Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!seismo!rochester!cornell!uw-beaver!mit-eddie!ll-xn!ames!lll-tis!ptsfa!ihnp4!alberta!bacon
From: bacon@alberta.UUCP (Dave Bacon)
Newsgroups: comp.arch
Subject: Re: chewing up mips with graphics
Message-ID: <393@pembina.UUCP>
Date: Mon, 22-Jun-87 22:25:17 EDT
Article-I.D.: pembina.393
Posted: Mon Jun 22 22:25:17 1987
Date-Received: Wed, 24-Jun-87 05:51:53 EDT
References: <8270@amdahl.amdahl.com> <359@rocky2.UUCP> <2194@mmintl.UUCP>
Organization: U. of Alberta, Edmonton, AB
Lines: 23
Keywords: simulation processors parrallelism MIPS
Summary: there would be no capacity loss

In article <2194@mmintl.UUCP>, franka@mmintl.UUCP (Frank Adams) writes:
> In article <6328@beta.UUCP> dzzr@beta.UUCP (Douglas J Roberts) writes:
> >Does somebody REALLY think we are going to run out of work before
> >we run out of MIPS? Fat chance.
> ....  It is true that a 100-fold increase in
> power in the next 3 months would create a temporary overcapacity; one it
> might take a couple of years to make much use of.  But I see no sign of such

	If the production of such machines were limited (which could
be expected with such a quantum leap). I think the one could easily
find the people who could use such machines, especially if they were 
priced in the same range as current machines and were capable of using
current software with minimal change. We had one person here, who used
~100 hours of cray time creating a molecular animation sequence, and
we would be QUIT happy to be able to do things like that in (almost)
real time.

  Granted: not EVERYBODY could make reasonable use of such machines but, if 
you give me an immensely powerful machine, I'm sure I could find at least
a dozen people who would be happy to use it within a week (assuming the 
software could be ported over that fast).

    Stephen Samuel {ihnp4,seismo!mnetor,vax135}!alberta!edm!steve