Molybdenum

[EzioMagnifico]

Molybdenum v3.0 (Linux and Windows) .
Thanks to the Author rn5f107s2

Molybdenum is a chess engine written in C++ .

Download : https://github.com/rn5f107s2/Molybdenum/releases/tag/v3.0

This release features a bunch of Elo gains related to search for a total of
~85 STC self play Elo:

Elo | 84.75 +- 4.99 (95%)
Conf | 8.0+0.08s Threads=1 Hash=8MB
Games | N: 10000 W: 3857 L: 1465 D: 4678
Penta | [59, 532, 1886, 2004, 519]
http://aytchell.eu.pythonanywhere.com/test/226/
~70 LTC self play Elo:

Elo | 73.48 +- 9.62 (95%)
Conf | 40.0+0.40s Threads=1 Hash=64MB
Games | N: 2500 W: 880 L: 359 D: 1261
Penta | [8, 132, 527, 497, 86]
http://aytchell.eu.pythonanywhere.com/test/225/

Elo gains:

Transposition table improvements:
TT cutoffs in qsearch: #53
TT move ordering in qsearch: #54
Improved replacement scheme: #55
Smaller entries: #56

Fractional reductions:
In history pruning: #61
In move count pruning: #62
In futility pruning: #63

All the other stuff:
Introduce singular extensions: #51
Use singular search to improve move ordering: #52
Symmetrize history limits: #57
RFP more after bad moves: #59
Tune time management: #60
Tune search; #64
Adjust bestScore if inaccurate: #65

The provided binaries use avx2, bmi2... (v3) / sse3, popcnt.. (v2) optimization respectively, note that even ryzens before zen 3 should be able to use the bmi2 binaries without slowdown, as they contain no pext / pdep instructions

Contributors 2 :
@rn5f107s2
rn5f107s2
@aytchell
aytchell Hannes Lerchl

[EzioMagnifico]

Molybdenum v3.1 (Linux and Windows) .
Thanks to the Author rn5f107s2

Molybdenum is a chess engine written in C++ .

Download : https://github.com/rn5f107s2/Molybdenum/releases/tag/v3.1

This release features an Elo gain of around 150 mostly coming from NNUE improvements:

STC:
Elo | 142.76 +- 18.12 (95%)
Conf | 8.0+0.08s Threads=1 Hash=8MB
Games | N: 1002 W: 517 L: 127 D: 358
Penta | [1, 33, 153, 203, 111]
http://aytchell.eu.pythonanywhere.com/test/347/

LTC:
Elo | 152.60 +- 18.14 (95%)
Conf | 40.0+0.40s Threads=1 Hash=64MB
Games | N: 1000 W: 526 L: 113 D: 361
Penta | [3, 21, 130, 252, 94]
http://aytchell.eu.pythonanywhere.com/test/346/

Elo gains:

Search Improvements:
While this release I mostly focused on improving the NNUE there are a few search patches:

Adjust the singular extensions margin based on the rootmovecount of the current branch: #66
(Note that the commit message may seem confusing, that's because I intended exactly the opposite effect of what this currently does, I tried to "fix" it, but so far unsuccessfully)
Don't singular extend the rootnode: #67
Movecount prune more when not improving #68
NNUE improvements:

Increased hidden layer size of 128 alongside an additional 40m fens: #69
New nets trained on additional data: #71, #72, #74, #78, #79
SCReLU instead of CReLU activation function: #76
Further increase in hidden layer size to 256: #77
All networks mentioned in this release were trained using the bullet trainer by @jw1912, without bullet this release would not have been possible, so thanks a lot for providing this awesome trainer

The provided binaries use avx2, bmi2... (v3) / sse3, popcnt.. (v2) optimization respectively. The avx2bmi2 binary is faster so it should be preferred if your CPU supports these instructions sets. Note that even ryzens before zen 3 should be able to use the bmi2 binaries without slowdown, as they contain no pext / pdep instructions

Contributors 2 :
@rn5f107s2
rn5f107s2
@aytchell
aytchell Hannes Lerchl