OpenROAD mpl2: improve annealing and debugger

When running CI for #4206, gf12/ariane failed in hierarchical macro placement for a specific macro cluster. Some debugging showed that some good results generated when SA temperature is high are being wasted. This PR is to make sure we keep the lowest cost valid result.

Also, we now know in which annealing step and perturb we are when debugging.

Some small rearrangement and cleaning as well.

Dec 06 '23 17:12 AcKoucher

Running Secure-CI.

Dec 06 '23 18:12 AcKoucher

clang-tidy review says "All clean, LGTM! :+1:"

Dec 06 '23 18:12 github-actions[bot]

clang-tidy review says "All clean, LGTM! :+1:"

Dec 07 '23 16:12 github-actions[bot]

CI is showing some failures, but I'll way to dig until #4357 is merged. I think it can have some meaningful impact.

Dec 08 '23 14:12 AcKoucher

clang-tidy review says "All clean, LGTM! :+1:"

Dec 08 '23 22:12 github-actions[bot]

Running Secure-CI with updates.

Dec 08 '23 22:12 AcKoucher

@AcKoucher what the status of this?

Dec 14 '23 17:12 maliberty

@maliberty I'm addressing the question in #4206, as it will be easier to not split the answer.

Dec 14 '23 18:12 AcKoucher

clang-tidy review says "All clean, LGTM! :+1:"

Apr 17 '24 00:04 github-actions[bot]

clang-tidy review says "All clean, LGTM! :+1:"

Apr 17 '24 01:04 github-actions[bot]

clang-tidy review says "All clean, LGTM! :+1:"

Apr 17 '24 01:04 github-actions[bot]