- New: In AllianceFramework constructor, now look for an XML configuration
file in the current directory (<cwd>/.environment.alliance.xml).
- New: DefParser, DefDriver & LefDriver are replaced by DefImport, DefExport
and LefExport (resp.). LEF/DEF parser/driver are no longer integrated
in among Alliance ones due to too much difference in data organisation.
Instead they are build as Import/Export utilities.
- New: Iccad04Lefdef import the ICCAD'04 benchmarks into the SxLib format
suitable for Knik/Kite. They are derived from the ISPD'98 (ibm01--ibm18)
benchmarks. Many corrections are applieds to ensure a more realistic
representation.
Generate a new library from the LEF file "on the fly".
- New: Ispd04Bookshelf import the ISPD'04 benchmarks in SxLib from bookshelf.
Thoses benchmarks can be placeds whith fastplace. Unfortunatly fastplace
do not manage blocks and reduce them to standard cell which is not
usable for detailed placement. Too bad, the parser exists still.
Make uses of the new vlsisapd/bookshelf parser. A whole library is
generated "on the fly" on the basis of "one node, one Cell" (so we have
a bijection between Cell and Instance). The result of this parsing is
different from the one of the stand-alone ispd04 binary in the sense it
fits the circuit to the SxLib gauge thus changing the routing capacity:
from 16 to 20.
- Change: In ApDriver, the lookup modify the lookup table for matching
Alliance layers against Hurricane ones, "OBSTACLEx" became "BLOCKAGEx".
- Change: In default <enviromnent.alliance.xml> adds the name of the
blockage net: "^obstacleNet$". Some work over the coherency of the
blockage managment remains to be done.
- Change: In AllianceFramework, partial rewrite of the get/creation of
AllianceLibrary to fit the requirements of the "on the fly" generation
from the benchmarks loaders. And correct some bugs by the way.
- Change: In Environment, suppress any reference to LEF technology as LEF/
DEF is now used as import/exports.
- Bug: In VstDriver, perform only one lookup for global signals among all
nets and not for each instance master cell. That was making the driver
quadratic! Now it's linear and reasonably fast.
+ change cmake OpenAccess macro we need 2 env variables :
OA_LIB_DIR where are the oa libs
OA_INCLUDE_DIR where are the oa headers
this way different headers version can be tested versus different OA compiled libs
- New: Added support for ACM/SIGDA ISCAS98 (.bench extension).
- Bug: In Parsers/Drivers changes the signature of the prototype, passes
file name string by value instead of by reference. In the cases of
reentrant P&D calls it may causes havoc.
- New: In Utilities, automatic setting of STRATUS_MAPPING_NAME. It's put
back into the environment from the configuration. This suppress the
need for the last environment variable under cgt.
- Change: New problem identified with the Python modules: each module seems
to be built as a complete binary, so all the static C++ initializers are
allocated in each module. In particular the C++ tree inheritance is built
for *each* module so we cannot longer uses the typeid() comparisons
across modules... It was used by boost::program_options to perform is
casts with boost::any and was starting throwing exceptions because of
bad casts. program_option was first initialized in "configuration"
first included by PyViewer then in PyCRL (see Utilities.cpp).
A first solution is to re-order the import of Python modules in
stratus1/st_model so that CRL is imported first.
The second is to not not link "configuration" with boost::program_option
as only the binary vlsisapd-conf-editor needs it.
That is a serious problem of which we must be aware and can cause further
strange behaviors.
Debug code used to diagnostic has been kept commented in the sources a
it may be needed again :-(
This behavior do not affect our singletons because they are part of
dynamic libraries that seems to be correctly shared between the various
Python modules.
* ./crlcore:
- In PyCRL, module method "getAllianceFramework()" moved as static object
method "get()" of AllianceFramework. Object AllianceFramework added
to module CRL.
cleaner separated functions converting Hurricane objects to oa
create 4 "Design CellView" of the cell : netlist, symbol, schematic and layout
remove bug with oaLib or oaTech at openning
adding oaLayerConstraint in oaLayer from the oaTech for the rule of the technology
- Bug: target_link_libraries() must be put back for OSX Snow Leopard
(doesn't seems to affect Leopard). As I do not have an OSX under
my hand it's untested and is likely to fail at that point.
* ./crlcore:
- New: PyCatalog & PyCatalogState python wrappers, mainly to have constants
values for stratus1 parsers/drivers.
- Library linking: there must not be "target_link_library()" for libraries,
only when building binaries. Avoid clashes between static module
or class variables, and strange reinitialisation of those variables.
- Change: Boost is now always linked staticly.
* ./crlcore:
- Change: In System, static initialisation problem again, there seems to
be a race between it and boost::filesystem, make the singleton creation
"at first call" again. Triggers the System singleton creation in
AllianceFramework to spare the user the need to explicitly doing it.
- New: In Utilities, automatically adds the "site-packages" PYTHONPATH
component, guessed from CORIOLIS_TOP.
Jean-Paul Chaput
Jean-Manuel Caba