Closed dnltz closed 1 month ago
Part of the motivation to implement metal fill in OR was to allow future support for timing aware fill. If fill can be accomplished in OR it might be nice to keep that door open.
Part of the motivation to implement metal fill in OR was to allow future support for timing aware fill. If fill can be accomplished in OR it might be nice to keep that door open.
I would let @KrzysztofHerman decide whether OR should fill the core. Maybe Klayout just fills everything else?
If you require FEOL fill beyond the usual fill cells that would have to be done in KLayout. I'll wait on this PR for @KrzysztofHerman 's comments
If you require FEOL fill beyond the usual fill cells that would have to be done in KLayout. I'll wait on this PR for @KrzysztofHerman 's comments
I'm gonna drop this patch because not adding USE_FILL
has the same result. BTW USE_FILL=0
still adds filler cells.
Latest update.
flow/Makefile
to proper support sg13g2.map
. There should be no ORFS specific relative path in the IHP PDK. See conversation here https://github.com/IHP-GmbH/IHP-Open-PDK/pull/151/filessg13g2_antennanp
is now allowed as CELL according to @KrzysztofHerman 's PR. I forgot to pick that change too.@dnltz since we have to generate very specific filler in FEOL we have to use klayout so in a case of IHP platform lests use this tool for generating all fill cells FEOL+BEOL.
As for the sg13g2.map
I have already seen the changes in .lyt
fila added in our dev branch. If any modifications are to be done we could handle it "on the fly", while running sg13g2_update.py
script.
@dnltz since we have to generate very specific filler in FEOL we have to use klayout so in a case of IHP platform lests use this tool for generating all fill cells FEOL+BEOL. As for the
sg13g2.map
I have already seen the changes in.lyt
fila added in our dev branch. If any modifications are to be done we could handle it "on the fly", while runningsg13g2_update.py
script.
@KrzysztofHerman I think ORFS should be synced with main and not dev. Therefore, I made this change manually to fix it in a proper way. The current solution is way cleaner ...
any concerns about this PR?
Does this contain RAMS as inquired about in https://github.com/The-OpenROAD-Project/OpenROAD-flow-scripts/discussions/2099 ?
What remains in order to merge this?
@maliberty we are waiting for some fixes in IOCells, after it is ready we can update and merge.
Do you have an ETA for that? If it will be a while it might be good to make a separate PR to pull in the RAMs.
@KrzysztofHerman is this something we could update later? I would like to continue with the other pull-request to finally send ElemRV...
@dnltz since I do not have an ETA for the IOCells we can separate the abovementioned PR's.
Any eta for this?
As this PR is proceeding slowly I've opened https://github.com/The-OpenROAD-Project/OpenROAD-flow-scripts/pull/2171 to merge the RAM related files (only typ corner) to enable the user in #2099
As this PR is proceeding slowly I've opened #2171 to merge the RAM related files (only typ corner) to enable the user in #2099
Rebased to master and pinged @KrzysztofHerman to take a look.
@dnltz please go ahead
@maliberty it failed to build an asap7 design. Still fine?
@maliberty it failed to build an asap7 design. Still fine?
Yes
@maliberty I removed the sg13g2_io_notracks.lef
file since it has no use case here.
Picked @KrzysztofHerman update script and modified it slightly. Synced files afterwards with this script and added the bondpad manually.
Should we remove metal fill here since KLayout is supposed to do this task?
This PR is based on #2002 with some changes and addresses PR #2066