Author: Jorge Pichard
Date: 08:32:33 08/30/04
Go up one level in this thread
On August 30, 2004 at 09:41:26, Kurt Utzinger wrote: >On August 30, 2004 at 09:18:47, Jorge Pichard wrote: > >>On August 30, 2004 at 09:03:38, Albert Silver wrote: >> >>>On August 30, 2004 at 08:30:34, Kurt Utzinger wrote: >>> >>>>On August 30, 2004 at 08:12:52, Jouni Uski wrote: >>>> >>>>>Eine FPGA-Karte untersucht momentan ca. 3 Millionen Positionen/Sekunde. 16 >>>>>Karten machen daher theoretisch 48 MPos/sec. (Donninger) >>>>> >>>>>Jouni >>>> >>>> If Hydra made 48 Mpos/sec this again proves (in comparison >>>> with the 2 Mpos/sec on Quad-Opteron server with 4 CPU's of >>>> Shredder) that the number of pos/sec can't be taken as a >>>> reliable value for the goodness of a chess program. It's >>>> of course simply impossible to compare apples and organes. >>>> Kurt [http://www.utzingerk.com] >>> >>>It would be necessary to have more games played to really judge the relative >>>strengths of the machines (software + hardware), but supposing the gain is 40 >>>Elo per ply at this point (diminishing returns) and that Hydra is gaining one >>>ply per 3-fold speedup that would mean it is roughly 120-150 Elo stronger than >>>Shredder on that hardware. >> >>The biggest problem that we have is that we don't really know if Hydra on a >>comparable hardware is the same strenght as Shredder, therefore, to say that >>Hyrdra is 120-150 Elo stronger than Shredder would be the wrong assumption. >>Now if you take the latest Nimzo and consider that it had four upgrades in >>private and each upgrade equal to 40 Elo which is an increase of 160 Elo, than >>we can say that Nimzo (Hydra )strength is almost equal to Shredder 8 on equal >>hardware. Even if Hydra is a special program that run on parallel, I believe >>that Dr. Donninger started from his own algorithm (clone Nimzo program) and >>simply made several changes. > > Hi Jorge > You might be wrong here. I guess that Hydra is a completely > new written program and not a simple "Nimzo-clone". > Kurt Even if the program has to be re-written to make it able to accept distributed clusters, the main idea behind Nimzo with a lot of modifications could still be used by Dr.Donninger. Jorge
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.