Author: Michel Langeveld
Date: 04:18:05 12/31/01
Go up one level in this thread
On December 31, 2001 at 07:07:58, Uri Blass wrote: >On December 31, 2001 at 06:43:33, Michel Langeveld wrote: > >>On December 31, 2001 at 05:28:54, Dann Corbit wrote: >> >>>On December 31, 2001 at 05:26:11, Michel Langeveld wrote: >>>>I want to check how efficient Nullmover is. >>> >>>By how often you fail high on your first choice. >> >>time=35.04 cpu=97% mat=0 n=14164653 fh=91% nps=404k >>ext-> chk=587453 cap=73868 pp=5119 1rep=28823 mate=293 >>predicted=0 nodes=14164653 evals=4235981 >>endgame tablebase-> probes done=0 successful=0 >>hashing-> trans/ref=28% pawn=98% used=21% >>SMP-> split=0 stop=0 data=0/64 cpu=33.99 elap=35.04 >> >>1) Crafty needs 2 million nodes more. >> >>2) I build with optimized settings now and my program needs 56.51 seconds for >>this position (with all statistal data switched on). The time crafty needs is >>about 40% less than my time. >> >>3) The number of nodes crafty needs is almost twice the number of nps I get. >>I have 216515,87 nps at the moment. >> >>4) I have no killers yet. >> >>Michel > > >There is a problem with comparing nodes with Crafty if you do not use the same >extensions and the same R for null move as Crafty. > >Uri Crafty uses 2 kinds of extensions I don't have. Pawn push and recaptures. Crafty uses some kind of adaptive nullmove with R=3. I use R=2 and even an other implementation of Crafty. I think comparing nodes can be still valid but crafty sees probably more in the same kind of depth at the moment by the extensions. Michel
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.