GyatsoYT wrote: ↑Sat Dec 13, 2025 6:05 pm
Hey there thanks for letting me know about the issue with threading.
Also if you don't mind what was the nps you got on single thread while searching like a depth of 20 vs multithreaded running on 12 cores / 24 threads if your cpu supports it?
I don't know if it is an issue of Gyatso with threading, it might as well be a Windows issue, Linux is better in this respect.
When I run Gyatso from the command-line it gives a little bit more than 1.4 Mn/s, with 12 threads it is approx. 16 Mn/s.
When I run it with 12 threads from icsdroneng, the speed varies between 5.5 and 10 Mn/s, sometimes even lower, which seems strange.
The CPU it runs on has 12 cores, so I use 12 threads, running with 24 threads on 12 cores will increase the n/s somewhat but it makes an engine play worse.
GyatsoYT wrote: ↑Sat Dec 13, 2025 6:05 pm
Hey there thanks for letting me know about the issue with threading.
Also if you don't mind what was the nps you got on single thread while searching like a depth of 20 vs multithreaded running on 12 cores / 24 threads if your cpu supports it?
I don't know if it is an issue of Gyatso with threading, it might as well be a Windows issue, Linux is better in this respect.
When I run Gyatso from the command-line it gives a little bit more than 1.4 Mn/s, with 12 threads it is approx. 16 Mn/s.
When I run it with 12 threads from icsdroneng, the speed varies between 5.5 and 10 Mn/s, sometimes even lower, which seems strange.
The CPU it runs on has 12 cores, so I use 12 threads, running with 24 threads on 12 cores will increase the n/s somewhat but it makes an engine play worse.
Since it shows a reasonable growth in nps when ran through command line , but doesn't show the same growth on icsdroneng seems to be a problem of either windows or the gui's. But I'll still look into it to find out if this is actually a code issue or not.
Belfish is still a little bit to strong for mortal engines and won the tournament. Marvin did pretty good too.
Eendje noticed too late that the number of players was odd, so she had to add a BYE instead of another (spare) engine.
This time there were some issues with engines refusing to move.
Joost Buijs wrote: ↑Sat Dec 13, 2025 10:31 pm
Belfish is still a little bit to strong for mortal engines and won the tournament. Marvin did pretty good too.
Eendje noticed too late that the number of players was odd, so she had to add a BYE instead of another (spare) engine.
This time there were some issues with engines refusing to move.
Bart Weststrate wrote: ↑Sun Dec 14, 2025 9:18 am
Joost sorry for my ignorance but what is Belfish?
Edit: you explained this earlier:
Belfish is a modified Stockfish run on a very small computer by Wieland Belka who is also the author of 'Eichoernchen'.
Acording to his finger-notes Belfish runs on bigger hardware than before:
1: BelFish - a StockFish derivative, modified, maintained and operated by belka
2: experimenting with search and evaluation function, with NNUE databases, with small techniques for speeding up code execution, with compiling options ...
3: running on an AMD mini comp at 8 cores, Win11 Pro x64, 32 GB RAM
