Page 9 of 16
Re: Sapeli 1.0 - New chess engine
Posted: Sat Nov 02, 2019 7:43 pm
by JohnWoe
Sapeli 1.63
Source code:
https://github.com/SamuraiDangyo/Sapeli
Release:
https://github.com/SamuraiDangyo/Sapeli ... /tag/v1.63
There was actually some bugs in perft function. The total perft suite is actually: 32,362,126,796 nodes.
Time management is greatly simplified.
AiLevel was much improved.
Now AiLevel is much better. Some quick tests.
AiLevel 0 as random mover loses all games as random mover should.
Level 1/100 scores 2%.
10/100 scores 9%. etc.
Seems to work great now.
Code: Select all
AiLevel: 0
Score of Sapeli 1.63 vs Fairy-Max 5.0b: 0 - 276 - 0 [0.000] 276
AiLevel: 1
Score of Sapeli 1.63 vs Fairy-Max 5.0b: 2 - 139 - 4 [0.028] 145
AiLevel: 10
Score of Sapeli 1.63 vs Fairy-Max 5.0b: 24 - 326 - 20 [0.092] 370
AiLevel: 30
Score of Sapeli 1.63 vs Fairy-Max 5.0b: 16 - 126 - 17 [0.154] 159
AiLevel: 50
Score of Sapeli 1.63 vs Fairy-Max 5.0b: 46 - 112 - 21 [0.316] 179
AiLevel: 80
Score of Sapeli 1.63 vs Fairy-Max 5.0b: 74 - 39 - 36 [0.617] 149
AiLevel: 100
Score of Sapeli 1.63 vs Fairy-Max 5.0b: 82 - 5 - 17 [0.870] 104
Re: Sapeli 1.0 - New chess engine
Posted: Wed Nov 06, 2019 11:41 pm
by JohnWoe
I wanted to check has there been any progress in terms of strength. It is faster and less buggy that's I know without testing.
I ran a little tournaments against older versions using Chess960 games. I test against Stockfish just to see there's no illegal moves/crashes.
Yes, the latest version is the strongest as it should be.
Code: Select all
Rank Name Elo +/- Games Score Draws
1 Sapeli 1.64 66 32 400 59.4% 13.3%
2 Sapeli 1.02 55 32 400 57.9% 14.2%
3 Sapeli 1.38 36 32 400 55.1% 13.3%
4 Sapeli 1.21 -61 32 400 41.4% 13.8%
5 Sapeli 1.16 -98 33 400 36.3% 12.5%
Finished match
Re: Sapeli 1.0 - New chess engine
Posted: Thu Nov 07, 2019 11:31 am
by Gabor Szots
JohnWoe wrote: ↑Wed Nov 06, 2019 11:41 pm
I wanted to check has there been any progress in terms of strength. It is faster and less buggy that's I know without testing.
I ran a little tournaments against older versions using Chess960 games. I test against Stockfish just to see there's no illegal moves/crashes.
Yes, the latest version is the strongest as it should be.
Code: Select all
Rank Name Elo +/- Games Score Draws
1 Sapeli 1.64 66 32 400 59.4% 13.3%
2 Sapeli 1.02 55 32 400 57.9% 14.2%
3 Sapeli 1.38 36 32 400 55.1% 13.3%
4 Sapeli 1.21 -61 32 400 41.4% 13.8%
5 Sapeli 1.16 -98 33 400 36.3% 12.5%
Finished match
Still no chance for a Windows compile?
Re: Sapeli 1.0 - New chess engine
Posted: Thu Nov 07, 2019 11:49 am
by Guenther
Gabor Szots wrote: ↑Thu Nov 07, 2019 11:31 am
...
Still no chance for a Windows compile?
I created one for my architecture only, old quadcore (yorkfield), no popcount, SSE 4.1.
If you are interested I can add it here, (if talkchess has no hickups anymore for connection)
(connection problems since around 15 minutes from here)
Re: Sapeli 1.0 - New chess engine
Posted: Thu Nov 07, 2019 12:02 pm
by Gabor Szots
Guenther wrote: ↑Thu Nov 07, 2019 11:49 am
Gabor Szots wrote: ↑Thu Nov 07, 2019 11:31 am
...
Still no chance for a Windows compile?
I created one for my architecture only, old quadcore (yorkfield), no popcount, SSE 4.1.
If you are interested I can add it here, (if talkchess has no hickups anymore for connection)
(connection problems since around 15 minutes from here)
Yes, I am interested. And sooner or later I may learn how to compile it myself.
Re: Sapeli 1.0 - New chess engine
Posted: Thu Nov 07, 2019 12:07 pm
by Guenther
Gabor Szots wrote: ↑Thu Nov 07, 2019 12:02 pm
Guenther wrote: ↑Thu Nov 07, 2019 11:49 am
Gabor Szots wrote: ↑Thu Nov 07, 2019 11:31 am
...
Still no chance for a Windows compile?
I created one for my architecture only, old quadcore (yorkfield), no popcount, SSE 4.1.
If you are interested I can add it here, (if talkchess has no hickups anymore for connection)
(connection problems since around 15 minutes from here)
Yes, I am interested. And sooner or later I may learn how to compile it myself.
Well, the quota for adding files is very low?
(140KB was too much already)
Sorry, the board attachment quota has been reached.
Get it here:
http://rwbc-chess.de/Downloads/Engines/Sapeli_164na.7z
Re: Sapeli 1.0 - New chess engine
Posted: Thu Nov 07, 2019 12:44 pm
by Gabor Szots
Re: Sapeli 1.0 - New chess engine
Posted: Mon Nov 25, 2019 6:28 pm
by Guenther
Gabor Szots wrote: ↑Thu Nov 07, 2019 12:44 pm
Thanks Günther.
Changed the name and thus the download now is called:
http://rwbc-chess.de/Downloads/Engines/Sapeli_164old.7z
Re: Sapeli 1.0 - New chess engine
Posted: Wed Dec 11, 2019 1:02 am
by JohnWoe
^ Thanks Günther! I added your build to the list:
https://github.com/SamuraiDangyo/Sapeli ... naries/bin
I released Sapeli 1.65:
Source code:
https://github.com/SamuraiDangyo/Sapeli
Release:
https://github.com/SamuraiDangyo/Sapeli ... /tag/v1.65
I have those other compilers installed on my laptop so I'll add Android + other builds later.
With this release I profiled and profiled Sapeli. And squeezed every NPS out I could. It is fast.
Some tuning to LMR. Slightly more aggressive pruning. Split search into smaller pieces.
1.65 searches a bit deeper than 1.64. But that wasn't the goal.
Source code polishing.
Strength wise on par with 1.64. Slightly stronger on my tests.
Code: Select all
benchmarks
1.65
= nodes 120041499 mnps 4.001 time 30.000
1.64
= nodes 114280936 mnps 3.808 time 30.010
some tests
Score of Sapeli 1.65 vs Sapeli 1.64: 414 - 393 - 193 [0.510] 1000
Elo difference: 7.30 +/- 19.34
Finished match
Re: Sapeli 1.0 - New chess engine
Posted: Wed Dec 11, 2019 6:02 pm
by Guenther
Compiled a few flavours for 1.65, with a big caveat, because I had never built versions before, I cannot run here myself,
which is a bit awkward.
Would be nice if someone could try the bmi or bmi2 version, if those run at all and with what speed.
The other two are are tested here and behave very similar at least on my old machine.
http://rwbc-chess.de/Downloads/Engines/Sapeli_165.7z