Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: a test position for computer programs(perpetual check)

Author: Mark Young

Date: 15:05:23 05/30/98

Go up one level in this thread


On May 30, 1998 at 09:32:49, blass uri wrote:

>the position:q5k1/4Q3/5r2/5P1B/6PK/8/8/8 white to move
>
>Qxf6 leads to perpetual check begin with Qh1+
>with 99968 Kbytes hash tables fritz5 found the draw after 32
>seconds computing 9823 Knodes
>
>Junior4.6 found the draw after 10 minutes and 21 seconds
>and it computed 92,632 Knodes.
>
>we see that the computers do not know what lines to compute
>because I can see the draw by looking at less than 1000 nodes.
>
>fritz5 is only very weak in this position while other programs I have
>are weaker than fritz5.
>
>is there a program that can see the draw
>after less nodes than fritz5?
>
>Uri

Position tested on a P II 300

Program       Hash         Time        Nodes

Fritz 5       50mb         21"         9,984KN  Fastest time.
Rebel 8       60mb         45"         5,100KN  Fast time. Lowest node
count.
Rebel 9       60mb         48"         5,100KN  Fast time. Lowest node
count.
Genius 5      32mb         3:34"       unknown
Junior 4.6    50mb         5:23"       92,674KN
Nimzo 98      50mb         >15'        unknown
Mchess 7.1    60mb         >15'        >35,328KN



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.