Global Shared Memory and SCI/Dolphin

James Cownie jcownie at
Wed Jul 16 11:36:23 EDT 2003

> > Because MPI is what most people want to achieve code- and
> > peformance-portability.

>   Partially I may agree, partially - not: MPI is not the best in the
> sense of portability (for example, optimiziation requires knowledge
> of interconnect topology, which may vary from cluster to cluster,
> and of course from MPP to MPP computer).

MPI has specific support for this in Rolf Hempel's topology code,
which is intended to allow you to have the system help you to choose a
good mapping of your processes onto the processors in the system.

This seems to me to be _more_ than you have in a portable way on the
ccNUMA machines, where you have to worry about

1) where every page of data lives, not just how close each process is
   to another one (and you have more pages than processes/threads to
   worry about !)

2) the scheduler choosing to move your processes/threads around the

> I think that if there is relative cheap and effective way to build
> ccNUMA system from cluster - it may have success.

Which is, of course, what SCI was _intended_ to be, and we saw how
well that succeeded :-(

-- Jim 

James Cownie	<jcownie at>
Etnus, LLC.     +44 117 9071438
Beowulf mailing list, Beowulf at
To change your subscription (digest mode or unsubscribe) visit

More information about the Beowulf mailing list