Path: utzoo!attcan!uunet!cs.utexas.edu!swrinde!emory!hubcap!shaffer
From: shaffer@crd.GE.COM (Phillip L. Shaffer)
Newsgroups: comp.parallel
Subject: Limits of sequential computation
Message-ID: <9490@hubcap.clemson.edu>
Date: 27 Jun 90 19:41:27 GMT
Sender: fpst@hubcap.clemson.edu
Lines: 31
Approved: parallel@hubcap.clemson.edu

I have heard discussions on the theoretical limits to the speed
of sequential computation, as one justification for research on
parallel processing (not that I think there are no other
reasons).  Such limits might be based on minimum size of
computing elements and the speed of light.  In a quick look, I
couldn't find any references on this topic.

Can anyone provide references or original thoughts on this
subject?

On a related topic, I saw a chart of computational speed (MIPS or
some other inadequate measure, but let's not quibble) versus
historical time (log scale for computational speed, linear for
time).  The data points represented various real microprocessors,
but the chart was far from complete.  A straight line was drawn
through the points, and was not too bad a fit.  I didn't see any
curvature in the data, but the author had drawn an extrapolation
(beyond 1990) as leveling off.  Has anyone seen or done a serious
study of trends of computing speed versus time, and if so, does
any evidence of leveling off exist, or is this a supposition
based only on the above mentioned theoretical limits to
computational speed?

Thanks for any responses.  If anyone e-mails to me, I will post a
summary.

--
     Phillip L. Shaffer                        shaffer@crd.ge.com
     GE Corporate Research & Development       uunet!crd.ge.com!shaffer
     Building KW, Room D211
     P.O. Box 8, Schenectady NY 12301