Xref: utzoo comp.sys.dec:3546 comp.unix.ultrix:3903
Path: utzoo!attcan!uunet!pyrdc!gmu90x!gmuvax2!rauletta
From: rauletta@gmuvax2.gmu.edu (R. J. Auletta)
Newsgroups: comp.sys.dec,comp.unix.ultrix
Subject: Is a Sun3/60 *really* Faster than a Vax3600?
Keywords: Subject says it all
Message-ID: <1759@gmuvax2.gmu.edu>
Date: 4 Jul 90 03:33:40 GMT
Followup-To: comp.sys.dec,comp.unix.ultrix
Organization: George Mason Univ., Fairfax, Va.
Lines: 36


I was trying to measure  relative disk performance
on a new disk drive on Sun3/60. So I tried cp'ing
a 1.5Meg file on three machines.

The csh time command reported the following numbers for
a 3/60, Vax3600, Vax8530 all on local disks.

3/60 with 25Meg of Memory, one user. (Sun OS)

0.0u 0.8s 0:01  74%

8530 with about 60Meg of Memory, 25 users (Ultrix).

0.0u .7s :01 76%


Vax3600, 16Meg of Memory, 1 user, but two inactive diskless VS2000
clients (no users logged on.) Drive is an RA82. (Ultrix)

0.0u 1.5s 0:07 20%

The most striking difference is the percentage of CPU's cycles 
each cp used over the period it was active with the VAX3600
quite low. I would interpret this  being due to a another CPU
intensive process running, but nothing showed up with top or ps.
Otherwise, if I read this right, the CPU is waiting on the
RA82?

My question is, is  the combination of a VAX3600 and RA82 really
that much of a dog, or is there someother explanation in
our configuration?


--Rich Auletta
  rauletta@gmuvax2.gmu.edu