TalkChess.com

shrapnel wrote:Is Komodo 8 faster than Komodo 7, kN/s -wise ? If so, by how much ?
Does it reach more Depth, faster ?
My my main gripe with komodo 7 was the very low kN/s speeds compared to H4 and even Stockfish !

I never understand this - how do you know the engines count nodes in the same manner?

What you state means that if a version of Komodo 8 was released where the nodes x 10 is shown, you would have been happy, because the kN/s would be high?

What matters is how well it plays.

shrapnel wrote:Is Komodo 8 faster than Komodo 7, kN/s -wise ? If so, by how much ?
Does it reach more Depth, faster ?
My my main gripe with komodo 7 was the very low kN/s speeds compared to H4 and even Stockfish !

This statement is very telling. It shows how little you know about computer chess. Kn/s can never be used as the only factor to judge the strength or quality of an engine. Infact often the exact opposite may be true. However, I am not saying that a faster .exe is not a great achievement. I am stating that Kn/s cannot be used to judge upgraded versions of the same engine. A larger much more intelligent .exe may have far smaller kn/s values, but it may crush the faster kn/s version in play. Capiche?

shrapnel wrote:Is Komodo 8 faster than Komodo 7, kN/s -wise ? If so, by how much ?
Does it reach more Depth, faster ?
My my main gripe with komodo 7 was the very low kN/s speeds compared to H4 and even Stockfish !

Hi,

First, never stop asking questions. That is how we learn! So people, pleas stop the attacks on someone just wanting to learn.

I just did a quick test on the opening position to a fixed depth, and the node rate for Komodo 8 is about 13% faster than Komodo 7a. Much of this is due to making the program more efficient, plus compiling it on a different compiler that boosted the speed by about 4%. Part of it is also to a slight change in how Komodo counts nodes (old Komodo did not count a node if it left the King in check, while new Komodo counts it, then takes back the illegal move). This was done to make the program faster overall for other reasons, although it might not seem intuitive. So the short answer is a minimum of 4% faster, but probably more like 9% with other speedups and code changes. And it is doing a lot more stuff now, making a direct comparison a bit hard.

Mark

Hmm... got my answer elsewhere, no thanks to anyone here !
Yes ! It is around 1000 kN/s faster than K7a in any given position and reaches More Depth faster than K7a !
Good ! I'll be using Komodo 8 more often now in my online Matches !
EDIT :-
Oh ! I didn't see your post, Mark.... thank you !

mjlef wrote:And it is doing a lot more stuff now, making a direct comparison a bit hard.

Mark

Yes, I think I know what you mean. I use Fritz 13 GUI, and I've observed that K8 sort of scrolls much faster( somewhat like Houdini) through possible move options than Komodo 7 !
I don't know how else to put it, but that's what I have observed !

Can the Komodo Android version use Syzygy endgame bases?

Hi Ted,

If the GUI supports setting the UCI options, yes. At the moment, only Chess PGN Master (https://play.google.com/store/apps/deta ... .pgnviewer) lets you do that, though, as far as I have found.

Jeremy

Komodo 8 plays strong and I am happy to have it on
My mobile and tablet.

Playing vs. stockfish5 there...

sockmonkey wrote:Hi Ted,

If the GUI supports setting the UCI options, yes. At the moment, only Chess PGN Master (https://play.google.com/store/apps/deta ... .pgnviewer) lets you do that, though, as far as I have found.

Jeremy

Thank Jeremy, Nice to know I can use the 345 piece bases. I will place them on SD card.

PS: I already own Chess PGN Master, and his other really good program ChessOCR. https://play.google.com/store/apps/deta ... h.chessocr

mclane wrote:Komodo 8 plays strong and I am happy to have it on
My mobile and tablet.

Playing vs. stockfish5 there...

+1 Just comparing analysis and testing out the new version.