From owner-freebsd-hackers  Tue Jun  3 10:42:38 1997
Return-Path: <owner-hackers>
Received: (from root@localhost)
          by hub.freebsd.org (8.8.5/8.8.5) id KAA11402
          for hackers-outgoing; Tue, 3 Jun 1997 10:42:38 -0700 (PDT)
Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.19])
          by hub.freebsd.org (8.8.5/8.8.5) with ESMTP id KAA11287
          for <hackers@FreeBSD.ORG>; Tue, 3 Jun 1997 10:41:15 -0700 (PDT)
Received: (from bde@localhost) by godzilla.zeta.org.au (8.8.5/8.6.9) id DAA24220; Wed, 4 Jun 1997 03:37:06 +1000
Date: Wed, 4 Jun 1997 03:37:06 +1000
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199706031737.DAA24220@godzilla.zeta.org.au>
To: gibbs@plutotech.com, luigi@labinfo.iet.unipi.it
Subject: Re: disk scheduling (was Re: DMA for IDE drives ?)
Cc: hackers@FreeBSD.ORG, toor@dyson.iquest.net
Sender: owner-hackers@FreeBSD.ORG
X-Loop: FreeBSD.org
Precedence: bulk

>My QUANTUM FIREBALL_TM3200S 300X with a 2940A Ultra SCSI does a
>mere 3MB/s writing (iozone 64 8192), on a PPRO-200. Not impressive,
>considering a WDC AC21600H does 4.9MB/s on the same hardware (in
>both cases, a P5-133 gives almost the same results, it is not
>cpu-limited).  The WDC AC21600H has a lower price-per-gigabyte,
>despite the FIREBALL_TM3200S is one of the cheapest SCSI units
>available.

My QUANTUM FIREBALL_TM2110A with Triton-1 IDE does 5572K/s writing
with bonnie -s 64 on the fastest file system that I measured (ext2fs
with a block size of 4K under FreeBSD; fs at offset 320MB) on a P5/133.
2.1GB Fireball_TM's are the same as 3.2MB ones except for fewer heads.

>> won't get anywhere near the throughput of a SCSI drive if you can't
>> schedule and have the drive reorder multiple transactions at a time. 

Right, IDE gives almost twice as much throughput with a Fireball_TM,
at least if it is not slowed down by using DMA :-).  (It seemed to be
only slightly slower using DMA under Linux.)

Of course, SCSI with busmastering DMA is better for systems that
do lots of i/o.  Such systems don't seem to be very common.  E.g.,
freefall has averaged 141K/sec total since boot 5+ days ago.  That's a
whole 35K/sec per disk.  It has only been 89.7% idle.  I don't think
the counters have overflowed.   The counters have overflowed on hub.
My home system has averaged 8K/sec per disk and 78.5% idle over one day.

Bruce