From owner-freebsd-hackers  Fri Jun  2 22:36:38 1995
Return-Path: hackers-owner
Received: (from majordom@localhost)
          by freefall.cdrom.com (8.6.10/8.6.6) id WAA29896
          for hackers-outgoing; Fri, 2 Jun 1995 22:36:38 -0700
Received: from crh.cl.msu.edu (crh.cl.msu.edu [35.8.1.24])
          by freefall.cdrom.com (8.6.10/8.6.6) with SMTP id WAA29882
          for <freebsd-hackers@freebsd.org>; Fri, 2 Jun 1995 22:36:31 -0700
Message-Id: <199506030536.WAA29882@freefall.cdrom.com>
Received: by crh.cl.msu.edu
	(1.38.193.4/16.2) id AA25376; Sat, 3 Jun 1995 01:36:15 -0400
From: Charles Henrich <henrich@crh.cl.msu.edu>
Subject: Re: A performance mystery
To: rls@kilroy.id.net (Robert Shady)
Date: Sat, 3 Jun 1995 01:36:15 -0400 (EDT)
Cc: henrich@crh.cl.msu.edu, rgrimes@gndrsh.aac.dev.com,
        freebsd-hackers@freebsd.org
In-Reply-To: <199506031308.JAA02922@hades.id.net> from "Robert Shady" at Jun 3, 95 09:08:30 am
X-Mailer: ELM [version 2.4 PL23]
Content-Type: text
Content-Length: 1571      
Sender: hackers-owner@freebsd.org
Precedence: bulk

> I missed the original message, but I'm guessing this is to an MFS mounted
> files system?  Or what..

Here's the situation, I have a Compaq Prolinea 575 (75mhz Pentium) w/ 24mb ram,
and a 3gig Seagate ST3400N (number may be a tad off) drive.  I also have a
Micron P90 Powerstation w/ 64mb ram, and a brand new connor drive. The Compaq
will compile a kernel in 8:30 that takes the Micron 9:30.  My question was, did
anyone have any idea why this is?  I just re-ran all the tests and here is the
results (slightly different kernel config)

                Compaq Proliant 575             Micron P90 Powerstation

time make       8:55                            9:39
ram-speed       10.85 Mb/S, 7.08 Mb/S           10.61 Mb/S, 20.66 Mb/S
iozone 200      2486434/4150208                 2446772/4203499
iozone 16 8192  3360694/2677660                 2711469/13256071

The make time is wall time, ram-speed tests the raw memory bandwidth, defeating
cache almost 100%.  iozone 200 tests raw read/write performance to disk, as it
far exceeds ram size, memory cache effect is insignificant.  iozone 16 8192
tests the memory speed with cache involved, because the test size is so small,
(as Rod pointed out) its essentially just a bunch of bcopy's.

As you can see by the ram-speed and iozone tests (and the fact that its a
Pentium-75!) the Compaq should be sorely beaten in build time over the Micron,
but yet its not.  Hence the mystery.

-Crh

    Charles Henrich     Michigan State University     henrich@crh.cl.msu.edu

                     http://rs560.msu.edu/~henrich/