Author: Gian-Carlo Pascutto
Date: 10:36:46 05/25/02
Go up one level in this thread
On May 25, 2002 at 13:18:12, Russell Reagan wrote: >I suppose it depends on how much "some shifting/anding" is, because you can >calculate attacks and mobility using "some adding/testing" (a.k.a ray tracing). >It seems to me like doing a little shifting and anding isn't clearly superior to >doing some adding and if testing. So, am I missing something? Perhaps it is a >slight bit faster and the couple of cpu cycles you save will eventually add up, >but even with 64-bit computing, this doesn't seem like it would give you this >giant boost that people talk about. Especially on a 32-bit machine, that >shifting and anding that you talk about is twice as expensive since it has to do >it twice for each 32-bit segment. So while it sounds nice when you hear someone >describe bitboards, is it really that much better? I still don't get it. 1) shifting/anding is completely predictable, the if's are not (big difference!) 2) the 2 x 32 bits operations are fast on superscalar machines (all modern CPUs) 3) I do not believe bitboards are necessarily faster on 32 bit machines -- GCP
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.