Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Debugging a chess program

Author: Anthony Cozzie

Date: 05:46:10 12/21/03

Go up one level in this thread


On December 21, 2003 at 07:24:37, Mridul Muralidharan wrote:

>On December 20, 2003 at 21:20:11, Anthony Cozzie wrote:
>
>>On December 20, 2003 at 18:53:45, Russell Reagan wrote:
>>
>>>I have a question for chess programmers. I'd like to know some of the methods
>>>that you use to find bugs in your chess programs. A few that come to mind are:
>>>
>>>Perft, to debug movegen/make/undo
>>>
>>>Mirroring/rotating/etc. the board to test evaluation
>>>
>>>Creating an independent version of your search that will run in parallel in a
>>>debug build, and verifying that the positions match after each move that is made
>>>or undone. For instance, the independent version could be a completely different
>>>board representation, 0x88 and bitboards (or whatever).
>>>
>>>Any other methods that you use to find bugs?
>>
>>zappa has a "debug log" option where all information is recorded to file, and a
>>graphical program to display _exactly_ what the program did: alpha, beta, m
>>oves, move ordering, transposition stuff, etc.
>>
>>anthony
>
>That is neat !
>My trace search will just spew out huge junk onto the console that I will then
>have to manually analyse and find errors with.
>This is a nice idea though !
>Do you use error codes or something in case you want to find out what the errors
>were ?
>
>Thanks
>Mridul

its not really for finding technical errors (like bugs in make_move or
something).  Its more for helpnig with move ordering / finding transposition
table fooleries.

anthony



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.