Author: Ulrich Tuerke
Date: 04:44:06 12/18/00
Go up one level in this thread
Hi Tim, I have been using the same position to test my transposition table for a long time. I have observed that the replacement algorithm for the table access is very important for positions of this kind. I mean the case that the key of a position maps to an index which is already in use, and you have to determine another index in order to store the position. IMO, it's very important to avoid "clustering" for these cases. I remember very well that one of my first attempts was to add some constant offset to the index. That didn't work very well because the new index, built this way, soon experienced the same problem as the old one and after a constant number of re-trys, the storage of the position was aborted. Now I use some bi-"or-ing" or "and-ing" IIRC, ,and it works much better. May be, you have a similar problem here ? Regards, Uli
This page took 0.01 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.