Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: 6 piece tablebase question

Author: GuyHaworth

Date: 04:49:27 03/08/04

Go up one level in this thread



Last I heard, Eugene had 22 4-2p DTM EGTs to do, plus all (65) 3-3p ones.

Don't know what his production rate is, but at 1 per week, that's ...


Also don't know Eugene's attitude to having the generator source in the public
domain:  presumably, as it is not, he's now not so in favour of that as he was.

The code is not fully UNIX-compatible, but John Tamplin (on UNIX) and Marc
Bourzutschky are sorting the occasional issue there.

There is no 'community' of people generating EGTs, though I guess there could be
if necessary.  Maybe we should have a list of volunteers with mega-kit
(particularly RAM).

The current EGTs being generated raise issues around:

1)  database-files > 2GB, now sorted by the multi-part format
2)  virtual stores > 2GB, another 32-bit integer problem
3)  index-ranges > 2 GB, ditto
4)  concurrency of many parallel tasks on highly-parallel (SMP) machines
5)  UNIX portability, as the '16-bit code' hasn't been so 'run in' as 8-bit


So, EGT-generation is not as easy as it looks, but all the above represent
progress.

g


PS:  Thanks to Eugene's help and Marc Bourzutschky's generalisation of Eugene's
code, I am completing the generation of 6-man EGTs for 'Mini Chess', 6x6 board,
which historically was the domain for the first well documented computer chess
game.

... and some endgames are deeper than in 8x8 chess!








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.