Yes, it works fine with 4 CPUs and I like Seer! Thank you for compiling it also for unsupported / older / strange systems
Regards, Alex
New engine releases & news 2021
Moderators: hgm, Rebel, chrisw
-
- Posts: 1533
- Joined: Sat Feb 06, 2021 8:06 am
- Full name: Alex Morales
Re: New engine releases & news 2021
Chess engines and dedicated chess computers fan since 1981 Mac mini M1 8GB-256GB, Windows 11 & Ubuntu ARM64.
ProteusSF Dev Forum TROLLS KINDERGARTEN
ProteusSF Dev Forum TROLLS KINDERGARTEN
-
- Posts: 543
- Joined: Sun Sep 06, 2020 4:40 am
- Full name: Connor McMonigle
Re: New engine releases & news 2021
Just to clarify, you're using my sse3 compile, correct? If so, it would be appreciated if you could mark it as such if it's not too much effort on your part (assign it the name "Seer 2.3.0-sse3" as you've done for Koivisto preferably). ThanksRebel wrote: ↑Fri Aug 13, 2021 10:38 amYou are going fastconnor_mcmonigle wrote: ↑Fri Aug 13, 2021 3:23 am I've released a new version (v2.3.0) of Seer here: https://github.com/connormcmonigle/seer ... tag/v2.3.0
It should be significantly stronger than the previous release (v2.2.0) and tested at >= 100 elo in self play. As with all previous versions >= v2.0.0 Seer is trained on unique data generated solely from Seer's search+EGTB and does not rely on training code/inference code etc. derived from other engines. Training code and data generation code can be found at https://github.com/connormcmonigle/seer ... e/selfplay.
I've attempted to simplify my naming conventions for binaries. As with the previous release, the SSE3+nopopcnt binary is marked as unofficial and I would prefer if testers refrained from using it for official rating lists/tournaments.
https://prodeo.actieforum.com/t358p25-g ... sults#4061
-
- Posts: 7025
- Joined: Thu Aug 18, 2011 12:04 pm
- Full name: Ed Schröder
Re: New engine releases & news 2021
I am using : seer_v2.3_x64_avx_popcnt.exe
How would you like to see it named?
How would you like to see it named?
90% of coding is debugging, the other 10% is writing bugs.
-
- Posts: 543
- Joined: Sun Sep 06, 2020 4:40 am
- Full name: Connor McMonigle
-
- Posts: 7025
- Joined: Thu Aug 18, 2011 12:04 pm
- Full name: Ed Schröder
Re: New engine releases & news 2021
Because it's no AVX2?connor_mcmonigle wrote: ↑Fri Aug 13, 2021 6:21 pmNevermind in that case- all is well. Thanks. However, I'm confused as to why Finn thought it necessary to send you a SSE compile of Koivisto if your hardware has AVX...
90% of coding is debugging, the other 10% is writing bugs.
-
- Posts: 543
- Joined: Sun Sep 06, 2020 4:40 am
- Full name: Connor McMonigle
Re: New engine releases & news 2021
AVX2 only differs from AVX due to the addition of FMA instructions. In fact, the same code path can be used for AVX and AVX2 as the compiler is smart enough to perform fma contraction. An AVX Koivisto binary would probably out perform an SSE binary by a significant margin. Hence my confusion.Rebel wrote: ↑Fri Aug 13, 2021 6:22 pmBecause it's no AVX2?connor_mcmonigle wrote: ↑Fri Aug 13, 2021 6:21 pmNevermind in that case- all is well. Thanks. However, I'm confused as to why Finn thought it necessary to send you a SSE compile of Koivisto if your hardware has AVX...
-
- Posts: 138
- Joined: Thu Jun 18, 2020 9:20 pm
- Full name: Finn Eggers
Re: New engine releases & news 2021
Yeah sorry. We brought out AVX2, AVX512 and SSE compiles. I realized that we don’t even require AVX2 specific code. This will be fixed with the next release.connor_mcmonigle wrote: ↑Fri Aug 13, 2021 6:27 pmAVX2 only differs from AVX due to the addition of FMA instructions. In fact, the same code path can be used for AVX and AVX2 as the compiler is smart enough to perform fma contraction. An AVX Koivisto binary would probably out perform an SSE binary by a significant margin. Hence my confusion.Rebel wrote: ↑Fri Aug 13, 2021 6:22 pmBecause it's no AVX2?connor_mcmonigle wrote: ↑Fri Aug 13, 2021 6:21 pmNevermind in that case- all is well. Thanks. However, I'm confused as to why Finn thought it necessary to send you a SSE compile of Koivisto if your hardware has AVX...
-
- Posts: 268
- Joined: Fri Mar 17, 2006 8:01 am
- Location: Russia
- Full name: Vladimir Medvedev
Re: New engine releases & news 2021
Experimental version of GreKo with well-known Stockfish's NNUE:
http://greko.su/greko-2021.08-nnue.zip
Requires AVX2.
http://greko.su/greko-2021.08-nnue.zip
Requires AVX2.
-
- Posts: 135
- Joined: Tue Mar 05, 2019 3:43 pm
- Full name: Archimedes
Re: New engine releases & news 2021
It only works, when network.nnue is in the current working directory (tested on DroidFish). On DroidFish this is the logs directory. Moving the network file to another directory or changing the name of the network file and change that in the uci options doesn't work (GreKo do not find the network file). It seems, GreKo ignores the setting for the NNUE file and looks always for network.nnue in the current working directory.WinPooh wrote: ↑Mon Aug 16, 2021 12:05 pm Experimental version of GreKo with well-known Stockfish's NNUE:
http://greko.su/greko-2021.08-nnue.zip
Requires AVX2.
-
- Posts: 268
- Joined: Fri Mar 17, 2006 8:01 am
- Location: Russia
- Full name: Vladimir Medvedev
Re: New engine releases & news 2021
Thank you for bug report. The problem was in incorrect parsing UCI options with spaces in their names. Changed names to "NNUE" and "Network", now it should work. Tested in Fritz GUI, file re-uploaded.Archimedes wrote: ↑Tue Aug 17, 2021 8:08 amIt only works, when network.nnue is in the current working directory (tested on DroidFish). On DroidFish this is the logs directory. Moving the network file to another directory or changing the name of the network file and change that in the uci options doesn't work (GreKo do not find the network file). It seems, GreKo ignores the setting for the NNUE file and looks always for network.nnue in the current working directory.WinPooh wrote: ↑Mon Aug 16, 2021 12:05 pm Experimental version of GreKo with well-known Stockfish's NNUE:
http://greko.su/greko-2021.08-nnue.zip
Requires AVX2.