TalkChess.com

Looks like you need to examine what the input parser is doing. It is clear that the input is "go ponder wtime 9675 btime 9833 winc 1000 binc 1000" Why it only saw "go ponder"? Sure the engine reads "go ponder wtime 9675 btime 9833 winc 1000 binc 1000" but it gives no more input after that. In the l...

BeyondCritics wrote: ↑

Thu Apr 15, 2021 12:06 pm

You have a bug that seems to break some basic abstractions, so it smells like UB.
Do you have several threads? If so, try to use only a single thread and see if the bug still happens.

I have tried with only 1 thread.
I have used the sanitizers to check the engine but there aren't errors.

I've tried my engine with cutechess-cli and ponder on and I've found a bug. These are the commands sent by cutechess: 111907 >Pedone(0): go ponder wtime 9675 btime 9833 winc 1000 binc 1000 111907 <stockfish(1): info depth 26 seldepth 30 multipv 1 score cp 0 nodes 1177874 nps 1280297 tbhits 0 time 92...

How do you build an android executable with pgo with the NDK? I've tried with these commands: $(NDK)/aarch64-linux-android21-clang ./main.c -s -static -pthread -O3 -flto -std=gnu11 -DNDEBUG -Wall -Wextra -lm -march=armv8-a -DNEON -fPIE -o ./Pedone_armv8 -fprofile-generate qemu-aarch64 ./Pedone_armv8...

Here you can find a translation of the original paper: http://talkchess.com/forum3/viewtopic.php?f=2&t=76250 And here an explanation of how works the neural network in stockfish: https://www.chessprogramming.org/Stockfish_NNUE Question 1: You are correct Question 2: It's efficient because you have a...

Yes it's original, I've not used stockfish code, I've implemented only a similar idea with small differences to the net and with my own training code.
The net has been trained with 300M positions evaluated by Pedone with traditional evaluation at depth 14.

I've just released a new version of my engine. These are the changes: - Pure neural network evaluation - Tuning search parameters - Fix a bug in chess960 mode - Improved futility pruning formula For this release I would like to thank Dale Weiler for the header incbin used to embed the neural network...

I have seen that the site has an old Pedone's version, to have always the latest one I suggest to add a link to Pedone's webpage https://sites.google.com/site/pedonechess

I don't know exactly why but this is the compilation command: Maybe -fPIE could make the difference but I'm not an expert of arm environments.

I would like to test some other things and maybe for January I will release a new version.