From owner-freebsd-hackers  Sat Nov  8 04:33:51 1997
Return-Path: <owner-freebsd-hackers>
Received: (from root@localhost)
          by hub.freebsd.org (8.8.7/8.8.7) id EAA05845
          for hackers-outgoing; Sat, 8 Nov 1997 04:33:51 -0800 (PST)
          (envelope-from owner-freebsd-hackers)
Received: from rah.star-gate.com (rah.star-gate.com [204.188.121.18])
          by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id EAA05839;
          Sat, 8 Nov 1997 04:33:39 -0800 (PST)
          (envelope-from hasty@rah.star-gate.com)
Received: from rah.star-gate.com (localhost.v-site.net [127.0.0.1])
	by rah.star-gate.com (8.8.7/8.8.5) with ESMTP id EAA02814;
	Sat, 8 Nov 1997 04:33:28 -0800 (PST)
Message-Id: <199711081233.EAA02814@rah.star-gate.com>
X-Mailer: exmh version 2.0gamma 1/27/96
To: Gordon Henderson <gordon@drogon.net>
cc: Stefan Esser <se@FreeBSD.ORG>, hackers@FreeBSD.ORG
Subject: Re: Seagate Cheetah performance... 
In-reply-to: Your message of "Sat, 08 Nov 1997 11:42:41 GMT."
             <Pine.LNX.3.95.971108113529.31803A-100000@unicorn> 
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Date: Sat, 08 Nov 1997 04:33:27 -0800
From: Amancio Hasty <hasty@rah.star-gate.com>
Sender: owner-freebsd-hackers@FreeBSD.ORG
X-Loop: FreeBSD.org
Precedence: bulk


Hi,

I think I now have my answer:
dd if=/dev/rsd0f of=/dev/null bs=16k 
^C41776+0 records in
41776+0 records out
684457984 bytes transferred in 51.149338 secs (13381561 bytes/sec)

dd if=/dev/rwd0f of=/dev/null bs=16k
^C15902+0 records in
15902+0 records out
260538368 bytes transferred in 32.176591 secs (8097140 bytes/sec)

The magic in my case was to fine tune my scsi adapter:

options AHC_TAGENABLE
options AHC_SCBPAGING_ENABLE
options AHC_ALLOW_MEMIO


Hope that this is an FAQ for web servers or file server configurations

	Amancio


> On Sat, 8 Nov 1997, Stefan Esser wrote:
> 
> > Please test all three drives with Bonnie. Use a test file at 
> > least twice as much as your system's RAM, and report your 
> > findings ...
> > 
> > You'll get a different picture, I assume !
> 
> As Stefan says, you have to use a test file size more than your memory
> system has, and the bigger the better, and bonnie is fairly universally
> accepted, isn't it? 
> 
> To show you how misleading things can be, when a memory cache gets in the
> way, heres a bonnie outout on a 32MB file (on a 256MB machine) 
> 
>       -------Sequential Output-------- ---Sequential Input-- --Random--
>       -Per Char- --Block--- -Rewrite-- -Per Char- --Block--- --Seeks---
>    MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU  /sec %CPU
>    32  6321 96.0 26077 76.1 14800 91.8  6127 94.8 40595 94.4 3320.7 90.0
> 
> 40MB/sec read rate! Impressive, huh? (interesting to note how innefficient
> character reading/writing is over block reading/writing!) 
> 
> Is there a Bonnie that copes with > 2GB files yet? I'm using one that I've
> had for several years now... 
> 
> Gordon
>