Lion wrote:Will there still be a game 5 and 6 ?
If yes, what will be the conditions of these 2 games and when will they be played ?
Thank you again for organizing that.
Rgds
Asking me they will be played!
Same conditions, different opening. When? I don't know yet. It might be some time as all participant have to have some time left again. (3 parties to coordinate ...)
Bye
Ingo
Is the result 3.5-.5 to Houdini? if so, i would personally believe that that Rybka Cluster is no match for that Houdini, and that the test is sufficient and therefore there is no need to spend more money for the cluster rental.
(But maybe you think it's negligable compared to other expenses? I don't know).
It seems to me that that cluster did not add anything to Rybka 4.1, but dropped from it. Perhaps the strongest cluster would also not be stronger than Houdini.
But i might be wrong. There are bigger experts than me. I HAVE been wrong in my life!
Houdini is completely outsearching and outplaying Rybka cluster. It is probably going at least 2 or 3 ply deeper for the same amount of time. Rybka cluster might have been the strongest chess entity when Rybka was #1 engine, but I think with 64 cores used in a non shared memory system ... it probably would be playing stronger had it used a fast 16 core shared memory single motherboard setup. It all depends on how the 64 cores were setup. If all the master and slave motherboards were 16 cores then the disadvantage would be much less, but if the 64 cores are made up of several 8 core machines, then it has no chance against a 16 core H3. I think for Rybka 4.1 to be competitive with H3, it would need 2 orders of magnitude superior hardware ... which would be a 64 core system on one motherboard and using shared memory. I don't think such a system exists. Of course another way to tighten the game is to upgrade the actual code of Rybka 4.1 and make its search and evaluation stronger. It doesn't look like much work has been done to improve Rybka code wise. Every single move played by Rybka Cluster was also chosen by my 8 core setup using 4.1
The slow squeezing technique I was talking about....
Watch closely now....
Dr.D
Dr., my good friend- all this is good fun, but what do you know now that you didn't know before the first game was played? Remember all the interest here is not over Houdini vs. Rybka, but the cluster vs. God knows how many cores Houdini was running. I contend it would not be a lot different if you just took the 2 engines and gave them 6 or eight cores each. Or just 1 core each. When you strip away all the glitter and the bands playing music, you still have Houdini 3 vs. a version of Rybka on its way to being 3 years old. No doubt- as I have said over and over, Vas and Robert are on a different planet from the rest of the chess programmers in this world. Houdini 3 is definitely Number 1- and it may end up being that Houdini is still Number 1 six months after Rybka 5 is released. Personally I would not put my money on that- but I will get excited when I see Houdini vs. a newly released Rybka.
One e5 xeon+4x12 westmere(1x5690+2x5680+1x5670) according to lukas earlier post in forum for the composition of 64 cores on the cloud. Not sure if that is still the configuration.
I dont think it has anything to do with this configuration as sjeng beats h3 on similar xeons vs h3 on 16 core e5 2687w.
It is the search that requires Vas attention then followed by evals. Hardware shared memory is pure lame excuses.
64 core machines do exisit. Amd 6200 will give that or e5-8800 will give you 80 cores but they are used as database servers not so good for chess due to lower clock speeds and crazy pricing.
Suj wrote:One e5 xeon+4x12 westmere(1x5690+2x5680+1x5670) according to lukas earlier post in forum for the composition of 64 cores on the cloud. Not sure if that is still the configuration.
The slow squeezing technique I was talking about....
Watch closely now....
Dr.D
Dr., my good friend- all this is good fun, but what do you know now that you didn't know before the first game was played? Remember all the interest here is not over Houdini vs. Rybka, but the cluster vs. God knows how many cores Houdini was running. I contend it would not be a lot different if you just took the 2 engines and gave them 6 or eight cores each. Or just 1 core each. When you strip away all the glitter and the bands playing music, you still have Houdini 3 vs. a version of Rybka on its way to being 3 years old. No doubt- as I have said over and over, Vas and Robert are on a different planet from the rest of the chess programmers in this world. Houdini 3 is definitely Number 1- and it may end up being that Houdini is still Number 1 six months after Rybka 5 is released. Personally I would not put my money on that- but I will get excited when I see Houdini vs. a newly released Rybka.
Merry Christmas-
george
Merry Christmas my friend
_No one can hit as hard as life.But it ain’t about how hard you can hit.It’s about how hard you can get hit and keep moving forward.How much you can take and keep moving forward….
