Report on SMP issues

Douglas Eadline deadline at plogic.com
Mon Mar 24 13:37:54 PST 2003


On Mon, 24 Mar 2003, Martin Siegert wrote:

> On Mon, Mar 24, 2003 at 07:58:06PM +0100, Joachim Worringen wrote:
> > Douglas Eadline:
> > 		> All
> > >
> > > I have just posted some results and discussion concerning SMP nodes
> > > and clusters.
> > >
> > > You can read and comment on the report at:
> > > http://www.cluster-rant.com/article.pl?sid=03/03/24/1731246
> > 
> > It's nice to have such a comparison -  but it does not make much sense to 
> > *only* compare speedups - you should add the wall clock times for the test 
> > runs, too.
> 
> Furthermore, if you are interested in the question whether it better to
> buy 2x single processor boxes or x dual processor boxes, this comparison
> does not tell you very much: you really should keep the total memory
> constant, i.e., configure the duals with twice as much RAM as the single.

You are correct. The price performance analysis is not finished.
Stay tuned for more data/issues etc. on this topic. After I/we
have gone though an analysis of the issues involved with SMP
nodes, I am going to try and develop some price/performance metrics
that can be used to design systems.
 

Doug


-- 
-------------------------------------------------------------------
Paralogic, Inc.           |     PEAK     |      Voice:+610.814.2800
130 Webster Street        |   PARALLEL   |        Fax:+610.814.5844
Bethlehem, PA 18015 USA   |  PERFORMANCE |    http://www.plogic.com
-------------------------------------------------------------------




More information about the Beowulf mailing list