Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: If nodes per sec decides playing strength

Author: Uri Blass

Date: 22:20:09 01/28/01

Go up one level in this thread


On January 29, 2001 at 01:06:11, jonas cohonas wrote:

>Goliath light would top all rating lists, but it dont!!
>Gandalf is the prog that seems to benefit most IMO from
>higher processor speed, and as we all know Gandalf does
>not have the highest node count. When i ran Gandalf on my
>old amd 500mhz k6-2 against fritz 6 and junior6 it would
>lose typically with a margain of 4-6 points in a 20 game
>match, but on my new Athlon T-bird 1.2 ghz it wins by 2-3 points
>in 40/120 games!! i think this goes to show that node speed
>is only part of a programs playing strength.

I did not claim that more nodes per second decides about playing strength.

I only claimed that increase in % in nodes per seconds decide about playing
strength.

I believe that gandalf is more optimized for the new processors than for the old
ones.

I guess that you may get the following numbers:

K6-2 500 mhz

Gandalf 50 kn/s Junior 300 kn/s

T- bird 1.2

Gandalf 200 kn/s Junior 750 kn/s

If this is the case it means that Gandalf is 4 times faster from the T- bird 1.2
when Junior is only 2.5 times faster so Gandalf earns more from the new
processor and the fact that Junior searches more nodes per second is not
relevant.

Uri



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.