Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Typical percentages spend in different parts of the algorithm

Author: Dann Corbit

Date: 07:29:59 10/22/03

Go up one level in this thread


On October 22, 2003 at 09:26:09, Renze Steenhuisen wrote:

>I was wondering about what the percentages are of different engines when looking
>at how much time is spent in some parts of the algorithm.
>
>Move generation:

Tiny fraction of the time.

>Evaluation:

Depending on the program anything from a tiny fraction to dominating

>making/unmaking:

Tiny fraction

>move sorting:

Often a surprising amount of time on move ordering (e.g. 1/3 of all time)

>table-lookups:

Hashing tends to be a very large fraction of the time for good programs.

>etc...?
>
>Another thing is whether these percentages changed in the past 10-20 years or
>so?

There is no way to make any sort of generalizations, because these things change
a lot from program to program.

The good programs tend to be balanced in this sort of format:

1. Hash operations
2. Evaluation
3. Search

4. (The rest)

Maybe 2 & 3 will be flipped.  On some programs with complicated eval, eval may
exceed the others.  On some programs with exotic search, search may dominate.



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.