TalkChess.com

Because the Neural Net in Lc0 is bigger than the one in SF NNUE, do we have any examples of where Lc0 evaluates a position better?
I'm thinking of positions where a deep AB search won't help much, such as blocked positions.

I remember Houdart saying it was fine. I also typically use an even number of cores doubling from 2, but that's more out of habit than something rational.

I wonder how a head to head would go between the 3090 and 2x2080Ti for Lc0. I guess running off one card would be quite a big advantage

Although it's not stated, surely that means FP16 on CUDA alone will be 71 TFLOPS, no? That's well over 2x performance increase for Lc0..

OK this looks BIG

https://www.anandtech.com/show/16057/nv ... 0-rtx-3090

Actually black moves first is being seen as an anti-racist idea, based on the assumption chess promotes white privilege (white moves first).

Nutty if you ask me...

Depressing.

Concerning speedup for Lc0: With Nvidia Ampere on its way pretty soon I've been looking into leaked and (take with a pinch of salt etc etc) early reports suggest actually it's nothing special. Performance per watt looks surprisingly similar to Turing, with the new cards sucking 350W each rather than...

With Lc0 a lot of the attention went on the NN and how it was better than a hand-crafted Evaluation Function. But I'm trying to understand the search side of things. I have a basic, layman's, working knowledge of alpha-beta search in chess and virtually no real understanding of MCTS, which I think ...

I wonder if 1.e4 g5 is lost for black