Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: hash entry replacement schemes

Author: martin fierz

Date: 17:14:45 09/16/02

Go up one level in this thread


On September 16, 2002 at 14:12:44, Dieter Buerssner wrote:

>On September 15, 2002 at 20:45:57, martin fierz wrote:
>
>>why do you have to make sure that the index is a multiple of N when you do N
>>probes?
>
>The idea it the following. The 1-N indeces are not handled all the same (so, I
>vision this like an N-table approach). Assume, they are somewhat sorted by
>importance, and you allways replace the least important, while others are
>supposed to be more priveledged, and should stay there longer. If the index is
>not a multiple of N, the "allways replace" entry of another index will overwrite
>the priveledged entries.

i don't order my entries in any way, i just get N indices and replace the one
with the lowest value. i can try adding that modulo statement to see if it makes
any difference, i'd be surprised if it does, but you never know :-)

aloha
  martin


>Regards,
>Dieter



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.