Re: Access to UNIX cluster-anybody want some processing time?(was:Re: Tesla simulation software project)

["Jim Lux" <jimlux-at-jpl.nasa.gov> (by way of Terry Fritz <twftesla-at-uswest-dot-net>)]

> > 
> > Some time ago I put together a time domain simulation of
> > 100 segments, but it was too much even for a 12 processor
> > cluster. For now I would be very happy to successfully model
> > the steady state AC input impedance upto and including the
> > 5/4 overtone.
> 
> I have access to 20 C180s and C400s HP-UX running 10.20 on a fibre net at
> Agilent for night time processing-If you or anybody else wants to send me
> your code or anybody else has a TC simulator that needs heavy
> processing-(How fast can you say E-Tesla 5000x5000 array-IF you can work
> out those anomalies, TF-and get it to integrate with Beowulf).

There are a number of EM codes (semishareware) out there that use MPI or
PVM, and would be suitable for running on a Beowulf cluster.  For that
matter, there are also large system solver codes that have run on Beowulfs
as well.

I'd start my researches at http://emlib.jpl.nasa.gov/ which has a list of
various EM programs.  There is also a fair amount of stuff on the Beowulf
discussion lists (archived somewhere, but I can't recall off hand).