Archives


- Beowulf
- Beowulf Announce
- Scyld-users
- Beowulf on Debian

PORTLAND COMPILERS

Many of your questions may have already been answered in earlier discussions or in the FAQ. The search results page will indicate current discussions as well as past list serves, articles, and papers.

Search

Craig Tierney ctierney at hpti.com
Fri Jun 14 11:15:30 PDT 2002


If all you want to do is compile and run MPI processes you can just
get the PGI workstation set for $560.  You can ask PGI about this
themselves.  I asked the same questions before and this is what
they told me.

Intel are very good for P4/Xeon systems.  I haven't tried their compiler
on P3s (but SSE support is very useful).  Intel did well on AMD systems
as well, but for MY codes the Intel boxes were still faster.

Craig

> Hi Craig,
> 
> I should be more specific in formulating my question: I DO NOT WANT to run
> compilation of programs on each node, all the developer's work is supposed
> to be done on master node. However, resulting MPI binary will be run
> across the cluster. I don't care how many users can compile programs
> simultaneously on master node, but what I care is how many MPI processes
> can I utilize during program execution. Do I need to buy 48 processor
> license bundled in PGI CDK cluster development kit for $3,600 or can I
> survive with just PGI workstation set for $560?
> 
> Ivan
> 
> P.S. People are complaining about quality of the PGI compilers, but still,
> I would like to have several compilers to compare performances for
> different applications.
> 
> ------------------------------------------------------------------------
> 
> On Fri, 14 Jun 2002, Craig Tierney wrote:
> 
> > It means that if you have 128 cpus in your cluster that you
> > have to buy 128 licenses of the CDK.
> > 
> > However, if all you want is their compilers you can buy licenses
> > for only where you compile.  You can download mpich or lam for
> > free and you can run on as many nodes as you like.
> > 
> > Disclaimer, I don't know if they changed the licensing for 4.0, but
> > I doubt it.
> > 
> > Craig
> > 
> ------------------------------------------------------------------------
> Ivan I. Oleynik                       E-mail : oleynik at chuma.cas.usf.edu
> Department of Physics
> University of South Florida
> 4202 East Fowler Avenue                  Tel : (813) 974-8186
> Tampa, Florida 33620-5700                Fax : (813) 974-5813
> ------------------------------------------------------------------------
> 
> _______________________________________________
> Beowulf mailing list, Beowulf at beowulf.org
> To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf

-- 
Craig Tierney (ctierney at hpti.com)



More information about the Beowulf mailing list