Received: from relay2.UU.NET (relay2.UU.NET [192.48.96.7]) by keeper.albany.net (8.7.1/8.7.1) with ESMTP id RAA27988 for <dwarner@albany.net>; Sun, 17 Dec 1995 17:45:55 -0500 (EST)
Received: from garcia.com by relay2.UU.NET with SMTP
id QQzuof25982; Sun, 17 Dec 1995 17:27:16 -0500 (EST)
Received: from (localhost) by garcia.com (5.x/SMI-SVR4)
id AA06898; Sun, 17 Dec 1995 17:27:23 -0500
Date: Sun, 17 Dec 1995 17:27:23 -0500
Errors-To: dwarner@albany.net
Message-Id: <199512160713.CAA14173@bort.mv.net>
Errors-To: dwarner@albany.net
Reply-To: lightwave@garcia.com
Originator: lightwave@garcia.com
Sender: lightwave@garcia.com
Precedence: bulk
From: Mark Thompson <mark@fusion.mv.com>
To: Multiple recipients of list <lightwave@garcia.com>
Subject: Re: Out to Video
X-Listprocessor-Version: 6.0c -- ListProcessor by Anastasios Kotsikonas
Status: RO
X-Status:
>Actually, I seriously doubt that scsi-2 will outperform PCI. PCI 64 bit has
>a maximum data transfer rate of 528 MB per second. thats megaBYTES, not bits.
Well, thats only in theory. In practice, things are extremely different.
First of all, that 528MB number is only a peak figure for fast and wide
PCI. Since NIETHER fast OR wide PCI is in common use today, that theorical
peak is 1/4 of that, or 132MB/s. Then when you take into account all the
overhead of running the PCI buss, coupled with all the other crap using
it (like the CPU), suddenly you don't have much dedicated bandwidth
leftover. Add to this that most PCI interface implementations are much
less than wonderful in their speed and efficiency. The upshot of this is
that you will generally get much better continuous bandwidth out of a
dedicated SCSI-2 port than over a shared PCI bus. That is why the PVR