Author: Uri Blass
Date: 22:10:26 01/10/03
Go up one level in this thread
On January 10, 2003 at 18:48:07, Russell Reagan wrote: >On January 10, 2003 at 17:22:59, Uri Blass wrote: > >>I guess that you are right and I guess that I will not use assembler after the >>advices that I got. >> >>The point is that it is even possible that assembler may do me slower and even >>if today it can do me faster tommorow it may do me slower because of some >>compiler optimization so it may be better to have 100% C code and not 99.9% C >>code. > >I think it would be very wise to use the method Tony gave, especially for >detecting protected passed pawns. In the vast majority of cases, there will be >no protected passed pawns, so Tony's code will exit almost immediately. In the >very few cases where there is a protected passed pawn, there will usually only >be one, and at most 8. That loop would be _fast_ always. The bitboard will >always be very sparse, so Tony's method should be more than fast enough. > >Plus the compiler is probably going to take care of any optimization, and you >can keep your program in C so you don't have to worry about changing your >assembler code if you want to run Movei on a newer processor in the future. > >Russell I already understood it but I remember previous discussion about the subject. You already posted similiar code that was copied from crafty but Dieter posted another code and based on my memory the claim was that it may be faster for programs that has not real 64 bit numbers. 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.