Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site geowhiz.UUCP
Path: utzoo!linus!philabs!cmcl2!seismo!uwvax!geowhiz!karsh
From: karsh@geowhiz.UUCP (Bruce Karsh)
Newsgroups: net.arch,net.math
Subject: Re: "The Shared Memory Hypercube"  Do you smell any smoke?
Message-ID: <182@geowhiz.UUCP>
Date: Sat, 4-May-85 22:29:16 EDT
Article-I.D.: geowhiz.182
Posted: Sat May  4 22:29:16 1985
Date-Received: Tue, 7-May-85 20:44:25 EDT
References: <2132@sun.uucp> <1447@think.ARPA> <551@lll-crg.ARPA> <1483@think.ARPA> <560@lll-crg.ARPA> <1523@think.ARPA>
Distribution: net
Organization: UW Madison, Geology Dept.
Lines: 20
Xref: linus net.arch:961 net.math:1615

> 
> >> Very few programs actually use the n-cube directly
> >Take a look at the FFT,  its an important problem, it blows smoke on the
> >hypercube.  As I have said before, the wires is there.  The FFT uses all of
> >them.
> I was afraid you might bring up FFT.  Just as a matter of curiosity,
> what percentag of total computing time do you use up on FFT's?

  Well, in the geophysical industry, it's not uncommon to have mainframes
doing almost nothing but ffts.  The geophysical industry is probably one
of the biggest consumers of computers in the world.
  (I'm not implying that it's necessary for a parallel machine to do FFT's
quickly.  But the oil industry would be a very large purchaser of a parallel
machine that did FFT's quickly.)
-- 
Bruce Karsh                           |
U. Wisc. Dept. Geology and Geophysics |
1215 W Dayton, Madison, WI 53706      | This space for rent.
(608) 262-1697                        |
{ihnp4,seismo}!uwvax!geowhiz!karsh    |