Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

🦈 IMP: Switch to BS Thread Pool Library #37

Merged
merged 11 commits into from
Aug 17, 2023
Merged

🦈 IMP: Switch to BS Thread Pool Library #37

merged 11 commits into from
Aug 17, 2023

Conversation

TheBlackPlague
Copy link
Owner

🎯 Summary

This PR switches from the currently used Thread Pool Library to the BS Thread Pool Library. The new thread pool library supports more platforms (including MacOS).

Furthermore, Parallel PERFT computation has been re-enabled and partially implemented with the use of C++ Standard Library std::future and BS Thread Pool Blocks. The combination allows for much faster PERFT computation than before, with about 60Mnps being gained per thread.

👏 Acknowledgements

NA

📈 ELO

STC:

ELO   | 0.96 +- 2.62 (95%)
SPRT  | 10.0+0.10s Threads=1 Hash=16MB
LLR   | 2.97 (-2.94, 2.94) [-3.00, 1.00]
GAMES | N: 34984 W: 9105 L: 9008 D: 16871

LTC:

ELO   | 0.96 +- 2.52 (95%)
SPRT  | 60.0+0.60s Threads=1 Hash=256MB
LLR   | 3.02 (-2.94, 2.94) [-3.00, 1.00]
GAMES | N: 35464 W: 8669 L: 8571 D: 18224

@TheBlackPlague
Use the BS Thread Pool library instead.
ca0b875 
Bench: 6691842
@TheBlackPlague
Use compiler-provided prefetch if not on x86/x64.
c7be3d7 
Bench: 6691842
@TheBlackPlague
Fix issues in MantaRay with custom memory stream implementation.
e8dca57 
Bench: 6691842
@TheBlackPlague
Try workaround.
5f09680 
Bench: 6691842
@TheBlackPlague
Final changes to MantaRay Package.
a953da0 
Bench: 6691842
@TheBlackPlague
Try reimplementing PERFT.
7a02fe6 
Eventually should move to the BS Thread Pool for parallelization
(so threads get reused properly).

Bench: 6691842
@TheBlackPlague
Begin work on using Thread Pool Library for PERFT.
e21b32b 
Bench: 6691842
@TheBlackPlague
Thread pool doesn't work that well with this. Revert to using std::fo…
c8a0c65 
…r_each.

Bench: 6691842
@TheBlackPlague
Devise custom future-based parallel scheduling for PERFT.
ed0d39f 
Bench: 6601319
@TheBlackPlague
No need to use atomic.
d52e1e6 
Bench: 6601319
@TheBlackPlague
Use blocks to maximize threading capabilities.
04682b4 
Bench: 6601319
@TheBlackPlague TheBlackPlague added = ELO This change doesn't alter ELO. = DOC This change doesn't improve the documentation. labels Aug 17, 2023
@TheBlackPlague TheBlackPlague self-assigned this Aug 17, 2023
@TheBlackPlague TheBlackPlague marked this pull request as ready for review August 17, 2023 08:22
@TheBlackPlague TheBlackPlague merged commit 0faabb5 into master Aug 17, 2023
@TheBlackPlague TheBlackPlague deleted the bs-thread-pool branch August 17, 2023 08:23
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers
No reviews
Assignees

@TheBlackPlague TheBlackPlague

Labels
= DOC This change doesn't improve the documentation. = ELO This change doesn't alter ELO.
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
1 participant
@TheBlackPlague