Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Anybody believe me ?

Author: Jeremiah Penery

Date: 06:34:49 06/05/99

Go up one level in this thread


On June 05, 1999 at 04:32:04, blass uri wrote:

>
>On June 05, 1999 at 02:19:42, Micheal Cummings wrote:
>
>>
>>On June 05, 1999 at 01:59:54, Tania Devora wrote:
>>
>>>
>>>
>>>A poor result for fritz5.32.   why ?
>>
>>I assume you played it on one computer. I do not know enough about Fritz and
>>Hiarcs if you play them on one computer maybe one will be worse off.
>>
>>I know CM6K loves to eat all CPU power, so there would be a disadvantage there.
>
>The difference is that Hiarcs7.32 is an engine for Fritz(both program are sold
>by chessbase so chessbase should care that in engine-engine games it will not
>happen.
>
>I do not know about the case of Hiarcs7.32 but from my experience from
>engine-engine matches[Junior against crafty or Fritz5(16 bit)] there is no
>problem of eating CPU.
>
>It is possible that the fact that the sides could not use the permanent brain
>gave advantage fo Hiarcs7.32 because hiarcs7.32 does not clear the hash tables
>after every move and the fact that it remember information from previous search
>may be more significant without permanent brain.
>
>It is also possible that the Hiarcs7.32 was lucky.
>
>In computer games sometimes there are results of 8:2 between equal opponents
>because one player was unlucky to choose openings that it does not understand.
>
>Uri

Another thing is that Fritz needs big hashtables to perform better.  Since it
seems to have a bad hashing strategy, anything less than 64M isn't good for it.
I don't think Hiarcs has the same problem.

Jeremiah



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.