Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: overflow problem crafty in perft

Author: Uri Blass

Date: 12:07:02 12/31/02

Go up one level in this thread


On December 31, 2002 at 14:04:03, Uri Blass wrote:

>On December 31, 2002 at 13:53:13, Vincent Diepeveen wrote:
>
>>On December 31, 2002 at 12:54:36, Uri Blass wrote:
>>
>>it has nothing to do with generating at all of course.
>>if i use my incremental attacktable it will beat it of course.
>
>I know that it is possible to do it faster but no program that I know does it.
>
>Some points that are possible to improve in movei.
>
>1)Movei does not use assembler stuff.
>2)Movei use no profile optimization except optimize for speed.
>3)Movei use no special hash tables to calculate perft faster(of course hash
>tables cannot be used in a similiar way to chess but they can be used).
>
>Uri

I can add that special code for white and for black in the move generator can
also help movei.
Movei of today does not have it(I added special code for cases when there are no
pins but not for white and black).

There are other small improvement that may be done so I believe than a good
programmer can change movei to be twice faster even without hash tables.

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.