Computer Chess Club Archives


Search

Terms

Messages

Subject: Wu/Beal retrograde algorithm: is it faster than others?

Author: Guido

Date: 13:09:01 06/06/03


On the site  http://www.msri.org/publications/books/Book42/files/wu.pdf there is
the paper "A memory efficient retrograde algorithm and its application to
chinese chess endgames" from R.Wu and D.F.Beal.

In this paper there is a logical scheme of this algorithm that the authors say
valid also for chess endgames, obviously after modification of the specific
functions of the game.

I found just this method by myself some time ago, but I thought that it was not
so fast and abandoned it. If the authors say that it is fast, this could be due
IMHO to a very good implementation of the program.

I appreciate very much the linearity of the logical scheme and also the idea of
using bitmap and sequential reading of the disk files in order to reduce RAM
memory needs.

Does anybody know how fast is this method compared with usual forward
algorithms, as Nalimov's tbgen for example, that obtain the same results, i.e
the complete kwnoledge of any position?

It is possible, and where, to find the generation CPU/IO time of the TBs spent
by different programs (with obviously the PC used)?
In a message on CCC these times have been published for 6 men endings of the FEG
program but information obtained seem to me different from Nalimov TBs, so
comparisons could be misleading.

Guido



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.