✨ Added multi-threading support for QuickSim

[Drewniok]

No description provided.

[codecov]

Codecov Report

Merging #128 (d5240bd) into main (3ce5bf0) will increase coverage by 0.01%.
The diff coverage is 100.00%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #128      +/-   ##
==========================================
+ Coverage   93.87%   93.88%   +0.01%     
==========================================
  Files          75       75              
  Lines        6936     6955      +19     
==========================================
+ Hits         6511     6530      +19     
  Misses        425      425              

Impacted Files	Coverage Δ
...ion/algorithms/simulation/sidb/is_ground_state.hpp	100.00% <ø> (ø)
...on/algorithms/simulation/sidb/time_to_solution.hpp	96.87% <ø> (ø)
...de/fiction/algorithms/simulation/sidb/quicksim.hpp	98.00% <100.00%> (+0.94%)	⬆️
...fiction/technology/charge_distribution_surface.hpp	94.62% <100.00%> (+0.06%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update dc56a83...d5240bd. Read the comment docs.

[github-actions]

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

[marcelwa]

Why are you using std::thread instead of std::execution_policy?

[Drewniok]

Actually, my first step was to use std::execution_policy since you mentioned this once. However, I was not able to make it work. As far as I can see, Apple Clang does not yet support parallel algorithms and execution policies.

[marcelwa]

I can't say for sure and your milage may vary, but I had some luck writing a thin abstraction layer on top of std::execution_policy that makes the general concept available error-free on all platforms we're continuously testing on. That is, if policies are supported, they are used, if not, not. This might still not be the solution you were hoping for and it might not even work on Apple Silicon chips. That requires further testing. If you want to try it at least, please check out the execution policy macros on the placer branch.

[Drewniok]

Sounds really interesting! Maybe we can keep this in mind. For now, I would suggest using the current implementation since the performance increases as expected and I have already tested/analyzed all the different benchmarks. What is your opinion? I mean, you have more experience with multithreading than I do. If you think there is a benefit, I will have a look.

[github-actions]

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

[marcelwa]

I agree with you.

[marcelwa]

Could you please add a few tests for varying thread counts? After that, this PR is ready to be merged.

[github-actions]

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

[github-actions]

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

[github-actions]

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

[github-actions]

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

[github-actions]

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

[github-actions]

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