Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Decrease in Crafty kNPS

Author: Gabor Szots

Date: 02:42:34 10/14/05

Go up one level in this thread


On October 14, 2005 at 04:05:36, Dann Corbit wrote:

>On October 14, 2005 at 02:23:25, Gabor Szots wrote:
>
>>On October 13, 2005 at 17:34:42, Dann Corbit wrote:
>>
>>>On October 13, 2005 at 17:16:38, Gabor Szots wrote:
>>>
>>>>20.0 was around 1125 on my Athlon XP 2500+. 20.1 is around 800. Why?
>>>>
>>>>Gábor
>>>
>>>I see a speed drop, but nothing so dramatic.  I expect that there is an
>>>important difference in the build process used to create the two binaries you
>>>are using.
>>>=====================================
>>>Crafty v20.0
>>>
>>>White(1): bench
>>>Running benchmark. . .
>>>......
>>>Total nodes: 77856394
>>>Raw nodes per second: 1179642
>>>Total elapsed time: 66
>>>SMP time-to-ply measurement: 9.696970
>>>White(1): quit
>>>=====================================
>>>Crafty v20.1
>>>
>>>White(1): bench
>>>Running benchmark. . .
>>>......
>>>Total nodes: 71777655
>>>Raw nodes per second: 1010952
>>>Total elapsed time: 71
>>>SMP time-to-ply measurement: 9.014085
>>>White(1): quit
>>>=====================================
>>
>>20.0:
>>
>>Bryan Hofmann: 1151
>>Peter Skinner: 1101
>>
>>20.1:
>>
>>Hofmann: 836
>>Skinner: 795
>
>What do you get with this one:
>http://cap.connx.com/chess-engines/new-approach/Crafty-1921.exe.bz2

found computer opening book file [./bookc.bin].
hash table memory =   96M bytes.
pawn hash table memory =   16M bytes.
EGTB cache memory =   16M bytes.
draw score set to    0.00 pawns.
choose from book moves randomly (using weights.)
choose from 5 best moves.
book learning enabled
result learning enabled
position learning enabled
playing a computer!
resign after 6 consecutive moves with score < -7.
5 piece tablebase files found
13980kb of RAM used for TB indices and decompression tables

Crafty v20.1

White(1): Running benchmark. . .
......
Total nodes: 74031875
Raw nodes per second: 732988
Total elapsed time: 101
SMP time-to-ply measurement: 6.336634
White(1):



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.