Homec4science
Diffusion SCITAS examples on how to run on the clusters (master) Browse

scitas-examples/Advanced/OccupyOneNode/occupy_nodes_with_serial_tasks.sh
824d2e1a22d1master

Last Change
Raw File

occupy_nodes_with_serial_tasks.sh
View Options

#!/bin/bash -l
#SBATCH --nodes 1
#SBATCH --cpus-per-task 1
#SBATCH --mem-per-cpu 4G
# README
# ===================================================================
# In this example an imaginary chemistry code is executed. It will
# process 300 files: myinput.001 to myinput.300, and will produce the
# corresponding output files. The output (stdout) of the command is
# captured into a corresponding file under the submission directory.
#
# Slurm parameters:
# * --nodes 1
# as srun is used to launch the tasks this can be any number
# * --mem-per-cpu 4G
# this needs to be set for the '--exclusive' parameter of srun to
# work as expected. It should be adapted to each use case.
# * --cpus-per-task 1
# this can be adjusted to run (non-MPI) multithreaded code
#
# Note: This example differs from the accompanying ones in particular
# because xargs is used to manage the "queue" of tasks to execute. As
# such there is no need to 'wait' for any task as they are never sent
# into the background.
# ===================================================================
echo started at `date`
module purge
module load intel intel-mpi
module load chem2020
# calculate how many tasks can we run simultaneously
# (i.e. how many cores are available)
NR_TASKS=$(echo $SLURM_TASKS_PER_NODE | sed 's/\([0-9]\+\)(x\([0-9]\+\))/\1*\2/' | bc)
# execute the 300 tasks, NR_TASKS tasks at a time, redirecting the
# output to a task-specific output file
cat <(seq -w 1 300) | \
xargs -I{} -max-procs=$NR_TASKS \
bash -c "srun -N 1 -n 1 --exclusive chem2020.x -in myinput.{} -out myout.{} > ${SLURM_SUBMIT_DIR}/slurm-${SLURM_JOBID}.{}.out 2>&1"
echo finished at `date`
c4science · Help