Xref: utzoo comp.sys.super:377 comp.arch:23115
Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!swrinde!zaphod.mps.ohio-state.edu!menudo.uh.edu!karazm.math.uh.edu!jet
From: jet@karazm.math.uh.edu (J Eric Townsend)
Newsgroups: comp.sys.super,comp.arch
Subject: Re: Massively Parallel LINPACK on the Intel Touchstone Delta machine
Message-ID: <1991Jun6.161351.8577@menudo.uh.edu>
Date: 6 Jun 91 16:13:51 GMT
References: <1991Jun3.233741.8570@elroy.jpl.nasa.gov> <13301@pt.cs.cmu.edu> <16333@life.ai.mit.edu>
Sender: usenet@menudo.uh.edu (USENET News System)
Organization: University of Houston -- Department of Mathematics
Lines: 19
Nntp-Posting-Host: karazm.math.uh.edu

In article <16333@life.ai.mit.edu> lethin@raisin-scone.ai.mit.edu (Richard A. Lethin) writes:
>At what size N for massive linpack does the accumulated numerical fuzz
>swamp the precision available in a double precision number?

I dunno.  i860s will work with denormalized numbers (which kills
performance, IMHO).

>Aren't really large problems of interest sparse matrices anyway?  Who

Not really, talk to Boeing, for example.  I wish they'd publish on what they're
doing, but they're so damn secure that some of their employees almost
had a fit during an Intel class when it was mentioned that they
had a ipsc/860 and were running LOOCS.

--
J. Eric Townsend - jet@uh.edu - bitnet: jet@UHOU - vox: (713) 749-2126
Skate UNIX! (curb fault: skater dumped)

   --  If you're hacking PowerGloves and Amigas, drop me a line. --