AMBER Archive (2009)

Subject: RE: [AMBER] mairun

From: Ross Walker (ross_at_rosswalker.co.uk)
Date: Thu Jan 08 2009 - 12:25:26 CST

Hi Wen

> I am testing parallel programs which have been installed on our linux
> cluster:
>
> ls -l /opt/amber10/bin/mpirun
> -rwxr-xr-x 1 root root 228654 2009-01-06 15:12 /opt/amber10/bin/mpirun
> ls -l /opt/amber10/exe/sander.MPI
> -rwxr-xr-x 1 root root 5500906 2009-01-06 15:22
> /opt/amber10/exe/sander.MPI
>
> Run test/cytosine> /opt/amber10/bin/mpirun -np 4
> /opt/amber10/exe/sander.MPI -O -i in.md -c crd.md.23 -o cytosine.out
>
> --> no lamd running on the host
>
> run /opt/amber10/bin/lamboot
>
> --> LAM 7.1.3 - Indiana University
>
> then run the test again, and got the same message "no lamd running on
> the host"

This suggests a problem with the configuration on your machine. What does
the 'run' command you list above actually do? It is running it on your local
machine yes?

I would try a few simple things to check things.

1) Check your path and make sure mpirun and lamboot are the correct ones (in
/opt/...) and not in /usr/bin etc.

You can use: which lamboot

to see what it returns.

If need be add: /opt/amber10/bin/ to the 'beginning' of your path in your
login files (such as .bashrc)

2) Check MPI_HOME points to /opt/amber10/

3) Run 'ps aux' and see if any copies of lamd or lamboot are running and
kill them if they are.

4) As a regular user (NOT ROOT since lamboot cannot be run as root) do the
following:

lamboot
mpirun -np 2 ls

You should get 2 copies of ls run which will return 2 directory listings. If
this works then you can try again running an amber simulation.

You could also see if lamboot has a verbose mode you can run it in -
something like lamboot -v (I don't have lamboot installed on any of my
machines to check unfortunately).

I suspect though that your problem lies in either the version of lamboot
that is running not matching the mpirun command (due to path issues) or the
correct lamboot running but it running a different version of lamd due to
path and MPI_HOME issues. Then when you run mpirun the lamd quits silently
and then you are presented with the lamd not running error.

Just a guess - but it should give you some things to try.

All the best
Ross

/\
\/
|\oss Walker

| Assistant Research Professor |
| San Diego Supercomputer Center |
| Tel: +1 858 822 0854 | EMail:- ross_at_rosswalker.co.uk |
| http://www.rosswalker.co.uk | PGP Key available on request |

Note: Electronic Mail is not secure, has no guarantee of delivery, may not
be read every day, and should not be used for urgent or sensitive issues.

_______________________________________________
AMBER mailing list
AMBER_at_ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber