[openib-general] Problems running MPI jobs with MVAPICH

[Don.Albert at Bull.com]

Weikuan,

> Could you try to run MVAPICH (1 or 2) with tcp option first, or use
> mpirun_rsh initially? Just to be sure you can run any MPI to start with.
> Please let us know of the problems too, along with the same set of 
traces
> and machine specification.
> 
With no changes to the software, I tried "mpirun_rsh".  If I use a simple 
command like "uname" it works:

[koa] (ib) ib> mpirun_rsh -np 2 koa jatoba uname -a
Linux jatoba.az05.bull.com 2.6.9-34.ELsmp #1 SMP Fri Feb 24 16:56:28 EST 
2006 x86_64 x86_64 x86_64 GNU/Linux
Linux koa.az05.bull.com 2.6.9-34.ELsmp #1 SMP Fri Feb 24 16:56:28 EST 2006 
x86_64 x86_64 x86_64 GNU/Linux

but if I try to run the MPI job, it fails with a segmentation fault on 
both sides:

[koa] (ib) ib> mpirun_rsh -np 2 koa jatoba /home/ib/mpi/tests/cpi/cpi
bash: line 1: 16302 Segmentation fault      /usr/bin/env MPIRUN_MPD=0 
MPIRUN_HOST=koa.az05.bull.com MPIRUN_PORT=32945 MPIRUN_PROCESS
ES='koa:jatoba:' MPIRUN_RANK=1 MPIRUN_NPROCS=2 MPIRUN_ID=19842 
DISPLAY=dalbertpc.usnetwork.lan:0 /home/ib/mpi/tests/cpi/cpi
bash: line 1: 19866 Segmentation fault      /usr/bin/env MPIRUN_MPD=0 
MPIRUN_HOST=koa.az05.bull.com MPIRUN_PORT=32945 MPIRUN_PROCESS
ES='koa:jatoba:' MPIRUN_RANK=0 MPIRUN_NPROCS=2 MPIRUN_ID=19842 
DISPLAY=dalbertpc.usnetwork.lan:0 /home/ib/mpi/tests/cpi/cpi

The MVAPICH User Guide document talks about a script "make.mvapich.tcp" to 
build for running over IPoIB, but the script is not in the "mvapich-gen2" 
directory.  There is a script "make.mvapich2.tcp" under the 
"mvapich2-gen2" directory.  I will try building MVAPICH2 for TCP and see 
if that works.

        -Don Albert-
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openfabrics.org/pipermail/general/attachments/20060329/694d1c03/attachment.html>