Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Length of time for Move_Gen()?

Author: Russell Reagan

Date: 14:31:21 02/15/04

Go up one level in this thread


On February 15, 2004 at 17:19:53, Sune Fischer wrote:

>I think you are right, so it is a tiny loop processing the same branch
>conditionals over and over and benefitting a lot on cache hits.

This test, just like perft, isn't really a good measure of comparison between
two programs. I think it's a good test between different versions of your own
program however. For instance, if you rewrite the movegen, you can see if it is
faster. Even still, the speedup in a tiny loop may not be directly indicative of
the speedup you'll get in your search. For instance, if all the CPU cache can be
dedicated to storing rotated bitboard lookup tables, it'll run faster. During
the normal search you can't use the entire cache for that.



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.