[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Crays vs <?>



-----BEGIN PGP SIGNED MESSAGE-----


writes Timothy C. May:
>
>On a more serious note, strong crypto will allow "foreigners" to send
>computing jobs to sites nearly anywhere in the world and have them run
>on compute servers. So much for export controls on the physical
>hardware!
>

I tend to disagree strongly with this statement.  The physical export of
supercomputers will still be controlled, although it may be relaxed.  I think
that supercomputers have been technically raised to 2000 MFLOPS, but that's
another story.  Getting time on crays, CMs, Paragons, KSRs, nCubes, whatever
is not extremely difficult, but it's pretty improbable that you'll get time
on the new Cray EL at the Pittsburg Supercomputing Center if you're a 
scientist living in North Korea (for example).  I suppose it's possible that
some very rich person (new, fully configured Cray T3D computers are $75M!)
you could set up a Center that took digicash for CPU time... not really
probable considering the upkeep on a Cray.

Anyway, I don't see the unrestricted use of true supercomputers in the 
near future... very powerful desktop/deskside machines are another
story.  (afterall, if you buy up a bunck of 2GFLOP machines (not considered
supercomputers for export) and string them together adequately, that's a hell
of a machine!)

- -nate

- -- 
+-----------------------------------------------------------------------+
| Nate Sammons  [email protected]  (303) 491-1578                  |
|   Colorado State University -- Computer Visualization Laboratory      |
+-----------------------------------------------------------------------+