Hello everybody,
Sorry about that incomplete mail ;-)
I set up a 48 node Linux RedHat 9.0 Beowulf cluster and wanted to use
LAM/MPI on it. I've experience only with MPICH.
Installation of LAM/MPI 7.0.3 was easy and straight forward, but I
cannot use mpirun to run jobs. I have a precompiled version of a
parallel program using LAM/MPI that runs fine on other systems, but as
soon as I want to run it on my system, I get the following messages:
lars_at_star29:~/lamtest>lamboot -v -b machinefile
LAM 7.0.3/MPI 2 C++/ROMIO - Indiana University
n0<5028> ssi:boot:base:linear: booting n0 (star29)
n0<5028> ssi:boot:base:linear: booting n1 (star23)
n0<5028> ssi:boot:base:linear: finished
lars_at_star29:~/lamtest>mpirun -np 2
/usr/local/bin/dacapo_2.7.3_lammpi.camp.run
------------------------------------------------------------------------
-----
It seems that [at least] one of the processes that was started with
mpirun did not invoke MPI_INIT before quitting (it is possible that
more than one process did not invoke MPI_INIT -- mpirun was only
notified of the first one, which was on node n0).
mpirun can *only* be used with MPI programs (i.e., programs that
invoke MPI_INIT and MPI_FINALIZE). You can use the "lamexec" program
to run non-MPI programs over the lambooted nodes.
------------------------------------------------------------------------
-----
I also tried using the lam-6.5.8-4 RPMS that came with the system, but
the same errors. Has anybody an idea what could possibly be wrong here?
Any help is highly appreciated!!
Lars
|
