Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!wuarchive!mit-eddie!uw-beaver!tera.com!doc From: doc@tera.com (Dan Cummings) Newsgroups: comp.unix.large Subject: Re: Hello Message-ID: <1990Sep11.203723.24585@tera.com> Date: 11 Sep 90 20:37:23 GMT References: <6f7y02Ubc6wm01@amdahl.uts.amdahl.com> <25894@boulder.Colorado.EDU> <7914@gollum.twg.com> Sender: news@tera.com Organization: Tera Computer Company, Seattle, WA Lines: 24 In lowa@imtws3.im.se (Leif-Olof Wallin IMT) writes: >How about the following definition just to get the discussion started: > > 200 users > > 10 GB of disk > > 50 Mips All right, I'm game. Just for starters I would say that this is not a very well balanced system by today's standards. I am biased in this statement by a long term involvement with supercomputer Unix implementations. More than 200 users implies that you have a fair chunk of memory. With this much memory 10GB of disk is probably inadequate. 50 mips is probably also inadequate. The Cray-2 at NERSC in Livermore california supports 200 users at a time as do, I believe, the YMP installations at NASA Ames and the YMP at San Diego Supercomputer Center. Is this the class of machine we are talking about? If so then this is a supercomputer Unix forum. Such a forum is worthwhile now that there are other contenders besides Cray research in this market. Dan Cummings Tera Computer Company doc@tera.com