From owner-freebsd-hackers  Fri Mar  8 21:04:31 1996
Return-Path: owner-hackers
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.3/8.7.3) id VAA19287
          for hackers-outgoing; Fri, 8 Mar 1996 21:04:31 -0800 (PST)
Received: from emout09.mail.aol.com (emout09.mx.aol.com [198.81.11.24])
          by freefall.freebsd.org (8.7.3/8.7.3) with SMTP id VAA19276
          for <hackers@freebsd.org>; Fri, 8 Mar 1996 21:04:29 -0800 (PST)
From: StevenR362@aol.com
Received: by emout09.mail.aol.com (8.6.12/8.6.12) id AAA00979; Sat, 9 Mar 1996 00:05:19 -0500
Date: Sat, 9 Mar 1996 00:05:19 -0500
Message-ID: <960309000518_345746242@emout09.mail.aol.com>
To: luigi@labinfo.iet.unipi.it, hackers@freebsd.org
Subject: Re: IDE performance
Sender: owner-hackers@freebsd.org
X-Loop: owner-hackers@FreeBSD.ORG
Precedence: bulk

In a message dated 96-03-08 18:13:30 EST, luigi@labinfo.iet.unipi.it (Luigi
Rizzo) writes:

>Can someone tell me why my IDE disk (WDC AC31600, 1.6GB) mounted
>on the primary IDE (together with a slave ATAPI CDROM) runs at
>5.5MB/s (both with iozone 128 8192 and with bonnie, block transfer),
>and the same disk on the secondary IDE does "just" little more than
>3.1MB/s ?
>
>This is on a Pentium100, Intel Zappa MB with built-in IDE controller,
>32MB ram, 2.1R. The wdc driver is compiled with flags 0x80ff80ff on
>both primary and secondary interfaces.
>
>Is this a feature of the motherboard, or something in the kernel ?
>(or just because IDE sucks... :)

     I don't know about the Zappa MB's, but the Intel Plato motherboards
state
in their documentation that the primary IDE interface is PCI and that the
secondary one is ISA.  This would probably account for the performance
difference.

Steve