Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: significant math

Author: Robert Hyatt

Date: 13:53:48 11/19/02

Go up one level in this thread


On November 19, 2002 at 16:51:35, Gian-Carlo Pascutto wrote:

>On November 19, 2002 at 16:42:21, Robert Hyatt wrote:
>
>>in other words, no evidence is acceptable?
>
>I can't think of much that would do, certainly
>nothing that has been produced here. There are
>several ways I could be convinced without a real
>proof, even.
>
>If you, say, would rewrite Crafty in classical form,
>spend time optimizing it, invite Vincent to do the
>same and he ends up with something that's decidedly
>slower than what you have now, then that's pretty good
>evidence that for a program like Crafty bitboards are
>the superior approach.
>
>If the top 5 engine programmers speak out and all say
>they use or switch to bitboards I'm also going to be
>convinced it's the superior approach.
>
>If after the switch to 64-bit hardware you end up
>smashing me and Vincent by a significant margin then
>I'm also going to be convinced.
>
>Lots of ways to convince me, as long as they're not
>based on hand-waving (you like that word, don't you?)
>
>--
>GCP


Yep and I don't do it.  I've given you _real_ numbers.  Of course you can
_always_
run the test that Bruce and I did for yourself.  Crafty's source is available,
so you have
access to a bitmapper.  Yours isn't a bitmapper so you have access to one that
is not.

Compile both using the same compiler on a 32 bit machine and on a 64 bit machine
and
see if one speeds up _more_ than the other.  If so, you have to explain why that
isn't
attributed to the 64 bit architecture...



This page took 0.03 seconds to execute

Last modified: Thu, 07 Jul 11 08:48:38 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.