Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Refuting nonsense about 64 bits servers/supercomputer chips

Author: Robert Hyatt

Date: 05:32:01 09/04/01

Go up one level in this thread


On September 04, 2001 at 06:32:18, Francesco Di Tolla wrote:

>
>>Rank Manufacturer                       (GFlops) Processors
>>
>>1.   IBM ASCI White,SP Power3 375 MHz   [7226.00] [8192]
>>2.   IBM SP Power3 375 MHz 16 way       [2526.00] [2528]
>>3.   Intel ASCI Red                     [2379.00] [9632]
>>4.   IBM ASCI Blue-Pacific              [2144.00] [5808]
>>
>>11.  Cray Inc. T3E1200                  [1127.00] [1900]
>>
>>439. Presto III Athlon 1.333 GHz 	[77.40]   [78]
>>
>>Interesting is of course the AMD which gets almost 1 GFlop per processor where
>>the Cray by far does not get 1 GFlop per processor. Bottom line: isn't your
>>beloved Cray out-performed by nowadays micro processors available from your
>>local super market?
>>
>
>Well, this is a measure of the efficency of the pipline of the floating point
>operations on the CPU and of the bandwidth/efficency of the bus connecting the
>CPU.
>
>It has very little to do with chess.
>
>Of course a high bandwith is good for a parallel chess program, but the
>performance on integers in non vectorial calculcations is the number that
>matters for chess.
>
>ciao
>Franz


Those assumptions are not necessarily valid.  IE Cray Blitz used vectors
quite efficiently.  In move generation.  In attack detection.  In the
evaluation.  It only requires that vectors be included in the design
process...

If you use them, then a vector machine can scream right along on chess
problems.  We did...



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.