Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: MTD(f), Quiscient Search, and hashing Quiscient nodes

Author: Walter Faxon

Date: 11:11:34 01/20/03

Go up one level in this thread


On January 19, 2003 at 09:08:51, Joel wrote:

>Hello All,
>
>Just finished writing a simple Quiscient Search function, using MVV/LVA...
>
>Now, previously my engine was using MTD(f) to power it. This obviously relies
>heavily on the transposition table for good performance.
>
>So this brings me to the question of hashing Quiscient nodes. I would like to
>know what the generally accepted thing to do is?
>
>Btw, I am really envious of everyone who has an engine thats able to compete in
>CCT5. Hopefully mine will be in an acceptable state next time round. :)
>
>Thanks,
>Joel


Joel, all.

A suggestion:  have a very small second hash table devoted to hashing quiescence
nodes exclusively.  Simple replacement only.  Advantages:  (1) little memory
latency since this hash table can be mostly cached, and (2) no interference with
entries in the main hash table.  I don't have a program, so take this idea with
a grain of salt.  If you try it, please post your results.

Thanks.

-- Walter



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.