Author: Uri Blass
Date: 12:33:23 11/14/02
Go up one level in this thread
On November 14, 2002 at 14:44:30, José Carlos wrote: >On November 14, 2002 at 13:51:59, Uri Blass wrote: > >>On November 14, 2002 at 13:33:26, Russell Reagan wrote: >> >>>On November 14, 2002 at 11:04:42, José Carlos wrote: >>> >>> >>>> It's a completely new engine written from scratch, just for the fun of >>>>experiencing new things. It uses non-rotated bitboards and MTD(f). Actually, my >>>>personal versions of both :) >>> >>>How do you handle things like rook file attacks with your non-rotated bitboards? >>>In other words, what do you do instead of using rotated bitboards? Do you use >>>the Kogge-Stone algorithms? >>> >>>Russell >> >>I cannot speak for averno but I do not have special evaluation for rook file >>attacks and there is not big difference in strength between movei and averno. > > He means for move generation. Rook attacks gives you the set of squares where >a rook can move to. In this case I usually do not use them. I have attack tables but they do not give me the set of squares that the rook can move to. They give me only for every square the direction that it is attacked and for every relevant direction the squares that can move to it(I say square because the attack tables give me information about squares and not about the piece that is the attacker(I can calculate the piece directly from another array)) Uri
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.