Author: Alvaro Jose Povoa Cardoso
Date: 08:42:05 02/01/02
In my checkers (portuguese/spanish version with long jumps) program I use the four bitboard ray technique for the four directions for each of the 32 square for the king jumps (with 32bit bitboards). And then I use the BSF/BSR assembler instructions for finding the moves. Since the PC processor speed is evolving much faster than the memory/FSB speed I would like ask if there is a significant performance diference between this technique and the rotated bitboards with 'large' tables for chess. I did not started yet my chess engine, but I want to start with bitboards so I would like your opinions if this technique is fast enough for the sliding pieces. It seams to me that there is less data needed for these operations and consequentely all this will be on L1 cache most of the time. I also understand this technique better than rotated bitboards. Best regards, Alvaro Cardoso
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.