New J.A Compiles

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

Moderators: hgm, Rebel, chrisw

User avatar
Jim Ablett
Posts: 1485
Joined: Fri Jul 14, 2006 7:56 am
Location: London, England
Full name: Jim Ablett

Re: New J.A Compiles

Post by Jim Ablett »

chessica wrote: Sat Jun 01, 2024 11:03 pm I think that I have already used your compiles very successfully under Windows 98, and I would like to thank you very much for that too.
You're welcome :)

Jim.
User avatar
Jim Ablett
Posts: 1485
Joined: Fri Jul 14, 2006 7:56 am
Location: London, England
Full name: Jim Ablett

Re: New J.A Compiles

Post by Jim Ablett »

peter wrote: Sat Jun 01, 2024 7:22 pm
Jim Ablett wrote: Sat Jun 01, 2024 4:14 pm Stockfish Plus 2022 64 ja

Windows/Linux & Android clang profiled builds.

https://drive.proton.me/urls/PFTA2JRKMR#GTek6jWkUPbA
Thanks for that too. Firefather's one yet is already of the kind of ShashChess and Brainlearn, that uses MCTS always in combination with A-B- search only, one thread has to be free for this one main line, so no "pure" MCTS- search is to be got into output except in debug- mode. The latest one with just MCTS (without A-B) I've read about so far, is Kelly Kinyama's:

https://github.com/Kellykinyama12/Montecarlo

BTW it's given at FireFather's github- site as example for MCTS- SF too.
I fear it's not fully bug- free neither, but already the compile from Nicolet's version you made, works well enough for me to get much information out of it, so a direct comparison to Kinyama's one would be very interesting to me too. Again bmi2 was the only compile, I'd need for my own use
:roll:
Stockfish Montecarlo 64 ja bmi2 (Kelly Kinyama)

https://drive.proton.me/urls/PGQ75WFTHR#UrG1M7iOQeVH

Jim.
peter
Posts: 3204
Joined: Sat Feb 16, 2008 7:38 am
Full name: Peter Martan

Re: New J.A Compiles

Post by peter »

Jim Ablett wrote: Sun Jun 02, 2024 12:42 am Stockfish Montecarlo 64 ja bmi2 (Kelly Kinyama)

https://drive.proton.me/urls/PGQ75WFTHR#UrG1M7iOQeVH
You're the very best!
Runs well in Shredder, but are you sure, it's really MCTS?
Shows as much n/s and time to depth as A-B SF of that time did afair, and I don't see any MCTS- specific UCI commands.
A benchmark:

Stockfish 010624 64 BMI2 by T. Romstad, M. Costalba, J. Kiiski, G. Linscott
bench 4096 30 25
===========================
Total time (ms) : 2435
Nodes searched : 11623820
Nodes/second : 4773642
Peter.
User avatar
Jim Ablett
Posts: 1485
Joined: Fri Jul 14, 2006 7:56 am
Location: London, England
Full name: Jim Ablett

Re: New J.A Compiles

Post by Jim Ablett »

peter wrote: Sun Jun 02, 2024 1:21 am
Jim Ablett wrote: Sun Jun 02, 2024 12:42 am Stockfish Montecarlo 64 ja bmi2 (Kelly Kinyama)

https://drive.proton.me/urls/PGQ75WFTHR#UrG1M7iOQeVH
You're the very best!
Runs well in Shredder, but are you sure, it's really MCTS?
Shows as much n/s and time to depth as A-B SF of that time did afair, and I don't see any MCTS- specific UCI commands.
A benchmark:

Stockfish 010624 64 BMI2 by T. Romstad, M. Costalba, J. Kiiski, G. Linscott
bench 4096 30 25
===========================
Total time (ms) : 2435
Nodes searched : 11623820
Nodes/second : 4773642

montecarlo.cpp in the src code, makefile and referenced in search.cpp,

Code: Select all

multiPV = std::min(multiPV, rootMoves.size());

  bool ABnetwork = network;
  if(!mainThread)
	network = true;


  if (USE_MONTE_CARLO
  && (mainThread || (!mainThread && ABnetwork))
  )
      MonteCarlo(rootPos).test();
  else

  // Iterative deepening loop until requested to stop or the target depth is reached
  while (   (rootDepth += ONE_PLY) < DEPTH_MAX
         && !Threads.stop
         && !(Limits.depth && mainThread && rootDepth /
Jim.
User avatar
Jim Ablett
Posts: 1485
Joined: Fri Jul 14, 2006 7:56 am
Location: London, England
Full name: Jim Ablett

Re: New J.A Compiles

Post by Jim Ablett »

Jim Ablett wrote: Sun Jun 02, 2024 1:58 am
peter wrote: Sun Jun 02, 2024 1:21 am
Jim Ablett wrote: Sun Jun 02, 2024 12:42 am Stockfish Montecarlo 64 ja bmi2 (Kelly Kinyama)

https://drive.proton.me/urls/PGQ75WFTHR#UrG1M7iOQeVH
You're the very best!
Runs well in Shredder, but are you sure, it's really MCTS?
Shows as much n/s and time to depth as A-B SF of that time did afair, and I don't see any MCTS- specific UCI commands.
A benchmark:

Stockfish 010624 64 BMI2 by T. Romstad, M. Costalba, J. Kiiski, G. Linscott
bench 4096 30 25
===========================
Total time (ms) : 2435
Nodes searched : 11623820
Nodes/second : 4773642

montecarlo.cpp in the src code, makefile and referenced in search.cpp,

Code: Select all

multiPV = std::min(multiPV, rootMoves.size());

  bool ABnetwork = network;
  if(!mainThread)
	network = true;


  if (USE_MONTE_CARLO
  && (mainThread || (!mainThread && ABnetwork))
  )
      MonteCarlo(rootPos).test();
  else

  // Iterative deepening loop until requested to stop or the target depth is reached
  while (   (rootDepth += ONE_PLY) < DEPTH_MAX
         && !Threads.stop
         && !(Limits.depth && mainThread && rootDepth /
Jim.

There were two versions of 'montecarlo.cpp' included in the src code and when looking at the code there are slight differences.
I have re-compiled again with the other version, same download link.

Jim.
peter
Posts: 3204
Joined: Sat Feb 16, 2008 7:38 am
Full name: Peter Martan

Re: New J.A Compiles

Post by peter »

Jim Ablett wrote: Sun Jun 02, 2024 2:44 am There were two versions of 'montecarlo.cpp' included in the src code and when looking at the code there are slight differences.
I have re-compiled again with the other version, same download link.
Trojan:Win32/Wacatac.B!ml
:(
:)
Peter.
chessica
Posts: 672
Joined: Thu Aug 11, 2022 11:30 pm
Full name: Esmeralda Pinto

Re: New J.A Compiles

Post by chessica »

peter wrote: Sun Jun 02, 2024 1:21 am
Jim Ablett wrote: Sun Jun 02, 2024 12:42 am Stockfish Montecarlo 64 ja bmi2 (Kelly Kinyama)

https://drive.proton.me/urls/PGQ75WFTHR#UrG1M7iOQeVH
...
Shows as much n/s and time to depth as A-B SF of that time did afair, and I don't see any MCTS- specific UCI commands.
...
Hello Peter, what kind of UCI commands would you expect? Can you please show us an example?
User avatar
Jim Ablett
Posts: 1485
Joined: Fri Jul 14, 2006 7:56 am
Location: London, England
Full name: Jim Ablett

Re: New J.A Compiles

Post by Jim Ablett »

peter wrote: Sun Jun 02, 2024 8:26 am
Jim Ablett wrote: Sun Jun 02, 2024 2:44 am There were two versions of 'montecarlo.cpp' included in the src code and when looking at the code there are slight differences.
I have re-compiled again with the other version, same download link.
Trojan:Win32/Wacatac.B!ml
:(
:)
I think it's because I zipped the src code. Your antivirus is triggered by seeing a zip file within a zip file.
I wont do that in future.

Jim.
peter
Posts: 3204
Joined: Sat Feb 16, 2008 7:38 am
Full name: Peter Martan

Re: New J.A Compiles

Post by peter »

chessica wrote: Sun Jun 02, 2024 10:23 am [what kind of UCI commands would you expect? Can you please show us an example?
Something like this from ShashChess

option name MCTS Multi Strategy type spin default 20 min 0 max 100
option name MCTS Multi MinVisits type spin default 5 min 0 max 1000

or that from Dragon

option name MCTS Explore type spin default 40 min 15 max 500
option name MCTS Explore Root type spin default 60 min 15 max 500

, even if Nicolet's one doesn't have anything of that kind neither.
Peter.
User avatar
Jim Ablett
Posts: 1485
Joined: Fri Jul 14, 2006 7:56 am
Location: London, England
Full name: Jim Ablett

Re: New J.A Compiles

Post by Jim Ablett »

Image
Vault 0.4.1 by Morgan Houppin
https://github.com/mhouppin/vault
Logo by by Dusan Stamenkovic

Vault 0.4.1 64 ja

Windows/Linux & Android profiled builds

https://drive.proton.me/urls/5YH680WVF0#UTdWl9qBvR8U

Jim.