- New: In LayerAssignement & GCell, perform layer assignment on whole net
trunk, instead on a segment by segment basis.
- New: In GCell, fragmentation analysis. When a layer change is requested,
it allows to check if the additionnal perpandiculars created to
maintain connexity will not over fragment the layer. This is particu-
larly critical for M3->M5 which create M4 in perpandiculars, and M4
may be subject to very dense configuration.
- New: In AutoSegment & LoadGrByNet, adds a flags telling if the net's
degree equal 2. Has not proven to be useful yet.
- New: In AutoSegment, adds a flag telling if the segment is the perpan-
dicular part of a dogleg. Has not proven useful yet.
- New: In GCell, compute a "fragmentation" indice, the average free track
length, the more fragmented, the smaller the indice.
- New: In AutoSegment::canMoveUp(), when requested, not only check for free
(parallel) track in the up layer but also check that the begin/end
local segments created to maintain layer connexity do not creates too
much fragmentation. This fragment could prove critical in higly dense
GCells.
- Change: In LayerAssign, disable the ::rpDesaturate() steps, after optimization
the router can handle 10 terminals per GCells in M2 without problem.
And ::rpDesaturate() must uses GCell::stepNetDesaturate() and *not*
directly ::stepDesaturate().
- 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).
- Change: In GCell & GCellGrid, unificated way of computing all the GCell's
densities.
- Change: In GCell::stepDesaturate(), uses shearUp() instead of moveUp().
- Change: In AutoSegment, new method ::shearUp() to avoid moveUp() saturating
GCells as side-effect.
- New: ChipTools, regroup all datas and utilities to manage a full-chip
design.
- Change: In LoadGrByNet/GCellConfiguration::_GCell_xG_1Pad(), uses straight
perpandicular wires on top & right pads. The GCells under those connectors
are fully saturated, wires must go out as straight as possible.
- Change: In AutoHorizontal/AutoVertical, update the "terminal" flag after
slackening. This is to avoid global that are no longer connected to
terminals behave as such.
- Change: In AutoSegment::canMoveUp() & canPivotUp(), prevent all local segments
to go up. This is to avoid cluttering upper levels with small segments.
- Change: In GCellConfiguration, for xG_xL3, detect straight vertical topologies
and for them to be fixed (new flag FIXED_GLOBAL set by the topological
builder instead of the constructor). This prevent the router to do
stupid things...
To sets the "Fixed" flag *after* the axis of global segments have been
positionned correctly adds a "_toFixGlobals" static table lookup.
- Change: In GCell::stepDesaturate(), if less than one free track remains in
the upper layer, do not move up the segment. This allows from a minimum
free room for expansion.
- Change: In GCell::_getString(), display indexes and layer names in various
saturation tables.
- Change: In AutoSegment, allow move up of local Segment. ::moveUp() and
::canMoveUp() arguments go from booleans to flags, which are more
explicits.
- Change: In loadGlobalRouting(), more exlicit message as to why a net is
filtered out the routing set (POWER, GROUND, CLOCK or BLOCKAGE).
- New: ChipTool, module with utilities specific to chip routing. Containing
a function to pre-break wires around a block.
- New: In loadGlobalRouting/GCellConfiguration, adds supports for RoutingPad
connected on chip's pad.
- New: In GCellGrid/GCell, can compute the density in three modes: average,
max of H/V and max of layer. Currently implemented in getMaxHVDensity(),
should be unificated with getDensity(). Used for display purposes.
- Bug: In AutoContact, when splitting a contact, add a specific check for
"one layer span".
- Bug: In AutoSegment::canMoveUp(), do not move up fixed segments.
- Change: In LayerAssign, desaturate until the last two top layers are reacheds
instead of only METAL2 & METAL3.
- Change: In Configuration/GCellGrid/GCell, correct computation of the edges
capacity (based on the numbers of avalaibles layers and not hard-wired to
4).
- New: To pass the ibmXX, new parameter "katabatic.saturateRp" which allows
to sets the threshold above which a GCell has too much RoutingPads in
it and must be further desaturated. Must be 8 for FPGA, and 10 for ibm.
- Bug: In LoadGrByNet(), when looking for the starting GCell, the computation
of unconnecteds RoutingPads was incorrect, too great and letting the
router believe that the routage was incomplete.
- 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.
- 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.
./metis,
./mauka,
./katabatic,
./kite :
- Change: switch toward the XML configuration system. Suppress the need
of a default static configuration (now built on demand from the XML
database).
In Mauka and Kite, suppress the ConfigurationWidget, wich are
obsoleted by the generic XML configuration widget.
- Change: adopt a tree layout compliant with the UNIX FHS.
* includes under TOP/include/coriolis2.
* shared datas under TOP/shared/coriolis2.
* docs under TOP/share/doc/coriolis2.
* configuration under TOP/etc/coriolis2
* ./crlcore:
- Change: In Environment, comply to the new tree layout, search configuration
files under TOP/etc/coriolis2/.
* ./knik:
- Change: In flute, comply to the new tree layout, get the "POW*.dat" files
from TOP/share/coriolis2/flute-2.4.
libraries gets installed in "lib64" instead of "lib".
buildCoriolis.py sets automatically LIB_SUFFIX for cmake.
coriolis2.spec modificated to uses lib64 on 64 bits.
- Change: In the CMakeLists.txt, in all the install commands remove all
the leading "/" as they prevents the CMAKE_INSTALL_PREFIX to be took
into account. It was nevertheless working because buildCoriolis.py was
using DESTDIR which is prepended anyway.
* ./goodies:
- Change: In buildCoriolis.py, no longer uses the DESTDIR but instead
CMAKE_INSTALL_PREFIX.
- Change: In KatabaticEngine, the containers used to store the "nets to
route" is changed from a vector<> to a set<> sorted by the Nets names.
Uses more memory but allows much faster Net deletion.
- New: In _loadNetGlobalRouting(Net*), detect non-routed Nets. A Net is
considered as unrouted if it's 10 first RoutingPads have no slave
Segment. The Net is withdrawn from the set of Nets to route.
- Bug: In GCellConfiguration::_GCell_1G_1L1(), correctly compute the
access flag passed to _GCell_rp_Access().
- Change: In AutoContact::restrictConstraintBox() adds "warnOnError"
parameter to shut down warnings.
- New: KatabaticEngine::_computeNetConstraints() now also returns a set<>
of segments on which constraints are incompatibles. All the NetConstraints
related functions are modificateds accordingly.
- Change: In KatabaticEngine::_loadNetGlobalRouting(), create dogleg
on overconstrained Segments computed by a call to _computeNetConstraints().
Overconstraineds Segments can occurs when a global is directly
connecting two punctual RoutingPads (see "ieee_division").
As a side effect, and to avoid looping each Net is revalidated
*before* making any dogleg. So Nets are revalidated one by one instead
of alltogether.
- Change: In AutoSegment::_toConstraintAxis(), do nothing if constraint
interval is empty.
- Bug: In KatabaticEnginge::_desaturate() (part of LayerAssign), reset
the moved up Segment pointer between iterations. Suppress the warning
about the GCell's DyKeyQueue beeing not revalidated on deletion.
- Change: In LoadGrByNet, manage the 4G+3L1 configuration with the generic
_GCell_xG_xL1_xL3(). Needed for "dct_lvl2".
- Bug: In top CMakeLists.txt the SETUP_PROJECT_PATHS was not inserting
X_USER_TOP *before* X_TOP, thus potentially allowing an obsolete
system-wide configuration to shadow an up-to-date local one.
* ./katabatic:
- Change: In Configuration/KatabaticEngine, introduce a static Configuration
object (singleton) to handle the default configuration. Configuration are
now created only through cloning operations. In KatabaticEngine, the
_configuration becomes a pointer (previoulsy an value).
Confifuration becames a singleton and a decorator.
Also adds a print method to nicely display configurations values on
the terminal.
- Change: In Grid/BaseGrid, add an Abutment Box attribute (to facilitate
segment truncation in Kite::TrackFixedSegment.
- Change: In Grid/BaseGrid::Axis, when computing a row/column index using
the graduation table, if the coordinate is exactly on the last
graduation, return the last GCell instead of "out of bound".
- Change: In GCell::hasFreeTrack(), AutoSegment::canMoveUp() &
AutoSegment::canPivotUp() adds a "reserve" parameter to modify the
amount of wirelength to remains free after the insertion.
- Bug: In GCell, bad cast between <size_t> and <unsigned int>, showed only
under 64 bits.
- New: In KatabaticEngine::printMeasures(), adds a support for measurments,
both time and memory size in Mb are stored for each tag.
- New: Adds measurements "Gates", "GCells", "load" & "assign".
- New: In GCell & LayerAssign, adds a supplemental desaturation steps when
a GCell containts too many (more than 7) RoutingPads in M1. This is
especially useful for the eFPGA matrix with it's custom made cells.
- Change: Uses a map instead of a sorted vector in the layer assignment
steps. It was used to order GCells by density, and causing terrific
slowdown on big examples. Namely the eFPGA 16x16 took 1h20, now it's
down to 21 seconds...
Devellop a new object <DyKeyQueue> to handle map with elements
whose key *can* change. Next step is to normalize the key caching
mechanism and to templatize.
- Change: In LayerAssign, account the number of globals AutoSegments
moved up.
- Change: <PROJECT>_SEARCH_PATH are put back into the *first* tool of
a project.
- Bug: In HURRICANE_CHECK_MACRO(), the quiet flag was not correctly
implemented. User ARGV instead of argv (case sensitivity!).
- Change: New structure for the installation & CMake system.
* Tools are now grouped in "projects". There are three projects:
1. - IO: Standalones parsers/drivers (IO_USER_TOP, IO_TOP).
2. - Coriolis: Base & digital tools (CORIOLIS_USER_TOP, CORIOLIS_TOP).
3. - Chams: Analogic tools (CHAMS_USER_TOP, CHAMS_TOP).
Each *project* has a two "TOP" environement variables, for
example: IO_TOP and IO_USER_TOP. Thoses variables are the only
ones useds to locate the tool (CMake modules, headers & libraries).
The local path always takes precedence over the global one.
The localisation process occurs in each tool top CMakeLists.txt
where the macro SETUP_PROJECT_PATH is to be defined. There is no
way to put it in a shared includes file as it's the macro precisely
used to locates the includes... You have to call the macro once for
each project you wants to uses:
SETUP_PROJECT_PATHS(IO)
SETUP_PROJECT_PATHS(CORIOLIS)
* In FindTOOL.cmake, supress the <TOOL>_DIR_SEARCH and uses the
<PROJECT>_DIR_SEARCH instead (example: CORIOLIS_DIR_SEARCH).
* buildCoriolis.py modificated according to the new "TOP" scheme.
- Change: Adaptation to cmake 2.8. No more "add_subdirectory()" when
the sub directory doesn't hold a CMakeList.txt.
For "add_definition()", set policy 00005 to NEW.
- Change: Integrate IO extenal libraries.
- Change: Small compatibilities between "size_t" and "unsigned int"
under 64 bits.
- Change: Temporarily disable Equinox/Solstice in unicorn.
./coriolis/src/crlcore,
./coriolis/src/knik,
./coriolis/src/katabatic,
./coriolis/src/kite,
./coriolis/src/equinox,
./coriolis/src/solstice,
./coriolis/src/ispd:
- SVN MOVE: Source tree simplification & uniformisation. Now all tools
are at the same level, directly under the root of the repository.
No more "coriolis/src".
Jean-Paul Chaput
The Coriolis Project