New engine releases 2020

Discussion of anything and everything relating to chess playing software and machines.

Moderators: Harvey Williamson, Dann Corbit, hgm

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
mhib
Posts: 18
Joined: Sun Apr 21, 2019 5:55 pm
Full name: Marcin Henryk Bartkowiak

Re: New engine releases 2020

Post by mhib » Tue May 19, 2020 6:32 pm

Combusken 1.2.0 released
https://github.com/mhib/combusken/releases/tag/v1.2.0

Changes:
  • Sending Time and NPS to UI
  • ProbCut pruning
  • Staged move generation
  • Save best move in TT, even if it did not exceed alpha


Unfortunately the binary for android does not support syzygy.

Should be ~100 ELO stronger than Combusken 1.1.0 .

User avatar
CMCanavessi
Posts: 910
Joined: Thu Dec 28, 2017 3:06 pm
Location: Argentina

Re: New engine releases 2020

Post by CMCanavessi » Wed May 20, 2020 5:52 am

100 elo! wow... :twisted:
Follow my tournament and some Leela gauntlets live at http://twitch.tv/ccls

User avatar
Graham Banks
Posts: 34652
Joined: Sun Feb 26, 2006 9:52 am
Location: Auckland, NZ

Re: New engine releases 2020

Post by Graham Banks » Thu May 21, 2020 8:25 am

gbanksnz at gmail.com

fabianVDW
Posts: 142
Joined: Fri Mar 15, 2019 7:46 pm
Location: Germany
Full name: Fabian von der Warth

Re: New engine releases 2020

Post by fabianVDW » Fri May 22, 2020 2:16 pm

I've released a new version of FabChess at https://github.com/fabianvdW/FabChess/releases.


Version 1.15 comes with a significant strength increase over the last released version 1.14.1
Changes include:
-A lot of background refactoring to make the code cleaner and faster
-All in all speed improvements of around 20%, more coming in the future
-Use tt info for avoiding null moves
-Reset killer heuristic for grandchildren
-More LMR,
-Late move pruning
-SEE Pruning of quiet moves
-Simplification of qsearch
-Fix of endgame evaluation for most common endgames. FabChess will no longer report bogus evaluation and lose games because of that (hopefully).

Selfplay tests show:
ELO | 62.80 +- 5.19 (95%)
SPRT | 60.0+0.6s Threads=1 Hash=64MB
LLR | 0.07 (-2.94, 2.94) [0.00, 0.01]
Games | N: 8160 W: 2666 L: 1207 D: 4287

I didn't do enough testing against other engines to confirm that much of a gain, but It should be close to that.

A new type of executable will be available next to "normal" and "popcnt" namely "BMI2", which will use PEXT, so don't use that executable if your hardware does not support this instruction set natively. BMI2 will be the fastest exectuable, followed by popcnt and then normal will be very slow. Please let me know if there are any problems with the executables.
Author of FabChess: https://github.com/fabianvdW/FabChess
A UCI compliant chess engine written in Rust.
FabChessWiki: https://github.com/fabianvdW/FabChess/wiki
fabianvonderwarth@gmail.com

Gabor Szots
Posts: 656
Joined: Sat Jul 21, 2018 5:43 am
Location: Szentendre, Hungary
Full name: Gabor Szots

Re: New engine releases 2020

Post by Gabor Szots » Fri May 22, 2020 5:35 pm

Thanks Fabian.
Gabor Szots
CCRL testing group

MOBMAT
Posts: 327
Joined: Sat Feb 04, 2017 10:57 pm
Location: USA

Re: New engine releases 2020

Post by MOBMAT » Sat May 23, 2020 12:49 am

Bagatur zip file isn't recognized as an archive.
i7-6700K @ 4.00Ghz (using 6 threads), EGTBs on PCI SSD
Benchmark: Stockfish 11 64 bmi2 (nps): 2067669

MOBMAT
Posts: 327
Joined: Sat Feb 04, 2017 10:57 pm
Location: USA

Re: New engine releases 2020

Post by MOBMAT » Sun May 24, 2020 9:12 pm

mhib wrote:
Tue May 19, 2020 6:32 pm
Combusken 1.2.0 released
https://github.com/mhib/combusken/releases/tag/v1.2.0

Changes:
  • Sending Time and NPS to UI
  • ProbCut pruning
  • Staged move generation
  • Save best move in TT, even if it did not exceed alpha


Unfortunately the binary for android does not support syzygy.

Should be ~100 ELO stronger than Combusken 1.1.0 .
Does Combusken support UCI analysis mode? I tried to analyse a position in arena and the engine locked up. I tried the .exe from the command line, entered UCI and then 'go' and the program exited. i was using the x64 version
i7-6700K @ 4.00Ghz (using 6 threads), EGTBs on PCI SSD
Benchmark: Stockfish 11 64 bmi2 (nps): 2067669

jorose
Posts: 322
Joined: Thu Jan 22, 2015 2:21 pm
Location: Zurich, Switzerland
Full name: Jonathan Rosenthal

Re: New engine releases 2020

Post by jorose » Tue May 26, 2020 5:39 am

Winter v0.8 has been released!

For unclear reasons, clang is giving a huge N/s boost in this new release. As a result of not having this boost, it is actually a regression compared to v0.7 on my test system. That being said, I am confident you will find this new version to be a nice improvement.

I think stylistically Winter has become more positional and not quite as agressive, which many will not like. That being said I feel its style is still very appealing to me and I am interested in what your impressions are.

Tomorrow, I might make a more in depth post as I did for early Winter versions, especially as I kind of was under the radar for a couple months.

https://github.com/rosenthj/Winter/releases/tag/v0.8
-Jonathan

mhib
Posts: 18
Joined: Sun Apr 21, 2019 5:55 pm
Full name: Marcin Henryk Bartkowiak

Re: New engine releases 2020

Post by mhib » Tue May 26, 2020 7:35 am

MOBMAT wrote:
Sun May 24, 2020 9:12 pm
mhib wrote:
Tue May 19, 2020 6:32 pm
Combusken 1.2.0 released
https://github.com/mhib/combusken/releases/tag/v1.2.0

Changes:
  • Sending Time and NPS to UI
  • ProbCut pruning
  • Staged move generation
  • Save best move in TT, even if it did not exceed alpha


Unfortunately the binary for android does not support syzygy.

Should be ~100 ELO stronger than Combusken 1.1.0 .
Does Combusken support UCI analysis mode? I tried to analyse a position in arena and the engine locked up. I tried the .exe from the command line, entered UCI and then 'go' and the program exited. i was using the x64 version
I'll look into it.

matejst
Posts: 218
Joined: Mon May 14, 2007 6:20 pm
Full name: Boban Stanojević

Re: New engine releases 2020

Post by matejst » Tue May 26, 2020 2:29 pm

Jonathan,

Winter 0.7 was positionally excellent already, and I liked its evaluation. I use it to analyze my opening repertoire and whole games in these openings, and its output makes sense. I guess it is because it doesn't reach absurd depths too fast, so I can understand it. On the other hand, I do not use it for matches, so I am not influenced by the rating things etc.

The evaluation is at times a bit optimistic, but I thought it was mainly at lower depths. With enough time, I thought it was quite reasonable. It missed (misevaluated) tactical shots from time to time, but the alternatives were OK. It reminds me of Komodo 7 and 8. I missed the fact that it was not user friendly enough (no multi-pv, no learning, no real options), but it was all right nonetheless.

I followed the 2. league of the TCEC and I liked Winter's play. It was impressive in the last round win against Booot.

Winter opens a new path for chess engines development, and I hope you will continue to work on it.

Post Reply