Here is my job script. I can that two nodes cn1 and cn2 are assigned and
the job is running but I cannot see any output. What is the correct way of
running multi GPU job (using multiple nodes)? FYI- I have 2 GPUs and 4 GPUs
per node. Any feedback is appreciated.
#!/bin/bash
#PBS -N di_sho
#PBS -V
#PBS -j oe
#PBS -l walltime=100:00:00
#PBS -l nodes=2:ppn=2:gpus=2
export AMBERHOME=/usr/local/amber12
export CUDA_HOME=/usr/local/cuda-5.0
export
LD_LIBRARY_PATH=/usr/local/cuda-5.0/lib:/usr/local/cuda-5.0/lib64:/usr/local/libexec/mvapich2/:$LD_LIBRARY_PATH
export PATH=/usr/local/bin:$PATH
cd $PBS_O_WORKDIR
cat $PBS_NODEFILE
mpirun -machinefile $PBS_NODEFILE -np 4 $AMBERHOME/bin/pmemd.cuda.MPI -O -i
sander.in3 -o prod1.out2 -r prod1.rst -p solvated.prmtop -c density.rst -x
prod1.mdcrd
--
Regards,
Dr. Neha S. Gandhi,
Curtin Research Fellow,
School of Biomedical Sciences,
Curtin University,
Perth GPO U1987
Australia
LinkedIn
Research Gate
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Mar 26 2014 - 07:30:02 PDT
