PlentyChess - A UCI chess engine with a neural network based evaluation.
Rating Chess Engines Diary CEDR=3694
Building on Windows
For the build to work on Windows, you need to install MinGW as well as the required packages (make, g++, pthread, etc.).
Then you can use the makefile as you would on a unix system.
v.2.1.0:
This is mainly a bugfix release, since there is an issue with the UCI Chess960 / Ponder options in some GUIs.
Performance vs. PlentyChess 2.0.0
STC
Elo | 27.49 +- 7.20 (95%)
Conf | 8.0+0.08s Threads=1 Hash=16MB
Games | N: 2508 W: 685 L: 487 D: 1336
Penta | [12, 214, 605, 410, 13]
https://chess.aronpetkovski.com/test/1154/
LTC
Elo | 31.33 +- 6.67 (95%)
Conf | 40.0+0.40s Threads=1 Hash=64MB
Games | N: 2502 W: 697 L: 472 D: 1333
Penta | [5, 186, 640, 419, 1]
Improvements
This version includes the following improvements compared to version 2.0.0:
Correctly expose UCI "check" type options, fixing compatability with certain GUIs
Stronger neural network, trained longer and on more data
Much better thread voting, which makes use of the data of previous search iterations
A big search and time management tune
Improvements to move history updates
Binary guide
generic: Slowest
avx2: Faster, compatible with most modern CPUs
bmi2: Faster, includes avx2, not recommended for Zen1/2 CPUs
avx512: Fastest, includes bmi2
Windows/Linux & Android clang profiled builds
Jim Ablett compiles:
Comments
Post a Comment