Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: programmers: pawn hash tables

Author: Uri Blass

Date: 06:24:32 03/09/03

Go up one level in this thread


On March 09, 2003 at 08:50:12, Joel wrote:

>Hey Uri,
>
>If I can do it in 25 minutes, it won't take _anyone_ 100 hours! :)
>
>All I hash is a score, and a lock. The score is based on the stuff which is
>evaluated _only_ on pawn structure. Backward, isolated, doubled etc.
>
>At the moment I am re-writing the code to allow for the passed pawns and open
>file information to be stored there as well. (Thanks Bob)
>
>Its given me already more than a 10% speed increase. You have done all the hard
>work with getting the transposition table working, why not simply spend a night
>for that free 10%?
>
>Joel

It probably can give me more than 10% after I improve the speed of other
parts of my program but my thoughts are that it is better first to improve the
ealuation of pawn structure because I do not like to do the
same work twice and I may need later to change the hash structure.

I do not plan to evaluate backward pawns in the near future
and it is enough for me to improve evaluation of isolated pawns and passed
pawns.

I think that evaluating too much at a time is not a good idea and I think that
backward pawns when they are a weakness may be often evaluated by mobility.

Uri



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.