Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!usc!snorkelwacker.mit.edu!stanford.edu!leland.Stanford.EDU!elaine18.Stanford.EDU!dhinds
From: dhinds@elaine18.Stanford.EDU (David Hinds)
Newsgroups: comp.arch
Subject: Re: What's in the '586?
Message-ID: <1991May18.161515.22428@leland.Stanford.EDU>
Date: 18 May 91 16:15:15 GMT
References: <42347@cup.portal.com> <1340@argosy.UUCP> <5814@goanna.cs.rmit.oz.au>
Sender: news@leland.Stanford.EDU (Mr News)
Organization: Stanford University - AIR
Lines: 25

In article <5814@goanna.cs.rmit.oz.au> ok@goanna.cs.rmit.oz.au (Richard A. O'Keefe) writes:
>In article <42347@cup.portal.com> mmm@cup.portal.com (Mark Robert Thorson)
>wrote
>>for enabling the on-chip FPU.  Each password is good for 10 gigaflops,
>>i.e. you get 10,000,000,000 floating point operations.
>
>Note that at one megaflop (and if the 586 FPU can't achieve that, who
>wants it?) 10,000,000,000 floating point operations is about 3 HOURS
>of FP computation.  Whatever you think about Intel, they aren't going
>to expect people to buy passwords every few _hours_!

Seriously, though, (and apart from the issue of how to properly implement
such a scheme), is this really that outrageous in principle?  I'll bet it
would take most PC users an awful long time to use up a billion flops.  Is
a pay-per-use scheme that unreasonable?  You could have a phone service
type scheme, where you could pay a flat price for the chip for unlimited
service, or various other rates depending on your level of use.  Maybe
the chip would carry a meter, and someone could stop by to read it once
every few months.  I might really like having the latest, fastest CPU
available to me all the time, though I only expect to use it maybe 5% of
that time over all.  I confess I can't think of a really good way to do
this for PC's that would not be more trouble than it is worth.

 -David Hinds
  dhinds@cb-iris.stanford.edu