Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: Notesfiles $Revision: 1.6.2.17 $; site convex.UUCP
Path: utzoo!watmath!clyde!cbosgd!ihnp4!inuxc!pur-ee!uiucdcs!convex!masson
From: masson@convex.UUCP
Newsgroups: net.unix-wizards
Subject: Re: Convex C-1
Message-ID: <27300003@convex.UUCP>
Date: Tue, 20-Nov-84 19:27:00 EST
Article-I.D.: convex.27300003
Posted: Tue Nov 20 19:27:00 1984
Date-Received: Thu, 22-Nov-84 05:24:08 EST
References: <17150@arizona.UUCP>
Lines: 22
Nf-ID: #R:arizona:-1715000:convex:27300003:000:752
Nf-From: convex!masson    Nov 20 18:27:00 1984

>Has anyone had a chance to do any UNIX benchmarks on the Convex C-1?
>
>					Bill Mitchell
>					whm.arizona@csnet-relay
>					{noao,mcnc,utah-cs}!arizona!whm

We have run quite a number of benchmarks on the C-1.  The
C-1 is touted as "The World's First Affordable Supercomputer"
because of its Cray-like implementation.   Because of its
integrated vector processing, benchmarks that are highly
vectorizable tend to really shine.  Typical performance
for traditional supercomputer benchmarks (e.g. "Livermore
Loops") put the C-1 at about 1/4 of the CRAY, or about 
40-60x the VAX.

Note that these benchmarks are FORTRAN and take advantage of
the Convex FORTRAN compiler, which does vectorization and
optimization for you.  

Bob Masson
convex!masson