Path: utzoo!mnetor!uunet!lll-winken!lll-lcc!ames!ll-xn!mit-eddie!uw-beaver!cornell!rochester!crowl
From: crowl@cs.rochester.edu (Lawrence Crowl)
Newsgroups: comp.arch
Subject: Re: Sun-3/50 speed variations
Message-ID: <8055@sol.ARPA>
Date: 26 Mar 88 04:29:35 GMT
References: <4076@vdsvax.steinmetz.ge.com> <3505@cbmvax.UUCP> <4238@hoptoad.uucp>
Reply-To: crowl@cs.rochester.edu (Lawrence Crowl)
Organization: U of Rochester, CS Dept, Rochester, NY
Lines: 22

In article <4238@hoptoad.uucp> gnu@hoptoad.uucp (John Gilmore) writes:
>barnett@vdsvax.steinmetz.ge.com (Bruce G. Barnett) says:
>> I still don't understand how one Sun 3/50 can be 40% faster than another.
>
>Curiously enough, the video display eats about 40% of the memory bandwidth --
>unless you turn it off (/usr/bin/screenblank).  Then it goes down into the
>noise (just enough to refresh RAM).  ...  Perhaps the benchmark was run "over
>the net" on a screen-blanked machine?

I did just such a benchmark.  Here are the Dhrystone numbers:

SunOS, cc, pc                   C (noreg)        C (reg)         Pascal

Sun 3/50  on console            2267.0 (1.00)    2517.5 (1.00)   2486.3 (1.00)
with video display on
Sun 3/50  via rlogin            2898.6 (1.28)    3180.2 (1.26)   3231.7 (1.30)
with video display off

So turning the display off yields a (relatively) consistent 28% speedup.
-- 
  Lawrence Crowl		716-275-9499	University of Rochester
		      crowl@cs.rochester.edu	Computer Science Department
...!{allegra,decvax,rutgers}!rochester!crowl	Rochester, New York,  14627