#!/bin/sh # Script for running a simulation job in the SLURM system # # Written by Shyam Kumar Sudhakar, Ivan Raikov, Tom Close, Rodrigo Publio, Daqing Guo, and Sungho Hong # Computational Neuroscience Unit, Okinawa Institute of Science and Technology, Japan # Supervisor: Erik De Schutter # # Correspondence: Sungho Hong (shhong@oist.jp) # # September 16, 2017 ## Some parameters for running a SLURM job #SBATCH --job-name=GL_SIM #SBATCH --partition=compute #SBATCH --mem-per-cpu=10G #SBATCH --time=120:30:00 #SBATCH --ntasks=120 #SBATCH --cpus-per-task=2 #SBATCH --input=none ## Standard output and standard error files #SBATCH --output=SHAREDDIR/simulation.out.log #SBATCH --error=SHAREDDIR/simulation.err.log export PATH=... # Set paths for python, etc. here NEURONHOME=... # Set your NEURONHOME here export PATH=$NEURONHOME/nrn/x86_64/bin:$NEURONHOME/iv/x86_64/bin:$PATH export LD_LIBRARY_PATH=$NEURONHOME/nrn/x86_64/lib:$NEURONHOME/iv/x86_64/lib:$LD_LIBRARY_PATH echo PYTHONPATH is $PYTHONPATH echo "==============Starting mpirun===============" cd SHAREDDIR/model mpirun nrniv -mpi -python main.py echo "==============Mpirun has ended===============" ## Copy all the output data mkdir $HOME/work/output.$JOB_ID cp -v *.dat $HOME/work/output.$JOB_ID cp -v *.bin $HOME/work/output.$JOB_ID cp -R $PARAMDIR $HOME/work/output.$JOB_ID