*BSD News Article 36544


Return to BSD News archive

Path: sserve!newshost.anu.edu.au!harbinger.cc.monash.edu.au!msuinfo!gmi!zombie.ncsc.mil!MathWorks.Com!europa.eng.gtefsd.com!library.ucla.edu!ihnp4.ucsd.edu!ames!newsfeed.gsfc.nasa.gov!cesdis1.gsfc.nasa.gov!not-for-mail
From: becker@cesdis.gsfc.nasa.gov (Donald Becker)
Newsgroups: comp.os.386bsd.misc
Subject: Re: dual processor motherboards the way forward?
Date: 4 Oct 1994 16:39:37 -0400
Organization: NASA Goddard Space Flight Center -- Greenbelt, Maryland USA
Lines: 33
Message-ID: <36sei9$g3j@cesdis1.gsfc.nasa.gov>
References: <Cww6x9.1A0@gnome.co.uk>
NNTP-Posting-Host: cesdis1.gsfc.nasa.gov
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit

In article <Cww6x9.1A0@gnome.co.uk>,
Dr Chris Stenton  <jacs@gnome.co.uk> wrote:
>I have just had some information sent to me on a dual pentium PCI-EISA
>motherboard which is now availablle. The blurb says that there is some
>crude (my word) support under windows NT where you can select an
>application to run on the second processor. This made think what use
>this additional technology could have for intel based unix. Would it
>be better to just place the kernel on one processor and run all other
>processes on the other; or some form of dynamic load balance which
>will have some overhead? Anyway I am sure the FreeBSD and NetBSD teams
>could have some fun with this after they have tired with porting 4.4
>BSD.

I've look at a few of the dual P5 motherboards, and I've concluded that the
current generation will never be more than a niche product.  Their primary
advantage is low cost -- a two processor motherboard costs about the same as
two single processor motherboards.  Their primary disadvantage is low
scalability and relatively low performance.

The boards are designed so that the P5 processors share everything,
including the L2 cache and path to main memory.  This limits the performance
gain of adding the second process to well under a factor of two.  I'm sure
that there are a few applications where its possible to get an extra 25-50%,
but it hardly seems worth redesigning a whole software system for such
minimal gain.  IMO, more processors running in a loosely-coupled cluster
offers a much better performance gain, without locking you into a specific
hardware configuration.

-- 
Donald Becker					  becker@cesdis.gsfc.nasa.gov
USRA-CESDIS, Center of Excellence in Space Data and Information Sciences.
Code 930.5, Goddard Space Flight Center,  Greenbelt, MD.  20771
301-286-0882	     http://cesdis.gsfc.nasa.gov/pub/people/becker/whoiam.html