple pass number, add addMeasure<> templates for the various data-types. * Change: In CRL::gdsDriver & CRL::gdsParser(), comment out unused GDS record name constants. * New: Etesian::BloatProperty, property to attach to Instance occurrences that contains the extra number of pitch to add to the cell width. * Bug: In AutoSegment::CompareByDepthLength, the segment length comparison was wrong, it was always returning true, which broke the "strick weak ordering" of the comparison. This was producing a core-dump in GCell::updateDensity() when sorting a vector<>. The end() iterator was being dereferenced, leading to the problem. * Bug: In Katana::DataSymmetric::checkPairing(), the test for segments whose axis is perpandicular to the symmetry axis was wrong ("!=" instead of "-"). * New: In Katana/GlobalRoute, new ::selectSegments(), selectOverloadedgcells() and selectBloatedInstances() to automatically select segments from overloaded edges, overloaded GCells and bloated cells. * Change: In KatanaEngine, return a more detailed success state, distinguish between global and detailed. Add support for multiple routing iterations. * New: In cumulus/python/plugins/ConductorPlugin.py, embryo of routing driven placement.
2019-12-08 18:57:44 -06:00
collections = []
if self._devtoolset:
Groudwork for routing density driven placement. Compliance with clang 5.0.1. This commit contains two set of features that should have been commited separately. 1. Compliance with clang 5.0.1, tested with the RedHat collection llvm-toolset-7. This allow Coriolis to be compiled under Darwin (MacOS) with Xcode & macports. The bootstrap install system has been modificated accordingly. 2. The basic support for routing density driven placement. Related features are: * Bloat property. Each Occurrence of an Instance can be individually bloated. This property not attached to any tool to allow the placer and router to share it as wanted. Nevertheless, it is defined in Etesian. * BloatProfile in Katana, add individual Bloat properties to Instances occurrences based on the East & North overflowed edges of each GCell. * Support in ToolEngine for a "pass number" of a tool. This pass number is mainly used to make "per pass" measurements. The MeasureSet system is improved accordingly to support multiple values of a same measure. * Embryo of "P&R Conductor" to perform the place & route loop until the design is successfully placed. May be the first brick of a Silicon Compiler. * Change: In boostrap/FindBoostrap.cmake, in setup_boost(), added tag to the python component for macport (ex: python27). * Change: In boostrap/build.conf, put etesian before anabatic for instance occurrence BloatProperty dependency. Added option support for the "llvm-toolset-7" collection to build against clang 5.0.1. * Bug: In Hurricane::getRecord( const pair<T,U>& ), the getSlot<> templates for first & second arguments must be called with <const T> and <const U> as the pair itself is const (and not simply <T> & <U>). * Change: In Hurricane::getSlot() temlate, only use "string" arguments and not const string&, simpler for template argument deduction. * Bug: In Hurricane::AnalogCellExtension, the StandardPrivateProperty<> template has a static member "_name". Clang did show that the template for this static number has to be put inside the namespace where the template *is defined* (i.e. Hurricane) instead of the namespace where it is instanciated (i.e. Analog). * Bug: In Isobar, Matrix_FromListOfList(), PyInt_AsPlacementStatus() must be put outside the C linkage back in the Isobar C++ namespace (clang). * Bug: In Hurricane::DBo::~DBo, and derived add a throw() specification (clang). * Bug: In Hurricane::RegularLayer::getEnclosure() & setEnclosure(), change signature so it matches the one of the base class (clang). * Bug: In Hurricane::CellPrinter, use double brackets for initializer list (clang). * Change: In Hurricane::Breakpoint, reverse the meaning of the error level. Only error level *lesser or equal* than the stop level will be enabled. * Bug: In CRL/python/helpers/__init__.loadUserSettings(), must put the current working directory in the sys.path as in certain configuration it may not be included. * Bug: In CRL::ApDriver, DumpSegments(), no longer generate segments when encountering a RoutingPad on a top-level Pin Occurrence. The segment was generated in the wrong direction, creating DRC violations on the "mips_core_flat" example. * Change: In CRL::Measures, partial re-design of the measurements management. Now, each kind of measure can accept multiple values put in a vector. The index is intented to match a tool run number. * Change: In CRL::Histogram, add support for multiple sets of datas, indexeds with tool run number. * Change: In CRL::ToolEngine, add support for multiple pass number, add addMeasure<> templates for the various data-types. * Change: In CRL::gdsDriver & CRL::gdsParser(), comment out unused GDS record name constants. * New: Etesian::BloatProperty, property to attach to Instance occurrences that contains the extra number of pitch to add to the cell width. * Bug: In AutoSegment::CompareByDepthLength, the segment length comparison was wrong, it was always returning true, which broke the "strick weak ordering" of the comparison. This was producing a core-dump in GCell::updateDensity() when sorting a vector<>. The end() iterator was being dereferenced, leading to the problem. * Bug: In Katana::DataSymmetric::checkPairing(), the test for segments whose axis is perpandicular to the symmetry axis was wrong ("!=" instead of "-"). * New: In Katana/GlobalRoute, new ::selectSegments(), selectOverloadedgcells() and selectBloatedInstances() to automatically select segments from overloaded edges, overloaded GCells and bloated cells. * Change: In KatanaEngine, return a more detailed success state, distinguish between global and detailed. Add support for multiple routing iterations. * New: In cumulus/python/plugins/ConductorPlugin.py, embryo of routing driven placement.
2019-12-08 18:57:44 -06:00
collections.append( 'devtoolset-%d' % self._devtoolset )
print 'Using devtoolset-%(v)d (scl enable devtoolset-%(v)d ...)' % {'v':self._devtoolset}
if self._llvmtoolset:
collections.append( 'llvm-toolset-%d' % self._llvmtoolset )
print 'Using llvm-toolset-%(v)d (scl enable llvm-toolset-%(v)d ...)' % {'v':self._llvmtoolset}
if collections:
commandAsString = ''
for i in range(len(command)):
if i: commandAsString += ' '
if ' ' in command[i]: commandAsString += '"'+command[i]+'"'
else: commandAsString += command[i]
command = [ 'scl', 'enable' ]
command += collections
command.append( commandAsString )
sys.stdout.flush ()
sys.stderr.flush ()
child = subprocess.Popen ( command, env=self._environment, stdout=None )
(pid,status) = os.waitpid ( child.pid, 0 )
status >>= 8
if status != 0:
ErrorMessage( status, "%s (status:%d)."%(error,status) ).terminate()
return
def _enableTool ( self, tool ):
return
def _build ( self, tool ):
toolSourceDir = os.path.join ( self.sourceDir, tool.getToolDir() )
toolBuildDir = os.path.join ( self.buildDir , tool.name )
cmakeInstallDir = os.path.join ( self.installDir, "share", "cmake", "Modules" )
# Supplied directly in the CMakeLists.txt.
#cmakeModules = os.path.join ( self.installDir, "share", "cmake", "Modules" )
if not os.path.isdir(toolSourceDir):
print ErrorMessage( 0, "Missing tool source directory: \"%s\" (skipped)."%toolSourceDir )
return
if self._rmBuild:
print "Removing tool build directory: \"%s\"." % toolBuildDir
command = [ "/bin/rm", "-rf", toolBuildDir ]
self._execute ( command, "Removing tool build directory" )
command = [ 'cmake' ]
if self._ninja: command += [ "-G", "Ninja" ]
Groudwork for routing density driven placement. Compliance with clang 5.0.1. This commit contains two set of features that should have been commited separately. 1. Compliance with clang 5.0.1, tested with the RedHat collection llvm-toolset-7. This allow Coriolis to be compiled under Darwin (MacOS) with Xcode & macports. The bootstrap install system has been modificated accordingly. 2. The basic support for routing density driven placement. Related features are: * Bloat property. Each Occurrence of an Instance can be individually bloated. This property not attached to any tool to allow the placer and router to share it as wanted. Nevertheless, it is defined in Etesian. * BloatProfile in Katana, add individual Bloat properties to Instances occurrences based on the East & North overflowed edges of each GCell. * Support in ToolEngine for a "pass number" of a tool. This pass number is mainly used to make "per pass" measurements. The MeasureSet system is improved accordingly to support multiple values of a same measure. * Embryo of "P&R Conductor" to perform the place & route loop until the design is successfully placed. May be the first brick of a Silicon Compiler. * Change: In boostrap/FindBoostrap.cmake, in setup_boost(), added tag to the python component for macport (ex: python27). * Change: In boostrap/build.conf, put etesian before anabatic for instance occurrence BloatProperty dependency. Added option support for the "llvm-toolset-7" collection to build against clang 5.0.1. * Bug: In Hurricane::getRecord( const pair<T,U>& ), the getSlot<> templates for first & second arguments must be called with <const T> and <const U> as the pair itself is const (and not simply <T> & <U>). * Change: In Hurricane::getSlot() temlate, only use "string" arguments and not const string&, simpler for template argument deduction. * Bug: In Hurricane::AnalogCellExtension, the StandardPrivateProperty<> template has a static member "_name". Clang did show that the template for this static number has to be put inside the namespace where the template *is defined* (i.e. Hurricane) instead of the namespace where it is instanciated (i.e. Analog). * Bug: In Isobar, Matrix_FromListOfList(), PyInt_AsPlacementStatus() must be put outside the C linkage back in the Isobar C++ namespace (clang). * Bug: In Hurricane::DBo::~DBo, and derived add a throw() specification (clang). * Bug: In Hurricane::RegularLayer::getEnclosure() & setEnclosure(), change signature so it matches the one of the base class (clang). * Bug: In Hurricane::CellPrinter, use double brackets for initializer list (clang). * Change: In Hurricane::Breakpoint, reverse the meaning of the error level. Only error level *lesser or equal* than the stop level will be enabled. * Bug: In CRL/python/helpers/__init__.loadUserSettings(), must put the current working directory in the sys.path as in certain configuration it may not be included. * Bug: In CRL::ApDriver, DumpSegments(), no longer generate segments when encountering a RoutingPad on a top-level Pin Occurrence. The segment was generated in the wrong direction, creating DRC violations on the "mips_core_flat" example. * Change: In CRL::Measures, partial re-design of the measurements management. Now, each kind of measure can accept multiple values put in a vector. The index is intented to match a tool run number. * Change: In CRL::Histogram, add support for multiple sets of datas, indexeds with tool run number. * Change: In CRL::ToolEngine, add support for multiple pass number, add addMeasure<> templates for the various data-types. * Change: In CRL::gdsDriver & CRL::gdsParser(), comment out unused GDS record name constants. * New: Etesian::BloatProperty, property to attach to Instance occurrences that contains the extra number of pitch to add to the cell width. * Bug: In AutoSegment::CompareByDepthLength, the segment length comparison was wrong, it was always returning true, which broke the "strick weak ordering" of the comparison. This was producing a core-dump in GCell::updateDensity() when sorting a vector<>. The end() iterator was being dereferenced, leading to the problem. * Bug: In Katana::DataSymmetric::checkPairing(), the test for segments whose axis is perpandicular to the symmetry axis was wrong ("!=" instead of "-"). * New: In Katana/GlobalRoute, new ::selectSegments(), selectOverloadedgcells() and selectBloatedInstances() to automatically select segments from overloaded edges, overloaded GCells and bloated cells. * Change: In KatanaEngine, return a more detailed success state, distinguish between global and detailed. Add support for multiple routing iterations. * New: In cumulus/python/plugins/ConductorPlugin.py, embryo of routing driven placement.
2019-12-08 18:57:44 -06:00
if self._macports: command += [ "-D", "WITH_MACPORTS:STRING=TRUE" ]
if self._noSystemBoost: command += [ "-D", "Boost_NO_SYSTEM_PATHS:STRING=TRUE"
Groudwork for routing density driven placement. Compliance with clang 5.0.1. This commit contains two set of features that should have been commited separately. 1. Compliance with clang 5.0.1, tested with the RedHat collection llvm-toolset-7. This allow Coriolis to be compiled under Darwin (MacOS) with Xcode & macports. The bootstrap install system has been modificated accordingly. 2. The basic support for routing density driven placement. Related features are: * Bloat property. Each Occurrence of an Instance can be individually bloated. This property not attached to any tool to allow the placer and router to share it as wanted. Nevertheless, it is defined in Etesian. * BloatProfile in Katana, add individual Bloat properties to Instances occurrences based on the East & North overflowed edges of each GCell. * Support in ToolEngine for a "pass number" of a tool. This pass number is mainly used to make "per pass" measurements. The MeasureSet system is improved accordingly to support multiple values of a same measure. * Embryo of "P&R Conductor" to perform the place & route loop until the design is successfully placed. May be the first brick of a Silicon Compiler. * Change: In boostrap/FindBoostrap.cmake, in setup_boost(), added tag to the python component for macport (ex: python27). * Change: In boostrap/build.conf, put etesian before anabatic for instance occurrence BloatProperty dependency. Added option support for the "llvm-toolset-7" collection to build against clang 5.0.1. * Bug: In Hurricane::getRecord( const pair<T,U>& ), the getSlot<> templates for first & second arguments must be called with <const T> and <const U> as the pair itself is const (and not simply <T> & <U>). * Change: In Hurricane::getSlot() temlate, only use "string" arguments and not const string&, simpler for template argument deduction. * Bug: In Hurricane::AnalogCellExtension, the StandardPrivateProperty<> template has a static member "_name". Clang did show that the template for this static number has to be put inside the namespace where the template *is defined* (i.e. Hurricane) instead of the namespace where it is instanciated (i.e. Analog). * Bug: In Isobar, Matrix_FromListOfList(), PyInt_AsPlacementStatus() must be put outside the C linkage back in the Isobar C++ namespace (clang). * Bug: In Hurricane::DBo::~DBo, and derived add a throw() specification (clang). * Bug: In Hurricane::RegularLayer::getEnclosure() & setEnclosure(), change signature so it matches the one of the base class (clang). * Bug: In Hurricane::CellPrinter, use double brackets for initializer list (clang). * Change: In Hurricane::Breakpoint, reverse the meaning of the error level. Only error level *lesser or equal* than the stop level will be enabled. * Bug: In CRL/python/helpers/__init__.loadUserSettings(), must put the current working directory in the sys.path as in certain configuration it may not be included. * Bug: In CRL::ApDriver, DumpSegments(), no longer generate segments when encountering a RoutingPad on a top-level Pin Occurrence. The segment was generated in the wrong direction, creating DRC violations on the "mips_core_flat" example. * Change: In CRL::Measures, partial re-design of the measurements management. Now, each kind of measure can accept multiple values put in a vector. The index is intented to match a tool run number. * Change: In CRL::Histogram, add support for multiple sets of datas, indexeds with tool run number. * Change: In CRL::ToolEngine, add support for multiple pass number, add addMeasure<> templates for the various data-types. * Change: In CRL::gdsDriver & CRL::gdsParser(), comment out unused GDS record name constants. * New: Etesian::BloatProperty, property to attach to Instance occurrences that contains the extra number of pitch to add to the cell width. * Bug: In AutoSegment::CompareByDepthLength, the segment length comparison was wrong, it was always returning true, which broke the "strick weak ordering" of the comparison. This was producing a core-dump in GCell::updateDensity() when sorting a vector<>. The end() iterator was being dereferenced, leading to the problem. * Bug: In Katana::DataSymmetric::checkPairing(), the test for segments whose axis is perpandicular to the symmetry axis was wrong ("!=" instead of "-"). * New: In Katana/GlobalRoute, new ::selectSegments(), selectOverloadedgcells() and selectBloatedInstances() to automatically select segments from overloaded edges, overloaded GCells and bloated cells. * Change: In KatanaEngine, return a more detailed success state, distinguish between global and detailed. Add support for multiple routing iterations. * New: In cumulus/python/plugins/ConductorPlugin.py, embryo of routing driven placement.
2019-12-08 18:57:44 -06:00
, "-D", "BOOST_INCLUDEDIR:STRING=/usr/include/boost169"
, "-D", "BOOST_LIBRARYDIR:STRING=/usr/lib64/boost169"
]
if self._qt5: command += [ "-D", "WITH_QT5:STRING=TRUE" ]
if self._openmp: command += [ "-D", "WITH_OPENMP:STRING=TRUE" ]
command += [ "-D", "CMAKE_BUILD_TYPE:STRING=%s" % self.buildMode
#, "-D", "BUILD_SHARED_LIBS:STRING=%s" % self.enableShared
, "-D", "CMAKE_INSTALL_PREFIX:STRING=%s" % self.installDir
, "-D", "CMAKE_INSTALL_DIR:STRING=%s" % cmakeInstallDir
#, "-D", "CMAKE_MODULE_PATH:STRING=%s" % cmakeModules
, toolSourceDir ]
if not os.path.isdir(toolBuildDir):
print "Creating tool build directory: \"%s\"." % toolBuildDir
os.makedirs ( toolBuildDir )
os.chdir ( toolBuildDir )
self._execute ( command, "First CMake failed" )
os.chdir ( toolBuildDir )
if self._noCache:
cmakeCache = os.path.join(toolBuildDir,"CMakeCache.txt")
if os.path.isfile ( cmakeCache ): os.unlink ( cmakeCache )
command += [ "-D", "BUILD_DOC:STRING=%s" % self._enableDoc
, "-D", "CMAKE_VERBOSE_MAKEFILE:STRING=%s" % self._verboseMakefile
, "-D", "CMAKE_INSTALL_PREFIX:STRING=%s" % self.installDir
, "-D", "CMAKE_INSTALL_DIR:STRING=%s" % cmakeInstallDir
]
if self.libSuffix: command += [ "-D", "LIB_SUFFIX:STRING=%s" % self.libSuffix ]
if self._checkDatabase == 'ON': command += [ "-D", "CHECK_DATABASE:STRING=ON" ]
if self._checkDeterminism == 'ON': command += [ "-D", "CHECK_DETERMINISM:STRING=ON" ]
command += [ toolSourceDir ]
self._execute ( command, "Second CMake failed" )
if self._doBuild:
command = [ "make" ]
if self._ninja:
command = [ "ninja-build" ]
#command += [ "DESTDIR=%s" % self.installDir ]
command += self._makeArguments
print "Make/Ninja command:", command
sys.stdout.flush ()
self._execute ( command, "Build failed" )
return
def gitArchive ( self, projectName ):
rawArchive = self.tarballDir+'/'+projectName+'.tar'
os.chdir ( self.sourceDir+'/'+projectName )
command = [ 'git'
, 'archive'
, '--prefix=%s/' % projectName
, '--output=%s' % rawArchive
, 'devel'
]
self._execute ( command, "git archive of project %s" % projectName )
if not os.path.isdir ( self.archiveDir ):
os.mkdir ( self.archiveDir )
os.chdir ( self.archiveDir )
command = [ 'tar', 'xf', rawArchive ]
self._execute ( command, "unpacking raw archive %s" % rawArchive )
command = [ 'rm', rawArchive ]
self._execute ( command, "Removing raw archive %s" % rawArchive )
# Adds files neededs only for packaging purpose.
command = [ "/bin/ln", "-s", "./coriolis/bootstrap/Makefile.package"
, self.archiveDir+"/Makefile" ]
self._execute ( command, "link of %s failed" % "coriolis/boostrap/Makefile.package")
command = [ "/bin/ln", "-s", "./coriolis/bootstrap/debian", self.archiveDir ]
self._execute ( command, "Copying Debian/Ubuntu package control files" )
# # Remove unpublisheds (yet) tools/files.
# for item in self.packageExcludes:
# command = [ "/bin/rm", "-rf", os.path.join(self.archiveDir,item) ]
# self._execute ( command, "rm of %s failed" % item)
#
# # Adds files neededs only for packaging purpose.
# command = [ "/bin/cp", "-r", os.path.join(self.sourceDir ,"bootstrap","Makefile.package")
# , os.path.join(self.archiveDir,"Makefile") ]
# self._execute ( command, "copy of %s failed" % "boostrap/Makefile.package")
#
# os.chdir ( self.archiveDir )
# command = [ "/usr/bin/patch", "--remove-empty-files"
# , "--no-backup-if-mismatch"
# , "-p0", "-i", self.distribPatch ]
# self._execute ( command, "patch for distribution command failed" )
absSourceTarBz2 = '%s/%s' % (self.tarballDir,self.sourceTarBz2)
os.chdir ( self.tarballDir )
command = [ 'tar', 'jcf', absSourceTarBz2, os.path.basename(self.archiveDir) ]
self._execute ( command, "Creating composite archive %s" % absSourceTarBz2 )
return
def _setEnvironment ( self, systemVariable, userVariable ):
if not self._environment.has_key(systemVariable) or self._environment[systemVariable] == "":
if not self._environment.has_key(userVariable) or self._environment[userVariable] == "" :
self._environment[ systemVariable ] = self.installDir
print "[WARNING] Neither <%s> nor <%s> environment variables are sets." \
% (systemVariable,userVariable)
print " Setting <%s> to <%s>." % (systemVariable,self.installDir)
else:
self._environment[ systemVariable ] = self._environment[ userVariable ]
if not self._quiet:
print "Setting <%s> to <%s>." % (systemVariable,self._environment[systemVariable])
if self._environment.has_key(userVariable):
print "Transmitting <%s> as <%s>." % (userVariable,self._environment[userVariable])
return
def _commandTemplate ( self, tools, projects, command ):
if self._clang:
Groudwork for routing density driven placement. Compliance with clang 5.0.1. This commit contains two set of features that should have been commited separately. 1. Compliance with clang 5.0.1, tested with the RedHat collection llvm-toolset-7. This allow Coriolis to be compiled under Darwin (MacOS) with Xcode & macports. The bootstrap install system has been modificated accordingly. 2. The basic support for routing density driven placement. Related features are: * Bloat property. Each Occurrence of an Instance can be individually bloated. This property not attached to any tool to allow the placer and router to share it as wanted. Nevertheless, it is defined in Etesian. * BloatProfile in Katana, add individual Bloat properties to Instances occurrences based on the East & North overflowed edges of each GCell. * Support in ToolEngine for a "pass number" of a tool. This pass number is mainly used to make "per pass" measurements. The MeasureSet system is improved accordingly to support multiple values of a same measure. * Embryo of "P&R Conductor" to perform the place & route loop until the design is successfully placed. May be the first brick of a Silicon Compiler. * Change: In boostrap/FindBoostrap.cmake, in setup_boost(), added tag to the python component for macport (ex: python27). * Change: In boostrap/build.conf, put etesian before anabatic for instance occurrence BloatProperty dependency. Added option support for the "llvm-toolset-7" collection to build against clang 5.0.1. * Bug: In Hurricane::getRecord( const pair<T,U>& ), the getSlot<> templates for first & second arguments must be called with <const T> and <const U> as the pair itself is const (and not simply <T> & <U>). * Change: In Hurricane::getSlot() temlate, only use "string" arguments and not const string&, simpler for template argument deduction. * Bug: In Hurricane::AnalogCellExtension, the StandardPrivateProperty<> template has a static member "_name". Clang did show that the template for this static number has to be put inside the namespace where the template *is defined* (i.e. Hurricane) instead of the namespace where it is instanciated (i.e. Analog). * Bug: In Isobar, Matrix_FromListOfList(), PyInt_AsPlacementStatus() must be put outside the C linkage back in the Isobar C++ namespace (clang). * Bug: In Hurricane::DBo::~DBo, and derived add a throw() specification (clang). * Bug: In Hurricane::RegularLayer::getEnclosure() & setEnclosure(), change signature so it matches the one of the base class (clang). * Bug: In Hurricane::CellPrinter, use double brackets for initializer list (clang). * Change: In Hurricane::Breakpoint, reverse the meaning of the error level. Only error level *lesser or equal* than the stop level will be enabled. * Bug: In CRL/python/helpers/__init__.loadUserSettings(), must put the current working directory in the sys.path as in certain configuration it may not be included. * Bug: In CRL::ApDriver, DumpSegments(), no longer generate segments when encountering a RoutingPad on a top-level Pin Occurrence. The segment was generated in the wrong direction, creating DRC violations on the "mips_core_flat" example. * Change: In CRL::Measures, partial re-design of the measurements management. Now, each kind of measure can accept multiple values put in a vector. The index is intented to match a tool run number. * Change: In CRL::Histogram, add support for multiple sets of datas, indexeds with tool run number. * Change: In CRL::ToolEngine, add support for multiple pass number, add addMeasure<> templates for the various data-types. * Change: In CRL::gdsDriver & CRL::gdsParser(), comment out unused GDS record name constants. * New: Etesian::BloatProperty, property to attach to Instance occurrences that contains the extra number of pitch to add to the cell width. * Bug: In AutoSegment::CompareByDepthLength, the segment length comparison was wrong, it was always returning true, which broke the "strick weak ordering" of the comparison. This was producing a core-dump in GCell::updateDensity() when sorting a vector<>. The end() iterator was being dereferenced, leading to the problem. * Bug: In Katana::DataSymmetric::checkPairing(), the test for segments whose axis is perpandicular to the symmetry axis was wrong ("!=" instead of "-"). * New: In Katana/GlobalRoute, new ::selectSegments(), selectOverloadedgcells() and selectBloatedInstances() to automatically select segments from overloaded edges, overloaded GCells and bloated cells. * Change: In KatanaEngine, return a more detailed success state, distinguish between global and detailed. Add support for multiple routing iterations. * New: In cumulus/python/plugins/ConductorPlugin.py, embryo of routing driven placement.
2019-12-08 18:57:44 -06:00
self._environment[ 'CC' ] = 'clang'
self._environment[ 'CXX' ] = 'clang++'
if self._devtoolset:
self._environment[ 'BOOST_INCLUDEDIR' ] = '/opt/rh/devtoolset-%d/root/usr/include' % self._devtoolset
self._environment[ 'BOOST_LIBRARYDIR' ] = '/opt/rh/devtoolset-%d/root/usr/lib' % self._devtoolset
if self._macports:
self._environment[ 'BOOST_INCLUDEDIR' ] = '/opt/local/include'
self._environment[ 'BOOST_LIBRARYDIR' ] = '/opt/local/lib'
# Set or guess the various projects TOP environment variables.
for project in self.projects:
topVariable = "%s_TOP" % project.getName().upper()
topUserVariable = "%s_USER_TOP" % project.getName().upper()
self._setEnvironment ( topVariable, topUserVariable )
if tools:
# Checks if the requested tools are in the various projects.
self.standalones = tools
for project in self.projects:
self.standalones = project.activate ( self.standalones )
for tool in self.standalones:
print "[WARNING] Tool \"%s\" is not part of any project." % tool
if projects:
for projectName in projects:
project = self.getProject ( projectName )
if not project:
ErrorMessage( 1, "No project of name \"%s\"."%projectName ).terminate()
project.activateAll()
if not tools and not projects:
for project in self.projects:
project.activateAll ()
for project in self.projects:
for tool in project.getActives():
print "\nProcessing tool: \"%s\"." % tool.name
getattr(self,command) ( tool )
return
def enable ( self, tools, projects ):
self._commandTemplate ( tools, projects, "_enableTool" )
return
def enabledTools ( self ):
tools = []
for project in self.projects:
tools += project.getActives()
return tools
def build ( self, tools, projects ):
self._commandTemplate ( tools, projects, "_build" )
return
def gitTarball ( self, tools, projects ):
if self.gitHash == "x":
self._guessGitHash ( self.getProject(projects[0]) )
self._doSpec ()
# self._doDebChangelog ()
if os.path.isdir(self.tarballDir):
print "Removing previous tarball directory: \"%s\"." % self.tarballDir
command = [ "/bin/rm", "-rf", self.tarballDir ]
self._execute ( command, "Removing top export (tarball) directory" )
print "Creating tarball directory: \"%s\"." % self.tarballDir
os.makedirs ( self.tarballDir )
self.gitArchive ( projects[0] )
return
def userTarball ( self, tools, projects ):
self.enable( tools, projects )
userSourceTarBz2 = os.path.join ( self.tarballDir
, datetime.date.today().strftime('%s-%s-%%Y%%m%%d.tar.bz2'%
(self.packageName
,self.packageVersion)) )
excludes = []
for exclude in self.packageExcludes:
excludes += [ '--exclude='+exclude ]
os.chdir ( self.sourceDir )
command = [ "/bin/tar"
, "--exclude-backups"
, "--exclude-vcs"
, "--transform=s,^,%s/src/,"%self.projectDir ] \
+ excludes \
+ [ "-jcvf", userSourceTarBz2 ] \
+ self.enabledTools()
self._execute ( command, "tar command failed" )
return
def doRpm ( self ):
self.gitTarball ( [], self.packageProjects )
for rpmDir in [ "SOURCES", "SPECS", "BUILD", "tmp"
, "SRPMS", "RPMS/i386", "RPMS/i686", "RPMS/x86_64" ]:
rpmFullDir = os.path.join ( self.rpmbuildDir, rpmDir )
if not os.path.isdir(rpmFullDir):
os.makedirs ( rpmFullDir )
else:
for entry in os.listdir(rpmFullDir):
path = os.path.join( rpmFullDir, entry )
if os.path.islink(path):
realpath = os.path.realpath( os.readlink(path) )
if not os.path.isfile(realpath):
print 'Remove obsolete link: <%s>.' % path
os.unlink( path )
rpmSpecFile = os.path.join ( self.rpmbuildDir, "SPECS" , "coriolis2.spec" )
rpmSourceFile = os.path.join ( self.rpmbuildDir, "SOURCES", self.sourceTarBz2 )
sourceFile = os.path.join ( self.tarballDir , self.sourceTarBz2 )
if os.path.isfile ( rpmSpecFile ):
os.unlink ( rpmSpecFile )
os.symlink ( self.specFile, rpmSpecFile )
if not os.path.islink ( rpmSourceFile ):
os.symlink ( sourceFile, rpmSourceFile )
os.chdir ( self.rpmbuildDir )
command = [ "/usr/bin/rpmbuild"
, "--define", "_topdir %s" % self.rpmbuildDir
, "--define", "_tmppath %s" % self.tmppathDir
#, "--define", "_enable_debug_packages 0"
, "--with", "binarytar" ]
if self._devtoolset:
command += [ "--define", "scl devtoolset-%d"%self._devtoolset ]
command += [ "-ba", "--clean", rpmSpecFile ]
self._execute ( command, "Rebuild rpm packages" )
return
def doDeb ( self ):
self.svnTarball ( [], self.packageProjects )
if not os.path.isdir(self.debbuildDir):
os.makedirs ( self.debbuildDir )
os.chdir ( self.debbuildDir )
sourceFile = os.path.join ( self.tarballDir , self.sourceTarBz2 )
debOrigFile = os.path.join ( self.debbuildDir, "coriolis2_1.0.%s.orig.tar.bz2" % self.gitHash )
if not os.path.islink(debOrigFile):
os.link ( sourceFile, debOrigFile )
command = [ "/bin/tar", "jxf", debOrigFile ]
self._execute ( command, "Unpacking pristine sources" )
#command = [ "/bin/cp", "-r", self.debianDir, "." ]
#self._execute ( command, "Copying Debian/Ubuntu package control files" )
packageDir = os.path.join ( self.debbuildDir, "coriolis2-1.0.%s" % self.gitHash )
os.chdir ( packageDir )
self._environment["CFLAGS" ] = "-O2"
self._environment["CXXFLAGS"] = "-O2"
command = [ "/usr/bin/debuild", "-us", "-uc" ]
self._execute ( command, "Rebuild Debian packages" )
return
def getProject ( self, name ): return self._conf.getProject(name)
def loadConfiguration ( self, confFile ): self._conf.load( confFile )
def showConfiguration ( self ): self._conf.show()