Search found 111 matches

by jonkr
Sat Jan 23, 2021 4:29 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: More experiments with neural nets
Replies: 11
Views: 3945

Re: More experiments with neural nets

I guess that your NN topology with 705 inputs and 320 nodes in the first hidden layer is sufficient to roughly equal the best HCE. That's really cool! I do wonder how it actually compares, definitely performs better than my best HCE. Slow 2.5 still scores like -160 elo to Stockfish 11 in my 15s+.15...
by jonkr
Thu Jan 21, 2021 10:59 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: More experiments with neural nets
Replies: 11
Views: 3945

Re: More experiments with neural nets

For the general neural net, the 320 is the first hidden layer, so with input layer it's 705x320x32x32x1. (The endgame nets are Inputs Count x 256 x 32 x 32 x 1, where input count varies depending on how many different piece types can be in the net.) So the total inputs are 705, here's my general neu...
by jonkr
Wed Jan 20, 2021 6:12 pm
Forum: Computer Chess Club: Tournaments and Matches
Topic: CEGT - rating lists January 17th 2021
Replies: 5
Views: 1159

Re: CEGT - rating lists January 17th 2021

Hi Jonathan, did you receive my PM? My proposal for naming your engine in our lists would be: SlowChess BlitzNN 2.5 x64 . IIRC you wanted "Classic" to be deleted?! Just saw your PM now I don't always stay logged in. But yeah SlowChess BlitzNN 2.5 x64 is fine with me. To clarify to anyone wondering,...
by jonkr
Sun Jan 17, 2021 11:15 pm
Forum: Computer Chess Club: General Topics
Topic: SlowChess Blitz Classic 2.0
Replies: 93
Views: 28851

Re: SlowChess Blitz Classic 2.0

I added your v2.5 in my still running FCP Tourney-2021 and changed to v2.5 from today after round 4 out of 50. Now after 160 games place 3 / 41 with 116 / 160 points, 1.5 points more as Ethereal. Seems to be the same strength as Ethereal, but I haven't enough games. Thanks for updating it, glad to ...
by jonkr
Sun Jan 17, 2021 11:09 pm
Forum: Computer Chess Club: Tournaments and Matches
Topic: CEGT - rating lists January 17th 2021
Replies: 5
Views: 1159

Re: CEGT - rating lists January 17th 2021

I'm not sure if I'd consider SlowChess 2.5 NNUE. If it's meant to be a generic term for using a neural net to evaluate the positions on the CPU then it definitely fits, but It might cause some confusion since it's not based on the nodchip/Stockfish NNUE and not compatible. It's possible in future I ...
by jonkr
Sun Jan 17, 2021 10:56 pm
Forum: Computer Chess Club: General Topics
Topic: SlowChess Blitz Classic 2.0
Replies: 93
Views: 28851

Re: SlowChess Blitz Classic 2.0

Unfortunately, it does not work either. :( I've just now updated the zip file download again (slow64-sse.exe should be a newer modified time). I found the max 32-bit instruction is SSE4 also, while max 16-bit is SSE2. So it seems likely this is the remaining issue, but if it still doesn't work I'll...
by jonkr
Sun Jan 17, 2021 7:26 pm
Forum: Computer Chess Club: General Topics
Topic: SlowChess Blitz Classic 2.0
Replies: 93
Views: 28851

Re: SlowChess Blitz Classic 2.0

Thanks for this new version. The nopopcnt version does not work on my old PC (AMD Phenom II x4), 2.4 version works fine. When 2.5 starts thinking (out of the book), it crashes. +1 (Intel Pentium Penryn, Instructions sets: MMX, SSE, SSE2, SSE3, SSSE3, EM64T) Cheers, Tibono I updated the download for...
by jonkr
Sun Jan 17, 2021 3:01 am
Forum: Computer Chess Club: General Topics
Topic: SlowChess Blitz Classic 2.0
Replies: 93
Views: 28851

Re: SlowChess Blitz Classic 2.0

Tibono and Carlos777 Thanks for the report, I found what is likely to be the issue, I accidently used "_mm_packus_epi32" instruction in an optimization for the SSE neural net calculation, instead of "_mm_packs_epi32". The packus version is listed as SSE4, I think everything else I use is SSE2. I had...
by jonkr
Thu Jan 14, 2021 5:58 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Question regarding data structures
Replies: 8
Views: 1321

Re: Question regarding data structures

1 : Are you sure you have optimizations on / release mode? (Also I assume you want to compile x64.) I got 15000 KN/s just using your code, x64 with optimizations on in MS Visual Studio 2019, which seems like too big a difference to be just hardware. 2 : You might want to use a plain array for your m...
by jonkr
Wed Jan 13, 2021 6:47 pm
Forum: Computer Chess Club: Tournaments and Matches
Topic: My test for SlowChess Blitz 2.5 x64-AVX2
Replies: 1
Views: 612

Re: My test for SlowChess Blitz 2.5 x64-AVX2

Thanks for testing, I didn't realize Slow was getting even that close to Komodo 14... if there is next version then maybe will be fully competitive.