Author: Robert Hyatt
Date: 10:57:44 12/05/01
Go up one level in this thread
On December 05, 2001 at 12:45:54, Gerd Isenberg wrote: >On December 05, 2001 at 10:30:38, Robert Hyatt wrote: > >>On December 05, 2001 at 08:22:40, Gerd Isenberg wrote: >> >>>About the pawn hash key discussions: Why using Zobrist keys for pawn hashing at >>>all, at least in BitBoard programs? Isn't it smarter to use a unique 46 Bit >>>(Rank2-7) Difference of two colored Pawn-BitBoards instead of zobrist keys for >>>pawns? No key collisions and key and index (key mod nEntries) calculation on the >>>fly. >>> >>>Gerd >> >>The problem is that you have 56 bits of information. 48 possible pawn >>positions + 8 possible en passant target files. Which N bits of that are >>you going to use to produce the table index? _that_ is the problem. >>Because for hashing, _every_ pawn needs to influence the signature and >>table address. Using a real position, some pawns won't influence >>anything unless you take the position, fold it in half, and so forth, and >>then you discover that invites collisions like mad. > >Isn't it possible to choose a special prime number for nEntries to ensure that >every pawn or most combinations of pawns change the remainder in a pseudo random >way? I don't see how when you are going to squash a 56 bit field into something like 16-20 bits. Most people also _hate_ integer division. But in any case, for good distributions where a single pawn moved perturbs the hash signature signficantly, nothing beats Zobrist.
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.