Author: Vincent Diepeveen
Date: 06:14:42 04/07/04
Go up one level in this thread
On April 07, 2004 at 08:27:59, martin fierz wrote: >On April 07, 2004 at 06:49:59, Renze Steenhuisen wrote: > >> >>Hi all, >> >>could someone give me some numbers that are common with hashkey collisions? >>Because I guess my % is little too high... >> >>I'm getting like 0.03% [which is 1 every 3000, if I'm not mistaken] >> >>This is when using TT=32MB (haven't got the exact number of entries) >> >>If you think it is an error, any suggestions on where to start looking? >> >>Thanks! >> >> Renze > >hi renze, > >your number is much too high. a good estimate for the hash collision probability >is 1/sqrt(hashkey_range); which in your case comes out as 2^-32 or about >somewhere around one in a billion... Not at all. It's way less than that with 64 bits Zobrist. I measured it with 460 processors at a supercomputer with a shared hashtable and had major problems to get at 7 million nodes a second stored, some errors/collissions. >make sure you're not doing anything wrong with the hashkey updating like >suggested by others - compare with a computed-from-scratch key. also make sure >that your nullmove doesn't break your hashkey. and just to be sure, check that >you haven't got 32-bit numbers instead of 64 bit numbers by accident... >cheers > martin I'm sure it was some implementation bug with Renze.
This page took 0.02 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.