Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Crafty and 4Q tablebases

Author: Peter Marko

Date: 12:27:03 09/23/99

Go up one level in this thread


Thanks very much, Eugene. Does this mean that the KQQKQQ tablebase is ready to
use with Crafty 16.19? Looking at the size (over 400MB), I have no way of
downloading this to my home machine (33Kbit/s connection = 30 hours best case).
I may try it here at work, we have four times the throughput even during the
day, so I might be able to do it overnight. Would it be very inconvenient for
you to send the 4Q TB on a CD? I am willing to cover your costs, of course. By
the way, what are your thoughts on distributing your TB on a CD to a few people
on the World Team?

Thanks,

Peter


On September 23, 1999 at 14:54:23, Eugene Nalimov wrote:

>I uploaded Win32 executables to Bob's FTP site. Glitch in probing code is fixed,
>as well as one more EGTB-related bug that appeared in 16.19.
>
>Eugene
>
>On September 23, 1999 at 14:39:07, Robert Hyatt wrote:
>
>>On September 23, 1999 at 12:50:21, Peter Marko wrote:
>>
>>>Thanks for your quick response, Robert. Regarding CD distribution of the 4Q
>>>tablebases, I am more than willing to do it provided that
>>>
>>>(1) All parties involved give me permission to do it; and
>>>(2) I receive one copy of the tablebases on CDs.
>>>
>>>I can burn the duplicate CDs and mail them to a few people on the World Team to
>>>take that burden off the developers' shoulders.
>>>
>>>Please keep me posted on your progress with integrating Crafty and the 6-piece
>>>TBs.
>>>
>>>Thanks,
>>>
>>>Peter
>>>
>>
>>Crafty is finished...  (16.19).  I think Eugene might have a minor glitch in
>>the probe code, and he is looking at it, as we only need wtm for the kqqkqq
>>type files as they are symmetric with white and black, and it is just a matter
>>of inverting the board and probing to get the right score.  He thought he was
>>doing this, but I was only seeing hits on every other ply, which suggested that
>>it might not be working right.
>>
>>in any case, download 16.19 source, add the -DEGTB6 compile option, and you
>>are ready to go, although maybe not quite optimally just yet...  and I would
>>generally not run with just a couple of 6 man files in real games, as it will
>>try to probe on _any_ 6 man ending and most will be rejected, wasting time.
>>
>>
>>



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.