Author: Jesper Antonsson
Date: 02:38:01 06/01/02
Go up one level in this thread
Ok. It strikes me that this is something DB could've done in hardware at no performance cost. Perhaps they even did. On May 30, 2002 at 20:23:38, Robert Hyatt wrote: >On May 30, 2002 at 16:50:37, Jesper Antonsson wrote: > >>By the way, would it be really hard to write eval that finds such locked >>positions and evaluates them as draws? I get such positions every once in while, >>especially in my minds "search tree", so it should be worthwile to evaluate them >>correctly. With care, one should be able to make such eval somewhat lazy. >> >>/Jesper > > >It isn't hard to do. The problem is that it is a fairly "slow" operation >to verify that no pawn can move at all. If it were just the case where all >pawns are "rammed" it is trivial in bitmaps. shift all white pawns forward >one rank, OR them with all black pawns, and if the POPCNT() result is the >same as the number of white pawns, they are all blocked. But there are >special cases such as pawns at a4, b3 c4 (white) and a5, b6, c5 for black. >Factoring in the really useful cases makes it more time-consuming.
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.