From owner-freebsd-hackers  Sat Oct  7 16:10:20 1995
Return-Path: owner-hackers
Received: (from root@localhost)
          by freefall.freebsd.org (8.6.12/8.6.6) id QAA29685
          for hackers-outgoing; Sat, 7 Oct 1995 16:10:20 -0700
Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.34])
          by freefall.freebsd.org (8.6.12/8.6.6) with ESMTP id QAA29679
          for <hackers@freebsd.org>; Sat, 7 Oct 1995 16:10:16 -0700
Received: (from bde@localhost) by godzilla.zeta.org.au (8.6.9/8.6.9) id JAA05772; Sun, 8 Oct 1995 09:07:19 +1000
Date: Sun, 8 Oct 1995 09:07:19 +1000
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199510072307.JAA05772@godzilla.zeta.org.au>
To: bde@zeta.org.au, gibbs@freefall.FreeBSD.org
Subject: Re: VLB Disk Controllers
Cc: dennis@etinc.com, hackers@freebsd.org
Sender: owner-hackers@freebsd.org
Precedence: bulk

>>That's not saying much. An IDE controller has a much lower command
>>overhead than the buslogic:

>It also has much higher CPU overhead, but we've been down that road
>before. :)

:-).

>>A high command overhead causes slow file system operations for
>>everything except large i/o's.

>Yup, but your whole machine is affected by a PIO device.

That's mainly because IDE PIO is slow, not because it's PIO.  The
overhead for SCSI DMA is not insignificant.  For the BT455C, which
has a DMA speed of up to 40MB/sec (I doubt that my bus can keep up
with this), the overhead for reading from a Grand Prix at 5MB/sec
seems to be about 27%.  This is hard to measure accurately.  I
measured it by running a hog process in the background: on an idle
system it took 7.58 sec (real) and while transferring it took 10.42
sec.  7.58/10.42 = 73%.

Bruce