From: Bill Stewart <stewarts@ix.netcom.com>
Date: Fri, 19 Sep 1997 13:04:19 +0800
To: cypherpunks@Algebra.COM
Subject: Re: Supercomputer export link in National Security Committee report
In-Reply-To: <199709181634.JAA26324@ohio.chromatic.com>
Message-ID: <3.0.3.32.19970918215030.006c8990@popd.ix.netcom.com>
MIME-Version: 1.0
Content-Type: text/plain



>>Are we getting screwed by a false link to supercomputers?

Supercomputers have been part of the Export Controls for a long time.
Of course, a few years ago, "supercomputer == Cray-1" was the obvious
definition.  
Eugene Brooks's "Attack of the Killer Micros" has hit that industry hard.*
Some of the DEC Alpha machines hit that speed point, and lots of
lobbying was started that gradually moved the limit.  But now,
"Cray-1 ~~ Pentium-133" for basic number-crunching, though there 
are still applications requiring really high memory bandwidth where
the Cray-1 will outperform the Pentium, and the Crays have a much better 
I/O architecture than your basic PC (though PCIbus is no slouch, and
Ultra-Wide-Hyper-SCSI stuff keeps improving as well.)

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
* "Careful with that VAX, Eugene!".  The "Attack of the Killer Micros"
was Brooks's restatement of Moore's Law for the supercomputer industry,
probably 10 years ago when I was still reading comp.arch.

				Thanks!
					Bill
Bill Stewart, stewarts@ix.netcom.com
Regular Key PGP Fingerprint D454 E202 CBC8 40BF  3C85 B884 0ABE 4639