Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Hiarcs Hash-Table Retention?

Author: Robert Hyatt

Date: 12:50:23 06/24/99

Go up one level in this thread


On June 24, 1999 at 13:18:03, Melvin S. Schwartz wrote:

>
>I have been playing games at 40/2 against Hiarcs 7.32 optimized for the best
>settings. I have noticed that on numerous occasions even when I played the move
>expected by Hiarcs, the program took around 10 minutes to respond. I find it
>difficult to understand why it should have taken Hiarcs that long to respond to
>a move it expected.
>
>How do we really know that Hiarcs 7.32 retains the hash-tables between moves? Do
>you see the hash-table info on the screen? Does it say that in the manual on the
>CD? The manual on the CD is the manual for Fritz 5.32. Are you certain this is
>in fact Hiarcs 7.32?
>+


basically hash table size has nothing to do with how long it takes to do a
search.  That is more of a time allocation issue than anything else.  IE it
might correctly predict, but by doing so get to a deeper iteration, and it
might try hard to finish that iteration before timing out...

>
>As someone well-respected here often says: "Show me proof".
>
>Furthermore, all this talk about tablebases makes we wonder. I placed the CD
>under my breakfast table and it only supported one leg! I would need 3 more
>Hiarcs CD's to have full capability to support my breakfast table. Therefore, I
>conclude that Hiarcs tablebases are not only incomplete, but also not very
>practical.
>
>Any comments to the above will be given the utmost consideration. Please - only
>well-informed sources should reply. I will respect all opinions which disagree
>with mine even though they are wrong.
>
>Regards,
>Mel


:)



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.