Author: Tomas Casanovas Martinez
Date: 12:46:26 06/14/99
Go up one level in this thread
On June 14, 1999 at 13:31:56, Dave Gomboc wrote: >On June 14, 1999 at 12:45:49, James T. Walker wrote: > >>On June 14, 1999 at 00:15:00, Dave Gomboc wrote: >> >>>On June 13, 1999 at 23:44:04, Melvin S. Schwartz wrote: >>> >>>> >>>>On June 13, 1999 at 18:53:37, Tomas Casanovas Martinez wrote: >>>> >>>>>...sorry, Mel. I forgot to say that both programmes play with 32 M. hash tables. >>>> >>>>Tomas, >>>> >>>>If you read the Fritz manual, you should know the formula for HT size. 32 megs >>>>for Fritz is less than half what he needs on your processor. Fritz needs more >>>>than 72 megs on your hardware if you're testing at 40/2? I forgot to look back >>>>but if you're testing at 40/2, then that's another problem besides using the >>>>wrong book. >>>> >>>>Mel >>> >>>Let's keep things straight here: it's a problem with Fritz, not his computer. >>> >>>For instance: Fritz is playing the GMs on a quad machine with 4 gigs of ram. >>>I'm sure that in 10 years we'll all think that 4 gigs is peanuts, but today the >>>story is different. Meanwhile, those of us in reality with a limited paycheque >>>will notice that other chess software doesn't ask for a similar amount of >>>memory, and send the complaints to where they belong. >>> >>>Dave >> >>Sorry Dave but I don't understand your complaint! The fact that Fritz requires >>more RAM than most other programs is not a downside. It's the way Fritz was >>designed and is not a problem. If you can't give Fritz the RAM it needs to >>perform up to design then it's your fault not Fritz's. You can use that as an > >The amount of RAM it needs is unreasonably large. If it is not able to cope >well with less RAM, then it is a bad implementation. If it is able to cope well >with less RAM, then comments about "you should have given it 150 megs!" are >ridiculous. Your choice. > >>excuse when Fritz doesn't perform but it's not Fritz's fault it doesn't have the >>RAM it needs. When you purchase Fritz you understand what it needs to perform >>at it's best. So you either give it what it takes or settle for less than > >Really? How does the general user (who is not a computer chess freak) know >about this before they buy? > >>optimum performance from an excellent program. I run Fritz on a K6-2-400 with >>256 Meg RAM and it performs very well against Junior/Crafty. I'm still waiting >>for Hiarcs (It's in the mail) but hope to do my own test when it arrives. I >>admit that even with 256 meg ram, Fritz sometimes hits %100 of hashtable usage >>at 40/2. That's my fault! > >So what? Hitting 100% of hashtable use is normal in computer chess. A program >should be able to play well even with a saturated hash table. If it doesn't, >then its position replacement scheme needs improvement. > >>Jim Walker > >Dave Thanks to everybody for your comments. Although I don't believe that the opening book criteria is decisive for one or another program to win, as it is proven by the fact that the vast majority of games are not decided until moves 30 to 40(!), I have changed my 'environment' giving to Fritz 5.32 its own book. I have given also 128 Meg RAM to each program. I have done the same in the match H 7.32 - Junior 5, with the maximum hash allowed by my second machine -in this case I have loaded for Hiarcs 7.32 its own book. Unfortunately I have no idea about how to load the book of Crafty 16.6 . Anyway it is performing fantastically with the Fritz book!. I will let you know the results after and before the change!. Regards, :-) Tomas.
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.