Mystery engine at CCC

[ImNotStockfish]

As for search improvements and such, I'll try some things now and again. But mostly things that are already present in other engines I'm afraid. Any original ideas I have would go into Torch.

You say that the collaboration with chesscom will help the community, be a net positive, but that is just due to Openbench and with some caveats.

Openbench already existed before Torch, chesscom might improve it which is nice but not because they want to help the community in general but because of their own interests. As you already said before, you could close source Openbench but that is something that you wouldn't even have a reason to do before now.

Now you also made it clear that you guys are not willying to help the open source chess engine community anymore. If you guys come up with a good and original idea, you will not try to implement it in Stockfish or any other open source engine like you would before, now only in Torch, only for chesscom.

So, all of this brings competition (good), but said competition is a close source engine (bad), with an uncertain future, it might be paid or even exclusively available at chesscom (bad). This also brings improvements to Openbench (good), but you have a reason to close source it (bad). Chesscom is donating hardware to improve chess engines (good) but only their own engine (bad). To finish, the open source community lost a lot of great developers (horrible) and you guys make some money (good).

Now, is this really a net positive? I guess it depends on who you ask.

[Ajedrecista]

Hello Sam:

As for search improvements and such, I'll try some things now and again. But mostly things that are already present in other engines I'm afraid. Any original ideas I have would go into Torch.

You say that the collaboration with chesscom will help the community, be a net positive, but that is just due to Openbench and with some caveats.

Openbench already existed before Torch, chesscom might improve it which is nice but not because they want to help the community in general but because of their own interests. As you already said before, you could close source Openbench but that is something that you wouldn't even have a reason to do before now.

Now you also made it clear that you guys are not willying to help the open source chess engine community anymore. If you guys come up with a good and original idea, you will not try to implement it in Stockfish or any other open source engine like you would before, now only in Torch, only for chesscom.

[...]

I wish good luck with the new project and may the best win. Can we expect Rocky-like hard matches? The triplet SF-Torch-Lc0 might resemble Rocky, Apollo and the villain on duty, though without villains in the computer chess triplet. Just a bunch of draws and the few wins/loses being decisive like a good cross punch.

Time will tell if Torch is good or bad for computer chess. Remember that SF is seemingly hitting a wall —slower and slower progress— unless a major breakthrough appears. If Torch ideas do not reach SF framework, it would be a good stress test to see how SF team holds, but there are many people in SF team to just give up, IMHO, so I think they will more less keep the current slow pace of progress. Elo race becomes more difficult!

------------

[...]

So, all of this brings competition (good), but said competition is a close source engine (bad), with an uncertain future, it might be paid or even exclusively available at chesscom (bad). This also brings improvements to Openbench (good), but you have a reason to close source it (bad). Chesscom is donating hardware to improve chess engines (good) but only their own engine (bad). To finish, the open source community lost a lot of great developers (horrible) and you guys make some money (good).

[...]

A bell rang when I read this good-bad debate: are you sure you are not a scriptwriter? I have seen something similar before:

https://www.youtube.com/watch?v=i73-lpVV-Jo

From 'Treehouse of Horror III', The Simpsons S04E05 (code 9F04), first aired on October 29th, 1992.

Regards from Spain.

Ajedrecista.

[Eduard]

The number 1 engine is the one that wins the TCEC tournament. Everyone agrees: This is the best engine!

If Chess.com comes out with their own engine, then I don't want to trust Chess.com tournaments anymore. The ranked tournaments are all well and good too but Lc0 becomes really strong with strong GPU. Unfortunately, the testers don't offer that. The super final at TCEC is played on a classic Timecontrol and super hardware (for Lc0 important). Whoever wins this is No. 1.

[smatovic]

As far as I got it, Larry Kaufman and Mark Lefler are handing over the Torch to the OpenBench guys maybe it is not a bad thing, to have some commercial, closed-source chess programmers out there....bon chance.

--
Srdja

[jshriver]

Maybe it's Houdini 7

An update to Boris...

I dont know what Boris is but as a huge fan of vintage computer chess equipment I want one.

[ImNotStockfish]

The number 1 engine is the one that wins the TCEC tournament. Everyone agrees: This is the best engine!

I don't agree, now you may drop that "everyone"

If Chess.com comes out with their own engine, then I don't want to trust Chess.com tournaments anymore.

This doesn't make any sense. You need to elaborate.

The ranked tournaments are all well and good too but Lc0 becomes really strong with strong GPU. Unfortunately, the testers don't offer that. The super final at TCEC is played on a classic Timecontrol and super hardware (for Lc0 important).

CCC and TCEC have the exact same GPU

[jshriver]

each with its own FPGA acting as a chess coprocessor

Is this common? I have a few Xylinx I could use but even finding engines that utilize the GPU is kinda rare. Really only LC0 comes to mind, rest only use for NN training.

[jshriver]

Komodochess.com site doesn't allow log in anymore, does it?

I truly hope in memory of Don Dailey it keeps growing. I'just checked I can login to download links for previous purchases.



Noticed they did remove Komodo 8 which Mr. Dailey was kind enough to give me a free copy. Then again the older copes are still free to download anyway.

[dkappe]

each with its own FPGA acting as a chess coprocessor

Is this common? I have a few Xylinx I could use but even finding engines that utilize the GPU is kinda rare. Really only LC0 comes to mind, rest only use for NN training.

stoofvlees, a0lite, Scorpio and maybe a dozen more.

[Steve Maughan]

Can I ask, how long has Torch been in development? When was the first line of code written? It’s no small task to create the #2 engine.

— Steve