Path: utzoo!mnetor!uunet!husc6!bbn!rochester!PT.CS.CMU.EDU!G.GP.CS.CMU.EDU!butcher
From: butcher@G.GP.CS.CMU.EDU (Lawrence Butcher)
Newsgroups: comp.arch
Subject: Re: Tasting of Dhrystone 2.0 Results
Message-ID: <1141@PT.CS.CMU.EDU>
Date: 17 Mar 88 00:32:44 GMT
References: <503@pcrat.UUCP> <4074@vdsvax.steinmetz.ge.com> <506@pcrat.UUCP>
Sender: netnews@PT.CS.CMU.EDU
Organization: Carnegie-Mellon University, CS/RI
Lines: 20

I submitted two of the 3/50 numbers, the (reg)2859 and (reg)3452.
You may reproduce them.

The Sun 3/50 is a 68020 running at 15 MHz.  You would guess that it is
(15/16.666) times as fast as a 3/160, which has a 16.66 MHz clock.  However
the display steals memory cycles from the processor.  Under Sun Unix 3.5
(and maybe older versions) the display blanks after the machine has been
idle for a while.

To get my fast number I telnetted to an idle Sun from another machine and
did the measurement.  I then did "echo toads >/dev/console" and the display
unblanked.  Measuring again I got the slower number.

My numbers are probably +-10%.  Rick will have to get several independent
measurements to zero in on the correct number.  The Sun 3/50 should have two
ratings, and would exhibit the faster performance if it were used as a
server.

			Surprised me too:
				Lawrence Butcher