Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Generating attack tables: Is this code good?

Author: Anthony Cozzie

Date: 06:38:18 12/01/03

Go up one level in this thread


On December 01, 2003 at 08:30:21, Tord Romstad wrote:

>On December 01, 2003 at 07:42:52, Anthony Cozzie wrote:
>>
>>I experimented with some incrementally updated attack tables - I gave them up
>>because they were too slow.
>
>I had a conversation with Mridul on the ICC some time ago where this topic
>came up, and he told me he had the same experience.  Incremental updating
>was slower.
>
>>They took about 1000 ns to update per node on my
>>Athlon XP 1.5 [although this depends greatly on cache misses].  If you want,
>>I'll send you the code.
>
>Thanks, it would be interesting to see!  I'll let you know if I find any
>possible improvements, of course.
>
>Tord

Well Zappa is currently a bitboard based engine, and I was considering
switching.  Zappa's nps (in a lazy-eval position like WAC141) went from 550 ->
350 using these tables.  The big problem I had was that the bitboard code will
be considerably faster once I get an Opteron, while the attack table code will
stay slow.

anthony



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.