Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Rough comparison between rotated bitboards and 0x88

Author: Tom Kerrigan

Date: 10:06:34 06/20/00

Go up one level in this thread


On June 20, 2000 at 03:30:03, Inmann Werner wrote:

>Hello
>
>some rough ideas..
>
>What speaks against using both systems (bitboards and 0x88) in one engine.
>At make_move and undo_move, you have surely a speed decrease.
>But everywhere else, you can "mix" Bitboard and 0x88 code, whichever is faster
>or more efficient. Also in evaluation, it sometimes help, to have a "normal"
>boardrepresentation too.
>In my new version, I want to use 4 movegenerators.
>a) promising captures   b) rest of the moves c) if I am in check  d) quiet
>search
>
>Second idea: What speaks against Bitboards like
>
>BITBOARD BITSWHITE[6],BITSBLACK[6];
>
>where 1 to 6 are the material (BITSWHITE[5] are the white queens)
>you can "overdefine" the Bitboards, so you can access the wanted bitboard
>directly.
>
>I hope you understand my ugly english..
>
>Greetings
>
>Werner

Mixing bitboards and 0x88 is a bummer because they are associated with different
board dimensions.
-Tom



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.