Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Dr Hyatt : hash collisions and short keys

Author: martin fierz

Date: 17:06:23 08/20/02

Go up one level in this thread


On August 20, 2002 at 16:05:17, David Hanley wrote:

>I read with interest the experiments with crafty with introducing hash false
>hits artifically into the search tree.  I recall the experiments showed that
>adding the false hits, even at a very high rate 1 per 1000, had little of no
>effect on the search results.
>
>It would be good for me to use 32-bit signatures in my program, and i'm
>wondering if the above result indicates that 32-bit signatures won't matter so
>much in my program, which is a slow searcher.  Certinaly if i check hamming
>distance, 32 bit hash false matches should be at a far lower rate than the one
>per thousand that didn't adversely affect crafty.
>
>dave

just to make sure that the terminology is right: i use 64-bit hash signatures in
my checkers program, but only store the high 32 bit in the hashtable for
verification. i use the remaining 32 bits to get an index (or better said, about
21 bits of the remaining 32, doing a & with (1<<22)-1 or so).
i've tried using a mere 10 bits instead of 32 to store. works nearly as well.
with 20 bits, i see nothing different than with 32 in 80 positions i tested.

i'm sure that storing only 32 bits is perfectly ok. but if you want to actually
use a 32-bit signature, and need 21 bits for the index, that leaves you over
with 11 bits for verification. and that is definitely too little.

hamming distances have nothing to do with this. this is some weird folklore, i
have no idea where it comes from. i've seen this discussed here a number of
times, believe me, it is nonsense :-)

aloha
  martin



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.