Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!uwm.edu!psuvax1!hsdndev!spdcc!dirtydog!suitti
From: suitti@ima.isc.com (Stephen Uitti)
Newsgroups: comp.benchmarks
Subject: Re: Upper bound on numbers of users
Keywords: Workstations, Mini-computers, Main-frames
Message-ID: <1991Mar20.155013.18691@ima.isc.com>
Date: 20 Mar 91 15:50:13 GMT
References: <3468@naucse.cse.nau.edu>
Sender: usenet@ima.isc.com
Reply-To: suitti@ima.isc.com (Stephen Uitti)
Organization: Interactive Systems, Cambridge, MA 02138-5302
Lines: 44

In article <3468@naucse.cse.nau.edu> jdc@naucse.cse.nau.edu (John Campbell) writes:
>I'm trying to compile a table...
>What I'd like, is to hear from other sites and try to get an
>idea on the number of users actually using different machines 
>concurrently.  In other words, what machines are you running 
>and what's an "average" number of users?  I'm particularly
>interested in any sites that run "large" (>5) numbers of users
>on Workstations (we don't).

At Purdue in '82 or '83, I recall VAX 780s with 45-50 users, instructional.

They also had dual CPU VAX 780s, with 100-120 users, instructional.
	Response could be termed sluggish.

In about '86, Purdue had an IBM 3083 (roughly 10 MIPS).  I've no
idea what the average user count was, but one time I noticed that
a particular 40 CPU second job was completed in 40 seconds of
real time, give or take a tenth of a second.  Then I noted that
there were 300 users logged in.  Presumably, most of the users
were editing - a task mainly handled by the block-mode terminals.
It isn't EMACS, but I was starting to have some appreciation for xedit.

Three years ago at Harvard, we ran uVAX IIs with 20-35 users
each, instructional.
	Response was adequate, but not awesome.

Computer        Memory     Disk    #accts   Avg # users    (Mips)    O.S.
VAX 780             8Mb     1-2Gb    1200         45           1      4.2 BSD
Dual 780           16Mb     1-2Gb   >1000        100           2      4.2 BSD
uVAX II            13Mb  .9-1.5Gb     200         30           1   Ultrix 2.0

It depends quite a bit on what you are doing with the machine.
In some environments, there are no users - just some application
which trashes the machine.  Two developers may bring a system to
its knees.  If the project is different, ten developers might not
have any problems with user response.

I've seen very good consistent user-response on multi-CPU
machines, such as from Sequent & Multimax.  With lots of CPUs, it
looked like you could run huge numbers of users.  I never saw
that many.

Stephen.
suitti@ima.isc.com