Newsgroups: comp.parallel.mpi
From: lusk@mcs.anl.gov (Rusty Lusk)
Subject: Re: Running MPI without rsh/rexec access
Organization: Math and Computer Science, Argonne.
Date: 26 May 1995 18:33:02 GMT
Message-ID: <3q56su$rcr@milo.mcs.anl.gov>

In article <3q04fj$bu5@news.uncc.edu>, Stuart D Blackburn <sdblackb@uncc.edu> writes:
|> 
|> 	The professor that I am working with on this, has attended
|> workshops where MPI was discussed, and was interested to see if we
|> could try to use it here in addition to or instead of PVM for the
|> Parallel Programming class that he teaches. I was able to get the
|> mpich implementation of MPI installed, and it seems to work fine
|> using the ch_p4 device on a single machine. This of course isn't
|> very interesting though, and I can't seem to find any info on
|> starting the "secure server" for MPI without rsh capabilities.
|> 
|> 	I have beginning to believe that the this is not possible
|> given the lack of documentation on the subject. If anyone has any
|> experience with this using any freely distributed implementation of
|> MPI I would be VERY grateful to hear of it.  Thanks in advance for
|> your time and assistance.
|> 

You should be able to find instructions for starting the secure server
in the mpich-installation guide in the mpich/doc directory.  See also
the User's Guide in the same directory.  If you don't see it there
then I can send you a more up-to-date version of the installation
guide.  We use it here and it works.

Of course, for other machines on your network where you *do* have rsh
access, things are easier.  I presume that you did not have any
difficulties with this case.  If so, please let us know at
mpi-bugs@mcs.anl.gov, which is the normal place for problems related
to MPICH in particular.

Regards,
Rusty Lusk

