what architecture was MPI and PVM 1st designed for?

[Robert G. Brown]

On Tue, 16 Apr 2002, Jayne Heger wrote:

> 
> Hi,
> 
> Coulld anyone tell me what computer architecture MPI and PVM were first 
> designed for./written on.

See

  http://www.epm.ornl.gov/pvm/

and look under "documentation" for "PVM and MPI: A comparison of
features".  Read the "Background" section.

Among many other sources, but this is terse and probably adequate, close
to "horse's mouth" accurate for PVM (but "Project Overview" is also
there and IS horse's mouth:-) and of course I'm sure that the primary
MPI sites have similar historical stuff linked.

In a very terse nutshell, PVM was written for the kitchen sink (whatever
you happened to have handy and networked).  MPI was written by a
consortium of vendors and users to provide a common API for large,
expensive massively parallel computers.  As I understand it this wasn't
really the vendors' idea -- they would've been happy to continue
providing only their proprietary interfaces -- but the government
finally put its foot down as it learned just how much money it was
spending, first on the iron, then on porting code to run on the iron,
and then on NEW iron and RE-porting their ported code to run on the NEW
iron, etc.  Moore's law demanding that they rebuy everything every few
years or actually loose ground, of course...

   rgb

> Thanks,
> 
> Jayne Heger
> _______________________________________________
> Beowulf mailing list, Beowulf at beowulf.org
> To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf
> 

-- 
Robert G. Brown	                       http://www.phy.duke.edu/~rgb/
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone: 1-919-660-2567  Fax: 919-660-2525     email:rgb at phy.duke.edu