Hi,
I am new to amber. I recently installed amber12 in mac OS X 10.5
server based on the input I got from amber mailing list. I also tested both
serial and parallel runs. Both methods worked. For parallel run I gave
below command. But it only runs on the master node. (The job I am trying to
run is on a mac cluster of 15 nodes, with 8 cores for each node).
mpirun -np 8 $AMBERHOME/bin/sander.MPI -O -i min.in -o polyAT_wat1.out -c
polyAT_wat.inpcrd -p polyAT_wat.prmtop -r polyAT_wat1.rst
Then I used the SGE scripts to submit the parallel job (the SGE script is
given below). But my job is not running on any of the node not even in the
master node. I also provided the error output below the SGE script.
#!/bin/bash
## ------ EMBEDDED SGE ARGUMENTS ----
#$ -N amber_job
#$ -pe mpich 1
#$ -cwd
#$ -S /bin/bash
## ------------------------------------
# export all environment variables to SGE
#$ -V
MPIRUN=$AMBERHOME/bin/mpirun
sander=$AMBERHOME/bin/sander.MPI
export RSHCOMMAND=/usr/bin/ssh
echo " i got $NSLOTS slots to run on!"
$MPIRUN -np $NSLOTS -machinefile $TMDIR/mpd.hosts $sander -O -i min.in -o
polyAT_wat2.out -c polyAT_wat.incprd -p polyAT_wat.prmtop -r polyAT_wat2.rst
*Error output*
dyld: Library not loaded: /Volumes/data/common/amber12/lib/libmpl.1.dylib
Referenced from: /common/amber12/bin/mpirun
Reason: image not found
/common/sge/default/spool/node006/job_scripts/676: line 17: 620 Trace/BPT
trap
(core dumped) $MPIRUN -np $NSLOTS -machinefile $TMDIR/mpd.hosts
$sande
r -O -i min.in -o polyAT_wat2.out -c polyAT_wat.incprd -p polyAT_wat.prmtop
-r p
olyAT_wat2.rst
I have also added below lines to my .bash_profile file.
export AMBERHOME=/common/amber12
export PATH="$AMBERHOME/bin:$PATH"
export MPI_HOME=$AMBERHOME
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH\:$AMBERHOME/lib
I am not able to figure out what went wrong.
Thanks in advance
--
M. Prabu
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Mon Aug 06 2012 - 10:30:04 PDT
