From: Ramon Huerta <huerta@minerva.ii.uam.es>
Newsgroups: comp.parallel.pvm
Subject: PVM versus MPI
Date: Sat, 03 Apr 1999 15:44:49 +0200
Organization: Depatment of Computer Engineering
Message-Id: <37061B51.69234DB2@minerva.ii.uam.es>
Reply-To: huerta@minerva.ii.uam.es
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Xref: ukc comp.parallel.pvm:8213


I have a general question. I am currently writing my programs using PVM.
They work nicely on LINUX and SUN machines.

Now I am reading the MPI documentation and they claim to be standard.
It seems that they try to absorb PVM and many others. So at this point
I am confused. Shall I continue using PVM or turn into MPI?

Is there any efficiency issue behind?

I find PVM a natural way to go. Is MPI as natural as PVM?

Would you think that since MPI is a strongly funded project
will end up being the "STANDARD"?

I'd appreciate any insight
thanks in advance
Ramon