Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Fritzmark for 8.0.0.8 is higher between Fritz 8.0.0.8 and Fritz 8.0.0.23

Author: William Penn

Date: 00:22:39 12/06/03

Go up one level in this thread


I suspect that the Fritzmark (or Shreddermark, or whatever) is a series of known
chess problems, and the resulting number is a measure of how long it takes to
solve them as compared with an internal standard of some kind.

I have Shredder 7, not Fritz 8, and will give the suggestion of odd/even hash
table values a try. I doubt it will make a difference, but you never know. I did
a lot of experimenting trying to determine proper hash table sizes and concluded
that it's impossible to do. So I just set them for a reasonably big number,
while leaving enough free RAM for the op system to be comfortable.
WP

On December 04, 2003 at 13:19:59, Ernest Bonnem wrote:

>I think one should only compare the Kns.
>The FritzMark itself is just a rabbit pulled out of the Chessbase hat.
>Chessbase has never given any explanation of what it actually measures.
>
>On December 04, 2003 at 09:30:01, Brian Katz wrote:
>
>>Has anyone noticed that the Fritzmark when using Fritz 8.0.0.8 is higher by
>>approx. 110 to 140 Kns then when using Fritz 8.0.0.23?
>>
>>However, when just analysing positions, the Kns seems to be about the same.
>>
>>Also, what I found that was interesting......Was when setting the Hash with an
>>odd number, rather that by 2's or an even number as recommended, the value was
>>higher in many cases.
>>I am running an AMD Athlon XP 2600+ with 1 gig DDR RAM
>>Such as a reading for Fritz 8.0.0.8:
>> 59 MB Hash> 1339
>> 60 MB Hash> 1184 as opposed to 61 MB Hash> 1339
>> 66 MB Hash> 1318...............67 MB Hash> 1339
>>128 MB Hash> 1276..............129 MB Hash> 1297
>>Now if you try making comparisons by using Friz 8.0.0.23 you may find that you
>>get a reading of 110 to 140 Kns lower than with 8.0.0.8.
>> Has anyone found similar results?
>> If no one has tried making comparisons between odd vs even settings, give it a
>>shot, you may find it interesting. You may also find that there is a big
>>difference in the reading between engine versions.
>>Brian



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.