Author: Robert Hyatt
Date: 16:56:26 06/25/02
Go up one level in this thread
On June 25, 2002 at 17:33:07, Gian-Carlo Pascutto wrote: >On June 25, 2002 at 17:28:10, Russell Reagan wrote: > >>And as for the original testing if there is a pawn on E6, don't you keep a >>64-square array anyway along with your bitboards? > >IMHO That's not really bitboards any more. I do this because it is faster on 32 bit machines than hunting thru the various bitmaps to find what is on square E6, for example. It would be possible to change my move format and my move generator so that the captured piece is not included. Then this 64 word array could be discarded. And one day, for speed, that will happen. But at the moment, with all the fiddling around I do, the redundant data is very useful in that I have a debug mode that validates everything and if the bitmaps and the 64 char array don't agree, it pukes and tells me about it. > >>If so, is there anything that >>a classical approach can do more efficiently than a bitboard approach? Edge >>detection and attack detection seem to be two of 0x88's typical strong points, >>but I don't see how they would be faster than bitboards. Or is this a matter of >>the hardware you use? Ex. 32-bit hardware, it might be close...64-bit hardware, >>bitboards faster for everything...? > >Generating non-capturing moves. > >-- >GCP Maybe... I would expect that to be a break-even operation. But since captures are generated first at every ply in the tree, and often that is enough to stop the search with a fail high, or in the q-search captures are the only thing generated most of the time. So the GenerateCaptures() option is hit a lot more frequently... IE in a current profile run for Crafty, GenerateCaputures() was called 1.3M times, GenerateMoves() was called .4M times... a 3:1 ratio. This seems to hold up across several different sets of positions I tried in fact... 3:1 to 4:1 was typical...
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.