Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: perft records

Author: Reinhard Scharnagl

Date: 08:13:47 09/06/04

Go up one level in this thread


On September 06, 2004 at 10:55:24, Peter Fendrich wrote:

>On September 06, 2004 at 09:32:01, Reinhard Scharnagl wrote:
>
>- snip -
>>I do not know what you are thinking of. You can see that I am also are counting
>>mates, which have to be detected first. But indeed, I have a move-generator
>>which produces fully informed moves (capture, e.p., check, double check,
>>mate ...). So using that hardly calculated information for to optimize perft is
>>no cheating. I think, that using a pseudolegal move generator is a trick
>>instead.
>
>No, no I'm not implying that you are cheating but I think it is possible to
>optimise the code for perft only and that is not cheating. I can think of some
>techniques that would be very useful for perft but not useful for move
>generation in search. For instance just by skipping the sort (if one have a
>sort) in MoveGen would make perft way faster.
>Anyway it's hard to compare different approaches but I would like to get some
>figures despite that so thank you for your input!
>
>BTW. I'm using a pseudolegal move generator but have to go one ply deeper in
>order to capture the errors. The final perft counters are still correct so I
>don't think that's cheating either.
>/Peter

Peter,

this simply has been an reaction. I think it is very hard to compare different
generators. I invest a lot of efforts in the fully informed moves, you have to
inspect sub-plys later - there is no big difference.

But as long you do not update similar statistics, comparing would be odd. May
be I have to switch off the detecting of mates to have comparable results.

Regards, Reinhard.




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.