High Performance Computing

Hi, I have a difficulty running mpirun on my workstation, RHEL 4.

I have uninstalled lam-7.1.4 with 'rpm -e lam-7.1.4'. No problem.
The I downloaded openmpi-1.3.tar.gz. unzipped as normal. No problem.
Then did 'make clean'. Then './configure'. No problems.
Then I did 'make' and then 'make install'. No problems.

I put the path '/usr/local/bin/' and '/usr/local/lib/'into my .bashrc file, and did source .bashrc --> no problem.

The I tried to 'lamboot', and this is where the problem begins. Here is the response (I have swapped my real name for 'username':

username:2DCHANNEL$ lamboot

LAM 7.1.4/MPI 2 C++/ROMIO - Indiana University

So apparently, the old lam-7.1.4 is still hanging around? Still I have tried running a parallel job 'test' with: mpirun -np 4 ./test, and the result is startling:

username:2DCHANNEL$ mpirun -np 4 ./test
ssh: x1-6-00-17-a4-99-51-97: Name or service not known
--------------------------------------------------------------------------
A daemon (pid 7989) died unexpectedly with status 255 while attempting
to launch so we are aborting.

There may be more information reported by the environment (see above).

This may be because the daemon was unable to find all the needed shared
libraries on the remote node. You may set your LD_LIBRARY_PATH to have the
location of the shared libraries on the remote nodes and this will
automatically be forwarded to the remote nodes.
--------------------------------------------------------------------------
--------------------------------------------------------------------------
mpirun noticed that the job aborted, but has no info as to the process
that caused that situation.
--------------------------------------------------------------------------
mpirun: clean termination accomplished

and when I check to see th date of lamboot with 'll /usr/local/bin/lam*' I get:

username:2DCHANNEL$ ll /usr/local/bin/lam*
-rwxr-xr-x 1 root root 194709 Jul 3 2008 /usr/local/bin/lamboot
-rwxr-xr-x 1 root root 162203 Jul 3 2008 /usr/local/bin/lamcheckpoint
-rwxr-xr-x 1 root root 95131 Jul 3 2008 /usr/local/bin/lamclean
-rwxr-xr-x 1 root root 284740 Jul 3 2008 /usr/local/bin/lamd
-rwxr-xr-x 1 root root 122140 Jul 3 2008 /usr/local/bin/lamexec
-rwxr-xr-x 1 root root 211554 Jul 3 2008 /usr/local/bin/lamgrow
-rwxr-xr-x 1 root root 88140 Jul 3 2008 /usr/local/bin/lamhalt
-rwxr-xr-x 1 root root 691644 Jul 3 2008 /usr/local/bin/laminfo
-rwxr-xr-x 1 root root 89662 Jul 3 2008 /usr/local/bin/lamnodes
-rwxr-xr-x 1 root root 162296 Jul 3 2008 /usr/local/bin/lamrestart
-rwxr-xr-x 1 root root 93683 Jul 3 2008 /usr/local/bin/lamshrink
-rwxr-xr-x 1 root root 93373 Jul 3 2008 /usr/local/bin/lamtrace
-rwxr-xr-x 1 root root 193821 Jul 3 2008 /usr/local/bin/lamwipe

Note that these are dated Jul 3 2008 when I first installed lammpi.

What also worries me is the line which says, "ssh: x1-6-00-17-a4-99-51-97: Name or service not known"

Now, ever since I first got my workstation, as I login with my username I get a message like 'x1-6-00-17-a4-99-51-97 not recognised. It may degrade GNOME perfromance. Try putting it in /etc/hosts'. However, it gives an option to ignore this message and continue, which I always do and I have never suffered anything at all, till now apparently after 2-3 years.

What is going on? How do I remove lam-7.1.4 competely from my system, and install the new mpirun suitable for openmpi? (I have installed opempi in order to use mpif90 -- this is ok now and my f90 programs compile ok with mpif90, it is just the running bit that is not working properly.)

How do I rectify the login warning message?

Please advise -- I am not a systems manager, but a newbie.

Cheers.

You don't run mpirun with lam and vice versa. They can, however, peacefully co-exist. Once you set up openmpi, you must set up a cluster hosts file or something. Then for each process, openmpi will attempt to ssh to a host and run your program.