Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Bitboards

Author: Robert Hyatt

Date: 09:06:35 08/16/99

Go up one level in this thread


On August 16, 1999 at 09:11:06, Doug Beardsley wrote:

>I need some information about bitboards and how they relate to the rest of the
>chess program.  I have written what I call a "limited" bitboard representation
>which only uses two bitboards, white_occupied and black_occupied.  I use them in
>get_legal_moves() to speed up the generation by only checking occupied squares
>(pretty elementary).  I decided to convert it to a "full" bitboard
>representation similar to that used in crafty with bitboards for each individual
>piece.  I got that to work fine, but it caused a huge decrease in my nps (from
>~30,000 down to 2,000).  I'm pretty sure that this is because of increased
>overhead in make_move().  Now, here are my questions.  Does the full bitboard
>implementation slow things down like that, or am I doing something wrong?  Since
>the full bitboard implementation slows things down, how does the program use the
>bitboards to make the speed decrease worthwhile?  Comments would be greatly
>appreciated.


best test would be to compile crafty on your machine and compare the speed
it gets (using a full-bitboard representation) to yours...



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.