TalkChess.com

jdart wrote: ↑

Tue Jun 04, 2019 10:20 pm

A profiler might also be helpful. You could try OProfile for Linux (http://oprofile.sourceforge.net/news/).

oprofile? Is this 1999?

perf is really what you want these days; it has basically all the mindshare. If you want something more GUI-ish and commercial, there's stuff like VTune.

I tried prefetch for TT in Stockfish at some point, and couldn't measure a gain. As you say, the prefetch is really only effective if you can issue the prefetch some time before the actual access.

It is a joke you guys. If it helps, read the last page: "The author would also like to thank the anonymous referees of the Special Interest Group on Bafflingly Overdone Ventures In Chess journal for their careful reviews." It's a SIGBOVIK paper; SIGBOVIK is a joke conference. But Tom7 frequently su...

Here is what Stockfish says about the value of chessmen (note that pawns are not 100) during the midgame and during the endgame: PawnValueMg = 136, PawnValueEg = 208, KnightValueMg = 782, KnightValueEg = 865, BishopValueMg = 830, BishopValueEg = 918, RookValueMg = 1289, RookValueEg = 1378, QueenVal...

The RTX 2000 series has specialized hardware for neural networks. These “tensor cores” is the reason why your new GPU blows the old one out of the water—the number of CUDA cores simply doesn't matter, as they're not the ones doing the work.

I think it depends on your definition of “better”. I'm fairly certain Stockfish on a fast computer would have a positive endgame score versus Carlsen. I'm not convinced it's as good as him in swindling drawn endgames against other GMs. A computer doesn't really understand the concept of putting the ...

State is fine but it should be encapsulated in an object. Like so: rg=random_generator(seed) r=rg.rand() The state maintained by the rg object could be much more complicated than a single number (this would likely be the case for a cryptographically secure random generator). The rg object completel...

Rebel wrote: ↑

Sun Jan 06, 2019 4:09 pm

Today I had a new one, took me hours to see it.

Why didn't your compiler warn you?

It depends on what calculations you are doing. If you want to speed up big matrix multiplications, FFTs or other numerical code (and you don't want to push it to the GPU), MKL can help you. But an alpha-beta chess engine does none of that. NN engines do big matrix multiplications and convoltions, bu...

I don't seem to be able to find a list of MD5 hashes for 7 piece tables. I'd like to verify my downloads are not corrupt. Could someone point me to a list of hashes? Thanks Mike No MD5, but something better: http://tablebase.sesse.net/syzygy/7-DTZ/SHA256SUMS http://tablebase.sesse.net/syzygy/7-DTZ/...