run on 4 cores and timeout after 12 hours

[mikemhenry]

Checklist

• Added a news entry

Developers certificate of origin

• I certify that this contribution is covered by the MIT License here and the Developer Certificate of Origin at https://developercertificate.org/.

[mikemhenry]

testing here https://github.com/OpenFreeEnergy/openfe/actions/runs/10117518516

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 92.36%. Comparing base (dd0be23) to head (d185608).

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #909      +/-   ##
==========================================
- Coverage   94.17%   92.36%   -1.81%     
==========================================
  Files         134      134              
  Lines        9800     9800              
==========================================
- Hits         9229     9052     -177     
- Misses        571      748     +177     

Flag	Coverage Δ
fast-tests	92.36% <ø> (?)
slow-tests	?

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[mikemhenry]

This worked great! I just failed since it expected a GPU, so the nvidia-smi commands failed, = 2 failed, 855 passed, 40 skipped, 1 xfailed, 2 xpassed, 3594 warnings, 3 rerun in 4059.65s (1:07:39) = way faster using -n 4 than -n auto

[mikemhenry]

(or did I mess up and make sure the log tests run?)

[IAlibay]

This worked great! I just failed since it expected a GPU, so the nvidia-smi commands failed, = 2 failed, 855 passed, 40 skipped, 1 xfailed, 2 xpassed, 3594 warnings, 3 rerun in 4059.65s (1:07:39) = way faster using -n 4 than -n auto

@mikemhenry the code in this PR was 4h not 1h (I think you're looking at my logs 😅).

Main difference is that I used a newer runner and removed coverage options (which we wouldn't use anyways).

[mikemhenry]

That is exactly what happened, when we switch to GPUs we probably want the timeout to be 2x or 1.5x what the "normal" run length ends up being so we have a failsafe if things go loopy, but isn't too strict to allow for runtime fluctuations