KOMODO 1405 x64 6CPU v STOCKFISH 210315 x64 6CPU (Match)

[JJJ]

Come on. I m hoping for Komodo to be better than stockfish soon and get the realease when it ll come. But still not to soon I hope.

So I don't want to hear that Komodo will loose this one

You are not going to hear, you will see it.

Btw., all tests are interesting and providing us with useful information.

For example, what is it that makes Komodo perform worse than SF in bullet, than at equal strength in blitz, then again worse at rapid, and then finally better at very long time control?

Not a linear trend, in any case.

Komodo better at very LTC is over I think.

[Lyudmil Tsvetkov]

Come on. I m hoping for Komodo to be better than stockfish soon and get the realease when it ll come. But still not to soon I hope.

So I don't want to hear that Komodo will loose this one

You are not going to hear, you will see it.

Btw., all tests are interesting and providing us with useful information.

For example, what is it that makes Komodo perform worse than SF in bullet, than at equal strength in blitz, then again worse at rapid, and then finally better at very long time control?

Not a linear trend, in any case.

Komodo better at very LTC is over I think.

Not quite sure.

The last 2 matches run by Graham were AMD matches, where for some reason, at least on multiple cores, Komodo seems to perform worse in relation to SF.

[Graham Banks]

....AMD matches, where for some reason, at least on multiple cores, Komodo seems to perform worse in relation to SF.

Not so sure that this is fact. More of a rumour or assumption, I suspect.
Anyway, might run a match on my Intel octal at some stage.

[Lyudmil Tsvetkov]

....AMD matches, where for some reason, at least on multiple cores, Komodo seems to perform worse in relation to SF.

Not so sure that this is fact. More of a rumour or assumption, I suspect.
Anyway, might run a match on my Intel octal at some stage.

Most of the tests I have seen recently seem to indicate that.

For example, recently Bertil posted on fishcooking test data relating to latest SF SMP patches. According to the data, after more than a 1000 games, on 24 core AMD machines SF performs some 30-35 elo better than Komodo in comparison to the same sample size on Intel machines.

[Modern Times]

Current Score:

Komodo leads 111.0 - 109.0 ( +33 = 156 -31)

Very close indeed !

[JJJ]

Haha ! What will you say now Lyudmil Tsvetkov ?

[Modern Times]

Stockfish leads 136.0 - 132.0 ( +42 =188 -38)

[JJJ]

What will you say now Jean-Baptiste ?

[Modern Times]

With such a narrow margin, you were setting yourself up there !

Long way to go yet, just over half way through.

[JJJ]

With such a narrow margin, you were setting yourself up there !

I know