Author: Mathieu Pagé
Date: 16:08:05 06/20/04
Go up one level in this thread
On June 20, 2004 at 18:17:54, Dieter Buerssner wrote: >On June 20, 2004 at 17:29:28, Mathieu Pagé wrote: > >>You could also use any value for the size and take the remainder of the division >>as a index, but you will have to problem in this case: >> >>1) for some size you will get a non-uniform distribution of you entries > >Why? I see no reason, that any odd size should give a worse distribution than a >power of 2 size (at least when the original Zorbrist hash key is >> than the >range of the indices in the hash table). Hi Dieter, You are right. I was tinking about the case where the range of the indices is just a little smaller than the range of the keys (example: the key range from 0 to 99 and the indices from 0 to 66, then in this case the first half of the HT is targeted by more entries than the second half), but this is obviously not the case in computer chess. I should have read my post before sending it. Thanks for your comment. Mathieu P. >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.