riscv
/
coriolis
mirror of https://gitlab.lip6.fr/vlsi-eda/coriolis.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity
coriolis/crlcore/etc/symbolic/vsc200/kite.conf

53 lines
2.4 KiB
Plaintext
Raw Normal View History

More configuration sharing. * Change: In CRL Core, .../etc/, put all that's shared under .../common/, all layouts and some settings not technologically dependant.
2014-05-27 08:40:42 -05:00
# -*- Mode:Python; explicit-buffer-name: "kite.conf<vsc200>" -*-
import helpers
# Contains the layout (shared by all technologies).
execfile( helpers.sysConfDir+'/common/kite.conf' )
* All Tools: - A complete sweep of cleanup to suppress allmost all compiler warnings. * ./crlcore: - New: Extented Python support for all objects needed to configure Coriolis. - New: Configuration files and load helpers completly written in Python. An XML emulation mode is still avalaible. - New: UsersGuide, as a replacement for README, written in ReST. - New: Start of Doxygen documentation (mostly for documenting the initialisation objects & methods).
2012-11-16 06:49:47 -06:00
parametersTable = \
In CRL, update real conf. files. Smarter management of pin in LEF parser. * Change: In CRL Core, etc/, update the configuration files of real technologies. Mostly for FreePDK 45. This work is also done for AMS c35b4 (350nm) but in a private (SoC) git repository. Added a new parameter 'lefImport.minTerminalwidth' for the minimum size (width) of a metal1 terminal in standard cells. Corrected bug of the minimum trace level which must be initialized to a great value and *not* zero; * Change: In CRL Core, BlifParser, detect when there is no tie low or tie high defined, issue an error (connexion left open) but continue. * New: In CRL::RoutingLayerGauge, new overlad of getTrackPosition() with the parameter set of getTrackIndex(). Used to know if a terminal is on-grid or not. * New: In CRL::LefImport, smarter management of metal1 pins. Adds a _pinPostProcess() function to select the external components among the various shapes. If the gauge is VH, all the pin rectangles are translateds into vertical segments (even if the metal1 gauge says the tracks are horizontals). The _pinPostProcess() function adds as external components of a net, only the segments of a sufficent width as given in 'lefImport.minTerminalWidth' and that are on-grid.
2018-01-06 09:18:33 -06:00
( ('lefImport.minTerminalWidth' ,TypeDouble ,0.0 )
, ("katabatic.globalLengthThreshold",TypeInt ,1450 ) # Katabatic parameters.
More configuration sharing. * Change: In CRL Core, .../etc/, put all that's shared under .../common/, all layouts and some settings not technologically dependant.
2014-05-27 08:40:42 -05:00
, ("katabatic.saturateRatio" ,TypePercentage,80 )
, ("katabatic.saturateRp" ,TypeInt ,8 )
Express "xEdgeCapacity" ratio as "xTracksReservedLocal" a number of tracks. * Change: In Kite, Katabatic & Knik, express the number of tracks truly available to the global router by the number of tracks reserved to the local routage inside a GCell. Replace the ratio parameter "hEdgeCapacity" by an integer parameter "hTracksReservedLocal" (duplicate for verticals). It is more explicit to give directly the number of tracks that are to be used locally, and potentially saves us from rouding problems when calculating the number of availables tracks. Note: we cannot do that for the layer saturateRatio as it uses the density ratio that take account local wires, leading to fractional results. * Change: In Katabatic, in <GCellGrid>, rename checkEdgeSaturation() into checkEdgeOverflow(), more explicit. * Change: In Knik, in <Graph>, display the computed capacities of the lower left node edges (should be the same througout all the grid). * Change: In Unicorn, in <cgt.py>, uses the new parameters names for edge density.
2014-06-10 09:58:52 -05:00
, ('katabatic.topRoutingLayer' ,TypeString , 'METAL5')
* All Tools: - A complete sweep of cleanup to suppress allmost all compiler warnings. * ./crlcore: - New: Extented Python support for all objects needed to configure Coriolis. - New: Configuration files and load helpers completly written in Python. An XML emulation mode is still avalaible. - New: UsersGuide, as a replacement for README, written in ReST. - New: Start of Doxygen documentation (mostly for documenting the initialisation objects & methods).
2012-11-16 06:49:47 -06:00
# Kite parameters.
Express "xEdgeCapacity" ratio as "xTracksReservedLocal" a number of tracks. * Change: In Kite, Katabatic & Knik, express the number of tracks truly available to the global router by the number of tracks reserved to the local routage inside a GCell. Replace the ratio parameter "hEdgeCapacity" by an integer parameter "hTracksReservedLocal" (duplicate for verticals). It is more explicit to give directly the number of tracks that are to be used locally, and potentially saves us from rouding problems when calculating the number of availables tracks. Note: we cannot do that for the layer saturateRatio as it uses the density ratio that take account local wires, leading to fractional results. * Change: In Katabatic, in <GCellGrid>, rename checkEdgeSaturation() into checkEdgeOverflow(), more explicit. * Change: In Knik, in <Graph>, display the computed capacities of the lower left node edges (should be the same througout all the grid). * Change: In Unicorn, in <cgt.py>, uses the new parameters names for edge density.
2014-06-10 09:58:52 -05:00
, ("kite.hTracksReservedLocal" ,TypeInt ,4 , { 'min':0, 'max':18 } )
, ("kite.vTracksReservedLocal" ,TypeInt ,3 , { 'min':0, 'max':18 } )
, ("kite.eventsLimit" ,TypeInt ,4000002 )
, ("kite.ripupCost" ,TypeInt ,3 , { 'min':0 } )
, ("kite.strapRipupLimit" ,TypeInt ,16 , { 'min':1 } )
, ("kite.localRipupLimit" ,TypeInt ,9 , { 'min':1 } )
, ("kite.globalRipupLimit" ,TypeInt ,5 , { 'min':1 } )
, ("kite.longGlobalRipupLimit" ,TypeInt ,5 , { 'min':1 } )
Corrections in the Dijkstra global routing (ripup) mechanism. * Bug: In Anabatic::Dijkstra, the degree of a vertex (the number of neighbors belonging to the same net) was miscalculated. This was leading, in the materialize step to some feed-through vertexes not being broken. Leading in turn to incomplete transformation of the detailed routing. Also in _trackback(), the degree of the first vertex we were backtracking from was not incremented. * Bug: In Anabatic::Dijkstra::materialize(), systematically use GCell::breakGoThrough() on both source and target. This is needed when we are in the ripup phase as both source and target can be go-through. This was also leading to incomplete detailed routing transformation. * Change: In Anabatic::Edge::ripup(), ripup one third of the segments instead of thoses exeeding the global length threshold. This way we are sure to desaturate an edge. Needs to be further calibrated. * Change: In Aanabatic::GCell::breakGoThrough(), no longer return NULL. Return existing gcontact if any. Break if it is a go-through and create a new gcontact in last resort. Maybe rename this function. * New: In Anabatic::Configuration, new parameters: - anabatic.edgeHScaling, to adjust the length of the horizontal edges relative to the vertical ones (this is a ratio). - anabatic.globalIterations, set the maximum number of ripup passes of the global router. * New: In CRL/etc/*/kite.conf, added new parameters anabatic.edgeHScaling and anabatic.globalIterations. * New: In Katana::GlobalRoute::DigitalDistance, take into account the new edgeHScaling factor. Must be used when the capacity of V-edges differs greatly for H-edges (case of AMS 350nm c35b4 for instance). * Bug: In Katana::GlobalRoute::DigitalDistance, the historic cost is computed for an edge length of "1". Must be multiplicated by the current edge length to have any measurable effect. This bug is finally explaining why the ripup was producing the same solutions over and over, the historical cost was negligible!
2018-04-16 05:10:48 -05:00
, ("anabatic.edgeHScaling" ,TypeDouble ,1.0 )
, ("anabatic.globalIterations" ,TypeInt ,10 , { 'min':1, 'max':100 } )
* All Tools: - A complete sweep of cleanup to suppress allmost all compiler warnings. * ./crlcore: - New: Extented Python support for all objects needed to configure Coriolis. - New: Configuration files and load helpers completly written in Python. An XML emulation mode is still avalaible. - New: UsersGuide, as a replacement for README, written in ReST. - New: Start of Doxygen documentation (mostly for documenting the initialisation objects & methods).
2012-11-16 06:49:47 -06:00
)
New coriolis launcher. Configuration files cleanup. * Change: In CRL Core, simplify the loading sequence. The technology, both symbolic and real is now loaded directly from coriolisInit.py and not through the Alliance loader. This was a leftover from the time configuration was in XML. Remove others traces of XML loading. Remove SYMB_TECHNO_NAME, REAL_TECHNO_NAME & DISPLAY from the Alliance environement, as they was no longer used. Note that technology *still* need to be loader *after* Alliance framework has been initialized. Gauge information is moved from <alliance.conf> to <kite.conf>. * Bug: In Coloquinte, in optimization_subproblems.cxx static variables must not be inlined. Generate a problem when linking in debug mode (seems the symbol gets optimised out). * Bug: In Katabatic, in Grid::getGCell(), when the coordinate is *outside* the area, do not try to find a GCell, directly return NULL. * New: In Unicorn, create a generic command launcher named "coriolis" which automatically take cares of all environement setup, then run a command by default, it's <cgt>, but it can be anything. For example: <zsh>.
2015-04-13 11:54:09 -05:00
# Format of routingGaugesTable (dictionary):
# A list of entry of the form:
New "obstacleDw" for Kite. * New: In CRL Core, in RoutingLayerGauge, adds a new parameter "obstacleDw". It represents the distance to add from the *edge* of a blockage segment to the next *axis* of a routing track. This was previously computed as the pitch minus the mimimum half wire width. This works if the pitch is *equal* to the wire width plus the minimal distance (edge to edge). But if the wire width is less, than it leads to too great a distance around obstacles, making them bigger than they are. And potentially blocking everything. To summarize: before: obstacleDw = pitch - wireWidth/2 (deduced) after: obstacleDw = explicitly setup, usually: pitch - max(wireWidth)/2 Modify the Alliance configuration helper and the various kite.conf file accordingly. * New: In Kite, in BuildPowerRails, make use of "obstacleDw" to compute the footprint of a blockage.
2016-04-13 11:18:43 -05:00
# ( METAL_NAME, (Direction, Type, depth, density, offset, pitch, wire_width, via_width, obs_dw) )
Support for Net alias names. Blif parser enhancements. * New: In Hurricane, In Net & Cell, support for Net aliases names. Use a structure based on a simple ring of NetAliasHook. The Net holds a global map, sorted by names of all the aliases of all Nets. Elements NetAliasesHook of the map are slaves of ring whose master is an attribute of the Net (it is *not* in the map, as the primary name of the Net). In case of merge, the aliases of both Nets are merged and the name of the merged one become an alias. The Cell::getNet() looks in both the Net map and the aliases to find a Net by name. * Bug: In CRL Core, in coriolisInit.py, reoder the loading of the configuration files so the real technology is read as early as possible to set up the <gridsPerLambda> factor before any lambda is actually computed... * Bug: In CRL Core, in AcmSigda, do not try to fed the file when it has failed to be opened. Throw a clean exception instead. * New: In CRL Core, in Toolbox, add a NamingScheme object to convert a design into VHDL compliant names (mainly from Blif/Verilog). This is extensible in any case. * New: In CRL Core, in BlifParser, slightly more informative warning messages. Align the loading progress information on the other parsers. Add a capability to select which component of the design will be returned, if there are more than one. Use the "." as separator. For exemple you can request "Processor.Alu", which will load the "Alu" component from the design in "Processor.blif". To be able to save a Blif loaded design, systematically convert all the name for VHDL compliance, as it is the format used by the Coriolis native files (vst). Export the Blif parser to the Python interface. * New: In Kite, In NegociateWindow, add a counter of the number of remaining events. Gives an idea of the ETA... * New: In Unicorn, in cgt.by, add an option to load a Blif design from the command line. * New: In Cumulus, new RSave plugin to save both netlist & layout. Partly redundant with the previous one. Have to better organize that later.
2015-04-16 08:40:02 -05:00
New coriolis launcher. Configuration files cleanup. * Change: In CRL Core, simplify the loading sequence. The technology, both symbolic and real is now loaded directly from coriolisInit.py and not through the Alliance loader. This was a leftover from the time configuration was in XML. Remove others traces of XML loading. Remove SYMB_TECHNO_NAME, REAL_TECHNO_NAME & DISPLAY from the Alliance environement, as they was no longer used. Note that technology *still* need to be loader *after* Alliance framework has been initialized. Gauge information is moved from <alliance.conf> to <kite.conf>. * Bug: In Coloquinte, in optimization_subproblems.cxx static variables must not be inlined. Generate a problem when linking in debug mode (seems the symbol gets optimised out). * Bug: In Katabatic, in Grid::getGCell(), when the coordinate is *outside* the area, do not try to find a GCell, directly return NULL. * New: In Unicorn, create a generic command launcher named "coriolis" which automatically take cares of all environement setup, then run a command by default, it's <cgt>, but it can be anything. For example: <zsh>.
2015-04-13 11:54:09 -05:00
routingGaugesTable = {}
routingGaugesTable['vsclib'] = \
New "obstacleDw" for Kite. * New: In CRL Core, in RoutingLayerGauge, adds a new parameter "obstacleDw". It represents the distance to add from the *edge* of a blockage segment to the next *axis* of a routing track. This was previously computed as the pitch minus the mimimum half wire width. This works if the pitch is *equal* to the wire width plus the minimal distance (edge to edge). But if the wire width is less, than it leads to too great a distance around obstacles, making them bigger than they are. And potentially blocking everything. To summarize: before: obstacleDw = pitch - wireWidth/2 (deduced) after: obstacleDw = explicitly setup, usually: pitch - max(wireWidth)/2 Modify the Alliance configuration helper and the various kite.conf file accordingly. * New: In Kite, in BuildPowerRails, make use of "obstacleDw" to compute the footprint of a blockage.
2016-04-13 11:18:43 -05:00
( ( 'METAL1', ( Gauge.Vertical , Gauge.PinOnly, 0, 0.0, 4, 8, 3, 2, 6.5 ) )
, ( 'METAL2', ( Gauge.Horizontal, Gauge.Default, 1, 7.0, 0, 8, 4, 2, 6 ) )
, ( 'METAL3', ( Gauge.Vertical , Gauge.Default, 2, 0.0, 4, 8, 4, 2, 6 ) )
, ( 'METAL4', ( Gauge.Horizontal, Gauge.Default, 3, 0.0, 0, 8, 4, 2, 6 ) )
, ( 'METAL5', ( Gauge.Vertical , Gauge.Default, 4, 0.0, 4, 8, 4, 2, 6 ) )
#, ( 'METAL6', ( Gauge.Horizontal, Gauge.Default, 5, 0.0, 0, 8, 4, 2, 6 ) )
#, ( 'METAL7', ( Gauge.Vertical , Gauge.Default, 6, 0.0, 4, 8, 4, 2, 6 ) )
New coriolis launcher. Configuration files cleanup. * Change: In CRL Core, simplify the loading sequence. The technology, both symbolic and real is now loaded directly from coriolisInit.py and not through the Alliance loader. This was a leftover from the time configuration was in XML. Remove others traces of XML loading. Remove SYMB_TECHNO_NAME, REAL_TECHNO_NAME & DISPLAY from the Alliance environement, as they was no longer used. Note that technology *still* need to be loader *after* Alliance framework has been initialized. Gauge information is moved from <alliance.conf> to <kite.conf>. * Bug: In Coloquinte, in optimization_subproblems.cxx static variables must not be inlined. Generate a problem when linking in debug mode (seems the symbol gets optimised out). * Bug: In Katabatic, in Grid::getGCell(), when the coordinate is *outside* the area, do not try to find a GCell, directly return NULL. * New: In Unicorn, create a generic command launcher named "coriolis" which automatically take cares of all environement setup, then run a command by default, it's <cgt>, but it can be anything. For example: <zsh>.
2015-04-13 11:54:09 -05:00
)
# Format of cellGaugesTable (dictionary):
# A list of entry of the form:
# ( METAL_PIN, xy_common_pitch, slice_height, slice_step )
cellGaugesTable = {}
cellGaugesTable['vsclib'] = ('metal2', 8.0, 72.0, 8.0)