Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 21 Mar 1996 10:37:09 -0700 (MST)
From:      Terry Lambert <terry@lambert.org>
To:        msmith@atrad.adelaide.edu.au (Michael Smith)
Cc:        edmond@UWYO.EDU, msmith@atrad.adelaide.edu.au, questions@FreeBSD.org
Subject:   Re: Hardware Tailoring...
Message-ID:  <199603211737.KAA00307@phaeton.artisoft.com>
In-Reply-To: <199603210553.QAA11133@genesis.atrad.adelaide.edu.au> from "Michael Smith" at Mar 21, 96 04:23:55 pm

next in thread | previous in thread | raw e-mail | index | archive | help
> Andrew N. Edmond stands accused of saying:
> > > > Inter Pentium Pro 200mhz
> > > 
> > > This is a very poor choice.  None of the PPro motherboards on the open 
> > > market at the moment are suitable for use in servers.  You will either
> > > need to purchase a system from Intel's Server division (the 'Alder' board
> > > is known to work well), or go back to a high-end P5.
> > 
> > It *is* an Intel motherboard - not proprietary at all (though I did find 
> > many of them out there).  If it is an INTEL motherboard, I figure I 
> > should be fine. 
> 
> Read My Lips.  None of the PPro motherboards on the open market...
> I didn't make the point for the fun of it.  All of the commonly-available
> PPro motherboards are based on the Orion chipset.  This chipset is
> significantly flawed.  The flaw limits PCI bus bandwidth to about 4M/sec.
> This is Very Slow.

It might help him to know that there are several divisions in Intel
manufacturing motherboards, but only one division really selling
them: the "OEM Products Division".  Boards from this division are
uniformly inferior to boards from the other divisions in general,
and the Server division in particular.  The Server division _Simply_
Makes_Good_Boards_.

Previous examples from the OEM Products division include the Plato
and Zappa motherboards, which insist on sharing PCI interrupts --
a good way to spend your time demuxxing interrupts and serializing
operations and all that other "good stuff" that you would rather
not do, instead of calculating and pushing data.  Yes, it's "allowable"
under the PCI spec, but that doesn't mean it's a good thing, or even
desirable under any circumstances (ask Buslogic about the boards).


					Terry Lambert
					terry@lambert.org
---
Any opinions in this posting are my own and not those of my present
or previous employers.



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199603211737.KAA00307>