[Beowulf] RE: Compare and contrast MPI implementations

Brian D. Ropers-Huilman bropers at cct.lsu.edu
Thu Dec 15 12:03:06 PST 2005


-----BEGIN PGP SIGNED MESSAGE-----
Hash: RIPEMD160

One simple answer is to have your job scripts establish your environment
variables for you via exports (or setenv if you use those other shells).

David Mathog said the following on 2005-12-15 12:53:
> Thanks for the feedback so far.  
> 
> Now how about MPI coexistance?
> 
> LAM-MPI wants PATH set in .bashrc or some similar
> startup file.  If memory serves, that was also the method
> used by Mpich(1.x).  MPICH2 at least seems to have a way
> around this since mpdboot works with no twiddling of
> startup files on the remote nodes.
> 
> So then, if LAM-MPI and MPICH(1.x) both want PATH set to point
> to their own binary directories at login, from a single account,
> how does one run program A that wants MPICH and program B that wants
> LAM-MPI? Hopefully there's something better than "change the
> .bashrc file before each run"!
> 
> Thanks,
> 
> David Mathog
> mathog at caltech.edu
> Manager, Sequence Analysis Facility, Biology Division, Caltech

- --
Brian D. Ropers-Huilman, Director     .:.    High Performance Computing
Networking, Infrastructure, & Research / Information Technology Support
Frey Computing Services, Rm. 323                        bropers at lsu.edu
Louisiana State University                          +1 225.578.3272 (V)
Baton Rouge, LA 70803-1900  USA                     +1 225.578.5362 (F)
http://www.lsu.edu/                             http://www.cct.lsu.edu/
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFDocv6wRr6eFHB5lgRA/IYAKCWKrR/f01OMcf5CS792fYi92+S8QCg7/9N
aU+RDUlUaK42tT5H/RUG/8s=
=Em+0
-----END PGP SIGNATURE-----



More information about the Beowulf mailing list