Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Differences between 0x88 ,10x12 and Bitboards!?

Author: Robert Hyatt

Date: 08:04:27 11/20/02

Go up one level in this thread


On November 20, 2002 at 02:11:03, Anthony Cozzie wrote:

>On November 19, 2002 at 16:48:55, Robert Hyatt wrote:
>
>>On November 19, 2002 at 16:36:23, Gian-Carlo Pascutto wrote:
>>
>>>On November 19, 2002 at 16:34:46, Russell Reagan wrote:
>>>
>>>>On November 19, 2002 at 16:18:01, Gian-Carlo Pascutto wrote:
>>>>
>>>>>I don't think there's any way to prove that they are.
>>>>>
>>>>>I don't think it's a good idea to spew nonsense arguments
>>>>>in a fruitless attempt to convince someone who has another
>>>>>opinion.
>>>>
>>>>So in other words, you have just been a troll today.
>>>
>>>No, I've been specifically attacking people making unjustified
>>>claims they can never support with sound argumentation.
>>>
>>>--
>>>GCP
>>
>>
>>I found one bit of data.
>>
>>Bruce compiled ferret on a pentium pro 200, and then on a 500mhz alpha.  He ran
>>2.5X faster.
>>I compiled Crafty on the same pentium pro 200 using the same compiler as bruce
>>(MSVC)
>>and then on the alpha using the same compiler again and I ran 3.5x faster.
>>
>>How would you explain that difference?
>>
>>Everything was the same for the two programs.  Same compiler.  Same processor.
>>Same
>>everything.  Except that Bruce used his 0x88 or whatever at the time, while I
>>was using a
>>bitmap.  This was in 1997, pretty early.  I ran on that machine for the 1997
>>WMCCC
>>event.  Bruce used one of the cryo-machines at 767mhz that year...
>>
>>but 2.5X vs 3.5X has to come from somewhere.  Luck?  Or what?  I'm pretty sure I
>>_know_
>>the answer.  That is a 40% improvement if I did my math right.  Which is in line
>>with what
>>most knowledgable people claim a bitmap program should do since not all
>>instructions need
>>64 bits of data (branches, loop counters, a few others).
>
>/skepticism on
>remember, there are many many many differences between the Alpha and the P6, not
>just the fact that alpha is a 64 bit architechture.  Alpha is a true RISC,
>different caches, page sizes, etc.
>/skepticism off


OK... but both programs are compiled on the same machine using the same
compiler, so
both programs get exposed to the same architecture.  Hence it is the
_architecture_ that is
making the most significant contribution to speed gains...

That was my point all along...

A 64-bit program is _obviously_ going to gain more from moving to such a machine
than
a 32 bit program will gain.  And I do mean _obviously_...



This page took 0.02 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.