Finally great speed-up to Stockfish?

[Jouni]

https://nextchessmove.com/dev-builds

HQ attacks for AVX2:

Result of 100 runs
base (./stockfish.master ) = 1136025 +/- 2816
test (./stockfish.patch2 ) = 1171370 +/- 2848
diff = +35346 +/- 3275

speedup = +0.0311
P(speedup > 0) = 1.0000

[Eelco de Groot]

Speed pretty much meaningless, by your own posts Jouni, if doubling CPU time is now almost immeasurable, I think those were Mark Young's tests in ProDeo? It is not even pretty much, but totally unmeasurable by Elo because of noise. New Stockfish; Issues with race conditions still remaining if that is what Vondele is writing about (in the issues), and the Tablebase Mating issues probably. To just name two issues. But a lot of comments to a new net that (the comments) are absolutely incomprehensible these days. Complete alchemy If they hired some AI to write an advert that sounds impressive but nobody understands, they could not have done a better job. Would not be surprised if they actually did let an AI write it. Anyway, sounds good and certainly a gainer at fast timecontrols, that new net just two hours ago. I just saw it five minutes ago, I did not actually try to read any of the comments in question so please ignore my complaints totally...

[Jouni]

Patch gave free +4 elo in NCM and framework test.