Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: The hashtable formula for Hiarcs 7.32

Author: Melvin S. Schwartz

Date: 15:01:12 06/13/99

Go up one level in this thread



On June 13, 1999 at 14:09:24, Eran wrote:

>Is the formula below correct?
>
>HT[KB] = 0.2 * Pfreq[MHz] * t[s]
>
>HT is the hashtable size in kilobytes, Pfreq is the clock speed of the processor
>in MHz, and t the time available per move in seconds.
>
>Example: Average time per move of 30 minutes (1800 seconds) on a Pentium 200
>will require:
>
>HT = 0.2 * 200 * 1800 = 72000 KB = 72 MB
>
>Or
>
>Example: Tournament games with an average of three minutes (180 seconds) per
>move on a Pentium 200 will require:
>
>HT = 0.2 * 200 * 180 = 7200 KB = 7.2 MB

Hello Eran,

That formula you quoted above is the formula for Fritz 5.32. The CD containing
info about Hiarcs actually states Fritz, and I believe the manual is flawed. It
isn't clear to me why they begin the manual with Fritz at the headng. I find it
very difficult to believe Hiarcs uses the same formula for HT requirements as
Fritz. I requested information on this matter from Chessbase and am awaiting a
response to this and other questions.

Mel>
>However, Hiarcs 7.32 does retain information in the hashtable memory even
>between moves.  So it sounds to me that the information is built up until full
>in the hashtable memory while Hiarcs 7.32 is playing thru several moves ahead.
>Is that correct?  Is Hiarcs 7.32 very smart with hashtable memory?
>
>Eran



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.