The-OpenROAD-Project
DO NOT MERGE: Add slew violation check to size_down, re-enable size_down.
![github-actions[bot] avatar](https://avatars.githubusercontent.com/in/15368?v=4 "Published by a pub.dev verified publisher"){kind=small}
@mguthaus You need to amend d3d33f85e1a1cfd2db3153be03fff0e2abc94880 with a git rebase HEAD~2 -s then force push to make the DCO pass.
@povik Can you take a look? I think I've got general improvement in QoR in most cases.
@mguthaus I've started a run to evaluate QoR on ORFS. I will try to review the code
@mguthaus QoR comparison with a clean baseline (no other changes should be interfering): https://dashboard.precisioninno.com/compare?sourceAType=Commit&sourceBType=Commit&sourceBName=1171cb1d34e470c433f8eb5baef6b71213e6a920&sourceBID=1&sourceAName=4a04de24f4b308603ad92c9206633347015530cd&sourceAID=5753
I don't see any notable degradation.
Cool comparison tool.
It seems there is consistent finish__power__total and finish__design__core__area improvement in most designs (none got worse). This is expected since this is using downsizing to get improvement.
finish__timing__setup__ws and finish__timing__setup__tns results are a bit mixed but with more improvement than not. (Maybe I'm being subjective...). I think this also makes sense because other moves would get a similar result (albeit using more power).
Also, there is no degredatation in finish__runtime__total which was my main concern on larger designs.
@povik Can you take a look again? I believe I fixed all prior issues.
Results in setup slack are mixed tending to improved slack. But this is because the one path at a time repair setup needs to be improved and is in a local minima. It gets consistent power improvement with similar slack in all other cases.
I see merge conflicts to resolve. Is this otherwise ready for testing?
I rebased and pushed, but it fails clang-tidy due to the CI issues, I believe?