What happens when you try to mpirun an MPI application that was compiled with LAM's mpicc?
-----Original Message-----
From: K. Charoenpornwattana Ter [mailto:kcharoen_at_[hidden]]
Sent: Tuesday, May 22, 2007 09:11 PM Eastern Standard Time
To: lam_at_[hidden]
Subject: LAM: lamboot is ok, mpirun is not
Hi all,
I have some problems with lam/mpi. I have been searching around the net but
noone has same problem as me.
My cluster has 1 head node and 14 compute nodes. I installed centos 4.5-i386.
I used OSCAR 4.2.1 to help building this cluster. I completely uninstalled
lam/mpi that came with OSCAR 4.2 and installed lam/mpi 7.1.3 with blcr 0.5.1
.
The problem is I can successfully lamboot hosts, but can't execute mpi
application (even simple hello world) on multiple nodes. (I can lamboot on
single node and execute "mpirun -np 1 hello.out")
I can ping, tping, traceroute from head to every nodes and vice versa in the
cluster. I can execute any mpi applications on this cluster using MPICH.
[ter_at_uftoscar ~]$ which mpirun
/opt/lam-7.1.3/bin/mpirun
[ter_at_uftoscar ~]$ ssh oscarnode1 which mpirun
/opt/lam-7.1.3/bin/mpirun
[ter_at_uftoscar ~]$ echo $PATH
/opt/lam-7.1.3
/bin/:/usr/kerberos/sbin:/usr/kerberos/bin:/usr/local/sbin:/usr/local/bin:/sbin:/bin:/usr/sbin:/usr/bin:/usr/X11R6/bin:/opt/pbs/bin:/opt/pbs/lib/xpbs/bin:/opt/kernel_picker/bin:/opt/env-switcher/bin:/opt/pvm3/lib:/opt/pvm3/lib/LINUX:/opt/pvm3/bin/LINUX:/usr/local/apitest:/opt/c3-4/:/root/bin:/opt/lam-
7.1.3/bin/
[ter_at_uftoscar ~]$ ssh oscarnode1 echo $PATH
/opt/lam-7.1.3
/bin/:/usr/kerberos/sbin:/usr/kerberos/bin:/usr/local/sbin:/usr/local/bin:/sbin:/bin:/usr/sbin:/usr/bin:/usr/X11R6/bin:/opt/pbs/bin:/opt/pbs/lib/xpbs/bin:/opt/kernel_picker/bin:/opt/env-switcher/bin:/opt/pvm3/lib:/opt/pvm3/lib/LINUX:/opt/pvm3/bin/LINUX:/usr/local/apitest:/opt/c3-4/:/root/bin:/opt/lam-
7.1.3/bin/
I am sure that the older version of lam/mpi was completely removed. and I
set env switcher to none.
Any help would be greatly apprecated.
Thanks
Ter
|