Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Tablebase files and something sneaky

Author: Eugene Nalimov

Date: 14:05:02 09/24/99

Go up one level in this thread


egtb.cpp from Crafty 16.19 that is basically tbindex.cpp contains similar fix.
It also contains a fix in the probe function itself.

Eugene

On September 24, 1999 at 15:53:23, Peter Karrer wrote:

>There's a little problem with your tbindex.cpp (from tbgen_new.zip) code in this
>regard.
>
>FRegistered(tb, x_colorBlack) will fail for symmetric tables e.g. KQQKQQ.
>
>My workaround was
>
>if (rgtbdDesc[tb].m_fSymmetric && side == x_colorBlack) {
>        psqW = blackPieces;
>        psqB = whitePieces;
>        side = x_colorWhite;
>}
>
>On September 24, 1999 at 12:24:23, Eugene Nalimov wrote:
>
>>Pawnless files withe equal material ("symmetric") are absolutely identical. I
>>ignored that fact when the only such files were KNKN, KBKB, KRKR, and KQKQ, as
>>total saving were ~6Mb, and I decided not to complicate the program. But with 6
>>man TBs saving is much more, so I modified the code. Now for pawnless symmetric
>>TB only WTM file is necessary.
>>
>>For TB with pawns situation is more complicated: it's necessary to "flip" the
>>board for BTM. I had not done it yet, and I think that we have more than enough
>>time before that issue will become actual.
>>
>>Eugene
>>
>>On September 24, 1999 at 11:23:43, Dave Gomboc wrote:
>>
>>>On September 24, 1999 at 02:48:21, Dann Corbit wrote:
>>>
>>>>On September 24, 1999 at 02:43:55, Dave Gomboc wrote:
>>>>
>>>>>On September 24, 1999 at 02:39:34, Dann Corbit wrote:
>>>>>
>>>>>>And another thing:
>>>>>>I notice that we have only:
>>>>>>ftp://ftp.cis.uab.edu/pub/hyatt/TB/kqqkqq.nbw.emd
>>>>>>and
>>>>>>ftp://ftp.cis.uab.edu/pub/hyatt/TB/krrkrr.nbw.emd
>>>>>>
>>>>>>Is there a plan to create the .nbb files, or is this strictly for the Kasparov
>>>>>>match?
>>>>>
>>>>>Well, the data would be exactly the same, so why would you want to waste the
>>>>>space?  (Both sides have the same material.)
>>>>Does not seem to make any difference.  For some reason (unknown to me) it does
>>>>not seem to be symmetrical.  For example:
>>>>ftp://ftp.cis.uab.edu/pub/hyatt/TB/kbkb.nbb.emd
>>>>ftp://ftp.cis.uab.edu/pub/hyatt/TB/kbkb.nbw.emd
>>>>or:
>>>>ftp://ftp.cis.uab.edu/pub/hyatt/TB/kpkp.nbb.emd
>>>>ftp://ftp.cis.uab.edu/pub/hyatt/TB/kpkp.nbw.emd
>>>>
>>>>Strangly (to me at least because I don't understand why) they are not always
>>>>even the same size.
>>>>
>>>>
>>>>Same material, but two different files.
>>>
>>>In that case, I probably misunderstood the composition of the files.  Sorry! :)
>>>
>>>Dave



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.