Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Nodes/sec between programs.

Author: Thomas Mayer

Date: 05:17:32 09/20/03

Go up one level in this thread


Hi Teerapong,

>Suppose the nodes/sec(NPS) of program A is X and NPS of B is 3x.
>If both have the same type of move generation(). Can I infer that A
>has at least 4-5 times more complicated evaluation()?

simple answer: You can't. E.g. there are many different ways to count nodes,
some count each deepening (with nullmove, without nullmove), some count only
each evaluation, some count here, some count there... Then some have legal move
gens which might produce different (slower) node counts... So NPS does not mean
anything at all. You can only use it for benchmarking a PC or something like
that.
It is only an interesting measurement for programmers if they try to speed up
things - but usually time to depth or time to solution is used for that.

Greets, Thomas



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.