Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Test suite for move generation

Author: Uri Blass

Date: 15:38:45 03/03/04

Go up one level in this thread


On March 03, 2004 at 17:09:53, Andrew Wagner wrote:

>On March 03, 2004 at 16:54:45, Dann Corbit wrote:
>
>>On March 03, 2004 at 16:45:21, Andrew Wagner wrote:
>>
>>>Hmm, you guys are fast, but I'm not sure either of you actually completely read
>>>my post :). The two keys are
>>>
>>>1.) "mass-produce" (i.e. not copy and paste a fen and type 'perft 1', 'perft 2',
>>>etc.)
>>
>>Seems it should not be hard to have a program spit out the intermediate results.
>>So if you said:
>>Perft 5
>>It would also print out perft for 1,2,3,4
>>
>
>Yes, I *know* how to do that in my own program. I'm asking what program will
>read in fens or epd's from a text file and print out the perft values through 5
>or so.
>
>
>
>>>and 2.) "my own positions" - the ones on peter's site are great, but if your
>>>program is off in one of those positions it's very hard to track down a bug from
>>>it.
>>
>>On the other hand, the right answers are provided.
>>
>But...while it's useful to know the 'right answer' that my program sees 2 extra
>position in 5 million, it's a lot more useful to know if my program sees 2 extra
>positions out of 35.

You can find the relevant case when it sees extra positions out of 35.

If perft 5 is wrong then it is clear that perft 4 is wrong in one of the
positions that are one ply later so you can check the relevant positions one ply
later and find one when perft 4 is wrong and continue in this way.

I remember that sharper gives you also all the numbers of perft 4 when it
calculates perft 5 so you can find the number that is wrong faster and do not
need to calculate perft 4 for many positions only to find the wrong one.

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.