infinex wrote:
So Rybka wins.
Yes, Rybka wins. As of game 99, Rybka is in the lead by 2 points. Game 100 is still being played. I should know the results in the morning.
Deep Rybka 4.1 -- 100-game Match
Sort:
Where there any dramatic changes to the side who was winning (on the profile graph) during the 100 games or did one side maintain a slight advantage in the wins ?
ivoryknight71 wrote:
chessmaster102 wrote:
that is a strong possibility cause it just so happens the space on a PC can affect it's performance.
please help me understand this. i am *not* saying you are wrong. i just don't fully understand how this is possible, nor what you mean by "space". thanks.
as far as i know, engines only care about a few things: crunching power (speed of each processor and number of cores), time to crunch, and enough RAM for caching the variations. if you're talking about hard drive "space", then i'm curious to know more.
I think they both use page files if available, also, i think the fairest test would be to network 3 pc's together with the programs running on separate computers (to avoid interferance) with the 3rd computer running the UI.
Vacuous wrote:
Where there any dramatic changes to the side who was winning (on the profile graph) during the 100 games or did one side maintain a slight advantage in the wins ?
There were dramatic surprises a couple times. One engine, I think it was Rybka, totally didn't see it coming at least once. I don't remember the game numbers, however. I think I blogged about them on another site though. I may go back and check that tomorrow.
Vacuous wrote:
ivoryknight71 wrote:
chessmaster102 wrote:
that is a strong possibility cause it just so happens the space on a PC can affect it's performance.
please help me understand this. i am *not* saying you are wrong. i just don't fully understand how this is possible, nor what you mean by "space". thanks.
as far as i know, engines only care about a few things: crunching power (speed of each processor and number of cores), time to crunch, and enough RAM for caching the variations. if you're talking about hard drive "space", then i'm curious to know more.
I think they both use page files if available, also, i think the fairest test would be to network 3 pc's together with the programs running on separate computers (to avoid interferance) with the 3rd computer running the UI.
pagefiles are always available. they're also incredibly slow. hopefully your system has enough RAM so your pagefile *never* needs to be used during your competition, except in a rare case for the O/S.
ivoryknight71 wrote:
chessmaster102 wrote:
that is a strong possibility cause it just so happens the space on a PC can affect it's performance.
please help me understand this. i am *not* saying you are wrong. i just don't fully understand how this is possible, nor what you mean by "space". thanks.
as far as i know, engines only care about a few things: crunching power (speed of each processor and number of cores), time to crunch, and enough RAM for caching the variations. if you're talking about hard drive "space", then i'm curious to know more.
yes in rybka help file everythiing should be explained but its a lot of detail but to summarize it helps the engine absorb more from games and the hash tables are more efficient.
ivoryknight71 wrote:
Vacuous wrote:
ivoryknight71 wrote:
chessmaster102 wrote:
that is a strong possibility cause it just so happens the space on a PC can affect it's performance.
please help me understand this. i am *not* saying you are wrong. i just don't fully understand how this is possible, nor what you mean by "space". thanks.
as far as i know, engines only care about a few things: crunching power (speed of each processor and number of cores), time to crunch, and enough RAM for caching the variations. if you're talking about hard drive "space", then i'm curious to know more.
I think they both use page files if available, also, i think the fairest test would be to network 3 pc's together with the programs running on separate computers (to avoid interferance) with the 3rd computer running the UI.
pagefiles are always available. they're also incredibly slow. hopefully your system has enough RAM so your pagefile *never* needs to be used during your competition, except in a rare case for the O/S.
I've got about 4 gig of RAM, so I guess I'd have to setup 1 gig hashtables for each program to run a tournament.
I like how you use different openings, but 1 opening is played as white and black.
SerbianChessStar wrote:
I like how you use different openings, but 1 opening is played as white and black.
i did that in games 51-100...
[COMMENT DELETED]
ivoryknight71 wrote:
For the final 50 games, each engine will play the White side and the Black side of the same opening. In rounds 51 and 52, Houdini won as White and Black in a Queen's Indian.
http://www.westportchessclub.org/computer-chess/rybka-vs-houdini-100games.htm
This is enough evidence that Houdini is better; he can win with the same opening with both black AND white!
[COMMENT DELETED]
Even Rybka has been accused of allegedly being a clone of Fruit.
the link doesnt work can you give another
use Google
i did instead the more official 40 game match appeared so I have no complaints.
So Rybka wins.