Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Speedups for BitBoard programs on 64-bit machines

Author: Vincent Diepeveen

Date: 10:18:47 06/05/02

Go up one level in this thread


On June 04, 2002 at 17:52:47, Dann Corbit wrote:

>On June 04, 2002 at 16:28:39, Gian-Carlo Pascutto wrote:
>
>>On June 04, 2002 at 16:18:55, Gian-Carlo Pascutto wrote:
>>
>>>Because you are using a processor that is clocked at twice the clock
>>>frequency?  Why compare a 1ghz processor to a (nearly) 2ghz processor
>>>and conclude anything about efficiency there?  Is there anything that
>>>suggests that the alpha is simply more "efficient"?  To justify that
>>>clock frequency disparity?
>>>
>>>A machine twice as fast (clock freq) _should_ perform just as well as
>>>a 64 bit machine at 1/2 the frequency...  Less would suggest that the
>>>32 bit machine simply sucks badly.
>>
>>I don't agree with the validity of a clock-for-clock comparison,
>>but if you want to do it anyway, I'll again point to Vincent's
>>numbers:
>>
>>At the same clockspeed, Crafty only gets 33% faster on the 64-bits
>>machine.
>>
>>When you read this, keep in mind that most applications get _more_
>>than 33% faster on the 64-bits machine.
>
>All the new 64 bit chips in the discussion are pretty much beta stage right now.
>
>A 500 MHz 21264 hit 1M NPS for Crafty when Tim Mann did it.
>
>I think to extrapolate from instances has some value, but the general idea of a
>64 bit chip compared to 32 has so many advantages that surely they will clobber
>the 32 bit chips at some point.  There will also be a point when 10 GHz 64 bit
>chips are a cheap commodity item.  The 32 bit chips will pretty much be
>relegated to an embedded role by that time.

>So, I think that architecturally, it makes good sense to design for a 64 bit
>system right now.  It should also operate successfully on a 32 bit system.  But
>sometime in the not-too-distant future, the 64 bit chips are going to blow the
>doors off of the 32 bit chips and those who are not ready for the transition
>will be a large step behind.

That transition from 32 bits to 64 bits, for example for zobrist hashing,
that work requires only 1 week or so. So if in a few years we all have
cheap affordable 64 bits machine with a good compiler that can produce
64 bits code for it, we can put our hashing and a few other things in
64 bits if needed pretty much realtime then.

No reason to slow down datastructures a factor 2 now.

>There are several reasons that 64 bit performance data is sketchy for the new
>chips.
>
>First, they don't want performance data out yet because it gives a good target
>to the opposition.  If they release it before the opposition has had time to
>address it, by the time they do exceed the performance, the next generation will
>be ready anyway.
>
>Second, as beta chips, they still have opportunities for improvements and bug
>fixes.  No need for any embarassing 'forgot to download the division table
>completely' bugs.



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.