Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: EGTB question

Author: Robert Hyatt

Date: 07:08:26 01/28/01

Go up one level in this thread


On January 28, 2001 at 09:43:53, Dieter Buerssner wrote:

>On January 27, 2001 at 21:25:01, Robert Hyatt wrote:
>
>>On January 27, 2001 at 15:57:20, Les Fernandez wrote:
>>
>>>I have downloaded 117 out of 220 files taken from the 5 piece egtb from Bob's
>>>site.  I placed these 117 files into my crafty tb folder (also contains the full
>>>3 and 4 egtb set).  I am currently running crafty vs crafty with egtb=0.  All
>>>was running fine until the new incomplete 5 piece egtb was added.  The error I
>>>am getting is:
>>>
>>>Error: First chess program (crafty1.exe) exited unexpectedly
>>>Erro: Writing to first chess program. The pipe has been ended.
>
>It is possible, that you get a better error message, when you just start
>crafty1.exe without GUI and look at the screen, what it says.
>
>>>Does crafty check to make sure that the entire set of a particular egtb is
>>>present even if I have set egtb=0?? or does the error commnets suggest that of
>>>the 5 piece egtb's I have that one or more are corrupted???
>
>I believe, that it is possible to do a fast check of the TBs with the
>datacomp utility available from Robert Hyatt's ftp. It is included in
>the tbexe.zip. I have forgotten the path and the detailed instructions
>for the check, but remember, that it was not too difficult. Also, comparing
>the file sizes with the sizes in RobertHyatt's TB directory might be useful.
>
>>If you have a _partial_ TB, that is killing you.  Don't put partial files
>>in with the others.  Crafty will notice this and quit right after it starts.
>>Xboard won't be happy about it either...
>
>Perhaps there is a misunderstanding. I believe Les Fernandez asked about
>and incomplete set of 5-men TBs, and not about an incomplete single file.
>(Because he says, that he has 117 out of 220 files).
>
>But now to my question: What exactly is checked at the initialzation step
>in Eugene Nalimov's code? Is there some sanity check for the correctness
>of the found TBs done? If not, wouldn't a program only get problems, when it
>first accesses aa corrupt TB? If yes, what is checked for at the initialization
>step? I have the source, but nevertheless, the answer to those question is not
>obvious to me.
>
>Regards,
>Dieter


There really isn't a need to run datacomp.  He said that he had a _partial_
TB in the directory (one that he was downloading at the time).  That is
definitely going to kill Crafty.  It used to be that Crafty would just use
a partially-finished file and if it probed it, it might get bogus values.
THe 'sanity-test' was added to catch partial and mis-downloaded files (ie if
you use ASCII mode you will wreck the files.)

I think his problem was explained by the partial download comment he
made.



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.