rmp: Genetic algorithm for slack tuning #8908

b-chmiel · 2025-11-26T16:01:14Z

New resynthesis strategy for rmp.
Alternative approach to #8437 for optimizing ABC timing.
Utilizes a genetic algorithm with the following tunable parameters:

population size
mutation probability
cross probability
tournament size
tournament probability

Results from regression tests:

test	slack	execution time [s]
asap7 aes_genetic	+26.52	164.3
asap7 aes_annealing	+14.28	52.7

Marking as draft for a discussion, currently working on some minor implementation restructuring and algorithm parameter tuning.

Depends on #8889.

maliberty · 2025-11-26T16:04:06Z

aes on what pdk? Are the slacks negative or positive?

b-chmiel · 2025-11-26T16:27:17Z

Tested on asap7 PDK.
Results (in the regression test suite):
aes_annealing.ok
aes_genetic.ok

Slacks are positive, the higher the better. I'll adjust the description.

maliberty · 2025-11-26T16:41:58Z

@QuantamHD please assign a Google reviewer

github-actions

clang-tidy made some suggestions

There were too many comments to post at once. Showing the first 25 out of 72. Check the log or trigger a new build to see more.

src/rmp/src/annealing_strategy.cpp

src/rmp/src/annealing_strategy.h

src/rmp/src/genetic_strategy.cpp

src/rmp/src/genetic_strategy.h

QuantamHD · 2025-11-26T17:17:08Z

@calewis could you review this PR from antmicro?

calewis · 2025-12-01T23:21:18Z

@calewis could you review this PR from antmicro?

Almost done, I just need to go through src/rmp/src/gia.cpp tomorrow.

calewis

A few things: I think a lot of the code could benefit from a sample function. std::sample is a bit problematic though because it uses an implementation defined std::uniform_int_distribution which can lead to different behavior for the same seed on different compilers. There might be a boost::sample, but I didn't see it in boost.random. So I suggest making a utl::sample using the absl::uniform_int_distribution.

I didn't review the TCL changes and I don't know much about what ABC is doing so I only reviewed gia.cpp for the code in the file and not for what it is trying to accomplish.

In general it's close to LTGM, but we'll need @maliberty to give it a once over as well. I might be missing some style things since this is my first OpenRoad review.

src/rmp/README.md

src/rmp/src/genetic_strategy.h

src/rmp/src/genetic_strategy.cpp

src/rmp/src/slack_tuning_strategy.cpp

src/rmp/src/gia.cpp

src/rmp/src/slack_tuning_strategy.h

src/rmp/src/gia.cpp

maliberty · 2025-12-02T17:56:16Z

I'll review after your comments are addressed

github-actions

clang-tidy made some suggestions

src/rmp/src/genetic_strategy.cpp

src/rmp/src/gia.cpp

github-actions

clang-tidy made some suggestions

src/rmp/src/genetic_strategy.cpp

src/rmp/src/utils.cpp

src/rmp/src/utils.h

github-actions · 2025-12-09T13:20:00Z

clang-tidy review says "All clean, LGTM! 👍"

github-actions · 2025-12-09T13:42:47Z

clang-tidy review says "All clean, LGTM! 👍"

b-chmiel · 2025-12-09T14:26:13Z

The only errors I'm getting right now are related to ABC imports. For instance, importing header such as map/if/if.h causes include cycle warning (which is there as it includes map/if/acd/ac_wrapper.h which includes if.h) and redeclaration errors in cmake build.
Not including this header causes misc-include-cleaner warning as those functions are not directly included in ac_wrapper.h. I'm also getting warnings from the inside of ABC library too.

@maliberty Should I avoid including this header altogether or ignore all problematical warnings with lint comments and compiler pragmas?

I'm also running a benchmark for #8908 (comment) to decide which mutation approach will be best.

Besides that, I believe this PR is ready for a review.

calewis

Thanks for fixing things

maliberty · 2025-12-09T19:06:17Z

If you don't include if.h in ac_wrapper.h, what are "those functions" that it complains about?

github-actions · 2025-12-10T13:56:45Z

clang-tidy review says "All clean, LGTM! 👍"

b-chmiel · 2025-12-11T15:51:38Z

@maliberty The if.h header is required to be able to construct If_Par_t struct with custom parameters (defined there).
This header throws two warnings:

Wredundant-decls (for repeated externs)
Wunused-but-set-variable (for inline code in utilTruth.h)
I'd silenced those warnings for this header only, this resolved issues.

I'd also removed call to ABSL_UNREACHABLE as it requires bumping abseil to at least 2023. Then we may also add it to

OpenROAD/src/dft/src/clock_domain/ClockDomain.cpp

Line 38 in b57cfc6

// TODO replace with std::unreachable() once we reach c++23

and nix environment files would need to be bumped as well in a separate PR.

github-actions · 2025-12-11T15:52:42Z

clang-tidy review says "All clean, LGTM! 👍"

github-actions · 2025-12-12T23:02:01Z

clang-tidy review says "All clean, LGTM! 👍"

b-chmiel · 2025-12-15T13:27:49Z

@maliberty All tests pass now without warnings

Co-authored-by: Krzysztof Bieganski <kbieganski@antmicro.com> Co-authored-by: Ryszard Rozak <rrozak@antmicro.com> Signed-off-by: Bartłomiej Chmiel <bchmiel@antmicro.com>

Signed-off-by: Bartłomiej Chmiel <bchmiel@antmicro.com>

Co-authored-by: Krzysztof Bieganski <kbieganski@antmicro.com> Co-authored-by: Ryszard Rozak <rrozak@antmicro.com> Signed-off-by: Bartłomiej Chmiel <bchmiel@antmicro.com>

Signed-off-by: Bartłomiej Chmiel <bchmiel@antmicro.com>

github-actions · 2025-12-17T09:48:23Z

clang-tidy review says "All clean, LGTM! 👍"

github-actions bot reviewed Nov 26, 2025

View reviewed changes

b-chmiel force-pushed the rmp-genetic-slack-tuning-algorithm branch from 476e7f0 to e2e28e4 Compare December 1, 2025 14:16

calewis suggested changes Dec 2, 2025

View reviewed changes

github-actions bot reviewed Dec 2, 2025

View reviewed changes

b-chmiel force-pushed the rmp-genetic-slack-tuning-algorithm branch from e2e28e4 to 46ed168 Compare December 8, 2025 20:37

github-actions bot reviewed Dec 8, 2025

View reviewed changes

b-chmiel force-pushed the rmp-genetic-slack-tuning-algorithm branch 2 times, most recently from 77af6d8 to f7540ca Compare December 9, 2025 11:59

b-chmiel marked this pull request as ready for review December 9, 2025 14:26

calewis approved these changes Dec 9, 2025

View reviewed changes

b-chmiel and others added 5 commits December 17, 2025 10:37

rmp: eqy check

8b87945

Co-authored-by: Krzysztof Bieganski <kbieganski@antmicro.com> Co-authored-by: Ryszard Rozak <rrozak@antmicro.com> Signed-off-by: Bartłomiej Chmiel <bchmiel@antmicro.com>

rmp: check eqy in nightly only

89feee3

Signed-off-by: Bartłomiej Chmiel <bchmiel@antmicro.com>

rmp: genetic algorithm for slack tuning

f96ee4b

Co-authored-by: Krzysztof Bieganski <kbieganski@antmicro.com> Co-authored-by: Ryszard Rozak <rrozak@antmicro.com> Signed-off-by: Bartłomiej Chmiel <bchmiel@antmicro.com>

rmp: restructure, clang-tidy

cf89991

Signed-off-by: Bartłomiej Chmiel <bchmiel@antmicro.com>

rmp: review fixes

ca69fc3

Signed-off-by: Bartłomiej Chmiel <bchmiel@antmicro.com>

b-chmiel added 9 commits December 17, 2025 10:37

rmp: tidy fixes

1d2245f

Signed-off-by: Bartłomiej Chmiel <bchmiel@antmicro.com>

rmp: regenerate test output after stable sort

bcb148f

Signed-off-by: Bartłomiej Chmiel <bchmiel@antmicro.com>

rmp: regenerate aes test after stable sort

05082df

Signed-off-by: Bartłomiej Chmiel <bchmiel@antmicro.com>

rmp: clang-tidy

5a1e3bd

Signed-off-by: Bartłomiej Chmiel <bchmiel@antmicro.com>

rmp: handle no selected candidates in tournament

9b48838

Signed-off-by: Bartłomiej Chmiel <bchmiel@antmicro.com>

rmp: fix abc gcc compilation

36afc7c

Signed-off-by: Bartłomiej Chmiel <bchmiel@antmicro.com>

rmp: fix abc gcc compilation

86d5101

Signed-off-by: Bartłomiej Chmiel <bchmiel@antmicro.com>

rmp: remove ABSL_UNREACHABLE

2a5bab3

Signed-off-by: Bartłomiej Chmiel <bchmiel@antmicro.com>

rmp: fix abc gcc warning

9206880

Signed-off-by: Bartłomiej Chmiel <bchmiel@antmicro.com>

b-chmiel force-pushed the rmp-genetic-slack-tuning-algorithm branch from 7eb5e78 to 9206880 Compare December 17, 2025 09:38

rmp: Genetic algorithm for slack tuning #8908

Are you sure you want to change the base?

rmp: Genetic algorithm for slack tuning #8908

Conversation

b-chmiel commented Nov 26, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

maliberty commented Nov 26, 2025

Uh oh!

b-chmiel commented Nov 26, 2025

Uh oh!

maliberty commented Nov 26, 2025

Uh oh!

github-actions bot left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

QuantamHD commented Nov 26, 2025

Uh oh!

calewis commented Dec 1, 2025

Uh oh!

calewis left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

maliberty commented Dec 2, 2025

Uh oh!

github-actions bot left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

github-actions bot left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

github-actions bot commented Dec 9, 2025

Uh oh!

github-actions bot commented Dec 9, 2025

Uh oh!

b-chmiel commented Dec 9, 2025

Uh oh!

calewis left a comment

Choose a reason for hiding this comment

Uh oh!

maliberty commented Dec 9, 2025

b-chmiel commented Nov 26, 2025 •

edited

Loading