Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Perft -- Test position and data

Author: James Swafford

Date: 06:41:52 02/23/06

Go up one level in this thread


On February 23, 2006 at 09:10:47, Charles Roberson wrote:

>
>   Since I see others (Stuart) working on perft, I'll release a position that
>  is helping me find a bug.
>
>     7k/8/8/8/1p6/8/PP6/6K w - - 0 0
>
>    Here are the results
>      Depth    Hand Calculation    Telepath
>         1            6                 6
>         2           31                31
>         3          319              1288
>
>     Either my hand calculations are wrong or Telepath has a bug in the
>   EP code. The numbers represent the sum of all the nodes in the tree for that
>   depth. Thus, the 6 root nodes are counted in the 2 ply and 3 ply counts.
>
>    Given that some do not count the nodes that way subtraction should do the
>   trick with the depth counts being 6, 25 and 287 respectively.
>
>     I'd like to see somebody post perft results on this position for as many
>    depths as possible assuming they think their engine produces correct
>results.

Here is what I get with Prophet:

Initializing position...done.
setboard 7k/8/8/8/1p6/8/PP6/6K w - - 0 0

- - - - - - - k
- - - - - - - -
- - - - - - - -      white to move
- - - - - - - -      no ep
- p - - - - - -      no castling rights
- - - - - - - -
P P - - - - - -
- - - - - - K R

perft 2
running perft on movegen: bitmap movegen

29 moves (0 ms)
perft 3
running perft on movegen: bitmap movegen

225 moves (0 ms)
perft 4
running perft on movegen: bitmap movegen

1385 moves (1 ms)
perft 5
running perft on movegen: bitmap movegen

11494 moves (6 ms)
perft 6
running perft on movegen: bitmap movegen

77643 moves (56 ms)
perft 7
running perft on movegen: bitmap movegen

655713 moves (330 ms)
perft 8
running perft on movegen: bitmap movegen

639669 moves (331 ms)

--
James



This page took 0.01 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.