Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Hash replacement scheme

Author: Robert Hyatt

Date: 11:42:52 10/12/01

Go up one level in this thread


On October 12, 2001 at 14:39:30, Bob Green wrote:

>Dr. Hyatt or Mr. Moreland, Mr. Pascutto, Mr. Corbit, et. al.,
>
>Could you please comment on the relative advantages and disadvantages of:
>
>  1. Highest number node count
>  2. Deepest ply level
>
>when doing hash replacement?
>
>I would appreciate your insights.


They ought to be equivalent.  IE a deeper search produces a larger node
count.  But there are exceptions.  One position might lead to a part of
the tree where lots of extensions occur due to tactics, while another
position leads to a simple tree with few tactics and few nodes.

Which is the most important?  Technically, the tree with the larger number
of nodes in it represents more work to do, and by saving that entry, you
might avoid doing that work later.  Practically, I am not sure there is
an advantage to _either_.  One will be faster in some positions, the other
will be better in other positions.

I played with this a year or so ago when someone suggested it and I didn't find
any difference worth mentioning.  I like storing the "draft" (remaining depth)
as it requires fewer bits, which makes the hash entry smaller.  But other than
that, I suspect it is six of one, a half-dozen of the other...



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.