Author: blass uri
Date: 08:09:27 06/07/99
Go up one level in this thread
On June 06, 1999 at 19:28:03, Chuck wrote: >On June 05, 1999 at 03:55:03, eric guttenberg wrote: > >>Tania, >> >>I suspect Micheal is right. The results of the games between Fritz 5.32 >>and Hiarcs 7.32 that I have seen up till now show advantage to H7.32 >>but not by anywhere near the overwhelming margin in your results. >> >>It is likely that playing the games on one computer influences the performances >>and leads to unreliable results. I suspect that Fritz 5.32 and Hiarcs 7.32 >>are not very far apart although it appears that Hiarcs may be significantly >>stronger. Not many games have been played though. >> >>eric > >Your system setup plays into the hands of HIARCS, and I'm sure this affects >the results to some degree. Fritz relies on large hashtables, it is the way >the program is designed. A 16 MB hash table handicaps Fritz, even in blitz. I do not agree that 16 MB handicapps Fritz in blitz Fritz certainly have not enough time to fill 16 MB in blitz. I think that even in 30 minutes per game 16 MB are enough or almost enough(depends on your hardware). I do not think that there is a big difference. >For HIARCS this is not a problem, because it does not fill the hash tables >as quickly (I expect, I have not verified this yet). But one thing I have >seen is that Fritz really slows down once the hash table fills - this would >be a big problem for Fritz at very long time controls. I think that the difference is not very big. I expect the difference between 16MB hash tables and 64MB or 128MB at tournament time control for every program to be less than 40 rating points. Does someone has a proof that I am wrong for Fritz? Uri
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.