Author: Bas Hamstra
Date: 22:39:25 08/13/99
Go up one level in this thread
On August 13, 1999 at 06:26:16, Vincent Diepeveen wrote: >On August 12, 1999 at 21:51:16, John Eric Hanson wrote: > >>Hello All. New to this board - saw it mentioned in a response to a question on, >>I believe, sci.math. >> >>I noted a number of questions/comments here about theoretical issues involving >>algorithms. I was curious if most of the developers here were working from >>original code or using some open source chess program as starting point. >> >>Thanks much, -JEH > >Every byte of my program has been written by me. > >Yet i can imagine that a lot of programs share the same datastructure, >as some are simply fastest for now. Which are? I am redoing it all and am about to do some bitboards for fast attack generation. I experimented a lot. With incremental attacks, and/or incremental movegeneration. It is reasonably fast (say 10x Diep NPS or so, but with more eval). But with all that writing about movesorting, ok, I started experimenting with non-incremental better sorting. Disappointing, so far. What's the point of better sorting when the tree goes -10% and NPS -20%? I guess that's why hash is so utterly important for Fratz. The hashtable compensates a lot of the poor sorting. My target: 0.5 * Frutz with reasonable sorting.
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.