Quantum Espresso batch file.

RE: Quantum Espresso batch file.

2014-02-05T05:01:17Z

This will happen if your job script is loading the incorrect mpi modules and/or you're using the wrong mpi launcher. For example if you built QE with mvapich2_ib but load openmpi_ib and use mpirun. you will see a bazillion copies of the program running independently instead of having them all run as a single MPI instance.

You may want to use something like "vimdiff" to compare the job script you used to submit your first (successful) run to your second (unsuccessful) restart. Make sure the modules being loaded/unloaded/swapped are identical, and make sure you're using the same mpi launcher (mpirun_rsh/mpirun/mpiexec.hydra).

If nothing obvious jumps out, feel free to send your job script to help@xsede.org and we can take a closer look.

Glenn

RE: Quantum Espresso batch file.

2013-12-19T22:37:57Z

Hey Junyue,

First, I'd try and simplify things a bit. It seems like you're trying to do a lot of different things in your PBS script. I typically do place all of my data in /scratch rather than /home to run simulations. Just simplifies moving things around with the script(s).

So, try and re-write a much simpler PBS script to merely run the job, then add to it as you get more comfortable with the machine environment. Look at examples in the Gordon User Guide: https://portal.xsede.org/web/xup/sdsc-gordon

The error about not finding the .in file can likely be remedied by specifying the full path to the file rather than just the name itself. The problem with "==" I am not sure....maybe something that PBS does not support? Just a guess, but again...simplifying the script to do the bare minimum will help out in the beginning.

Quantum Espresso batch file.

2013-12-17T02:53:26Z

Dear users:

I am now using Gordon to do my calculations. I successfully compiled QE on it and could run jobs with command "mpirun PATH/pw.x -np 1 -in file.in > file.out".

However, I cannot run the jobs through batch script and I do not know why. Could anyone please help me out here?

The error is:
error while loading shared libraries: libfftw3.so.3: cannot open shared object file: No such file or directory

I tried to locate libfftw3.so.3 but could not find anything. And other libfftw3 files are in MVAPICH2 library which I did not use at all.

My batch file:

#!/bin/bash
#PBS -q normal
#PBS -l nodes=1:ppn=2:native
#PBS -l walltime=48:00:00
#PBS -N TEST
#PBS -o qe.mpi.log
#PBS -e qe.mpi.err
#PBS -A XXXXX
#PBS -M XXX@XXX.XXX
#PBS -m abe
#PBS -V
######################################################

module load mpich2_ib

/bin/mkdir -p /oasis/scratch/$USER/temp_project/${PBS_O_WORKDIR##*$USER/}
/bin/cp -prL $PBS_O_WORKDIR/* /oasis/scratch/$USER/temp_project/${PBS_O_WORKDIR##*$USER/}/.
cd /oasis/scratch/$USER/temp_project/${PBS_O_WORKDIR##*$USER/}

mpirun PATH/pw.x -np 1 -in file.in > file.out