Scripts for QE: Difference between revisions

From Wiki
Jump to navigation Jump to search
Hnguyen (talk | contribs)
New page: Here is a script for submitting a QE job to CSUC collserola <pre> #!/bin/bash # specifies the name of your job # BSUB -J jobname # # send standard output here # BSUB -o quantum_espresso.l...
 
Hnguyen (talk | contribs)
No edit summary
Line 1: Line 1:
Here is a script for submitting a QE job to CSUC collserola
Here is a script for submitting a QE job to CSUC collserola, usually for small jobs, shorter waiting queue


<pre>
<pre>
Line 13: Line 13:
#
#
# pick a queue
# pick a queue
# BSUB -q parallel16
# BSUB -q parallel4
#
#
# pick the machine
# pick the machine
Line 19: Line 19:
#
#
# Indicate the number of cores
# Indicate the number of cores
# BSUB -n 16
# BSUB -n 4
# All process should be executed in the same node
# All process should be executed in the same node
# BSUB -R span[hosts=1]  
# BSUB -R span[hosts=1]  
Line 44: Line 44:




EXE=/prod/QuantumESPRESSO/5.4.0/bin/pw.x
# Change to the work directory and copy all the necessary files to this folder
cd $TMPDIR
cp $INDIR/$INPUT $INDIR/*.UPF $TMPDIR
# Run the job
mpijob $EXE < $INPUT > $OUTPUT
# Create a new directory in the $SCRATCH directory, each new directory for each job specifying the JobID, and copy the output to this directory
mkdir -p $SCRATCH/output_$LSB_JOBID
cp -r * $SCRATCH/output_$LSB_JOBID
# copy in HOME and delete in scratch
mkdir -p $HOME/output_$LSB_JOBID
cp -r * $HOME/output_$INDIR_$LSB_JOBID
rm -r $SCRATCH/output_$LSB_JOBID
date
</pre>
To submit to pirineus, usually for very big jobs that requires a lot of nodes and therefore longer waiting queue
<pre>
#!/bin/ksh
# specifies the name of your job
# BSUB -J jobname
#
# send standard output here
# BSUB -o quantum_espresso.log 
#
# send standard error here
# BSUB -e quantum_espresso.err
#
# pick a queue
# BSUB -q parallel16
#
# pick the machine
# BSUB -R pirineus
#
# Indicate the number of cores
# BSUB -n 16
# All process should be executed in the same node
# BSUB -R span[hosts=1]
#
# Send an email notice once the job is finished
# BSUB -N -u MYEMAIL@iciq.es
INDIR=$HOME/Co3O4/GEOM_OPT # directory with your input and the .UPF files
INPUT=INPUT.in
OUTPUT=OUTPUT.out
date
# Set up the environment
. /opt/modules/default/init/bash
module load quantumespresso/5.4.0
unset OMP_NUM_THREADS
pseudo_dir= $INDIR/
export ESPRESSO_PSEUDO=$INDIR/
EXE=/prod/QuantumESPRESSO/5.4.0/bin/pw.x
EXE=/prod/QuantumESPRESSO/5.4.0/bin/pw.x



Revision as of 15:39, 29 March 2017

Here is a script for submitting a QE job to CSUC collserola, usually for small jobs, shorter waiting queue

#!/bin/bash
# specifies the name of your job
# BSUB -J jobname
#
# send standard output here
# BSUB -o quantum_espresso.log  
#
# send standard error here
# BSUB -e quantum_espresso.err 
#
# pick a queue
# BSUB -q parallel4
#
# pick the machine
# BSUB -R collserola
#
# Indicate the number of cores
# BSUB -n 4
# All process should be executed in the same node
# BSUB -R span[hosts=1] 
#
# Send an email notice once the job is finished
# BSUB -N -u MYEMAIL@iciq.es

export ESPRESSO_PSEUDO=$HOME/espresso_pseudo
export ESPRESSO_TMPDIR=$HOME/espresso_tmp


INDIR=$ESPRESSO_TMPDIR # directory with your input and the .UPF files
INPUT=INPUT.in
OUTPUT=OUTPUT.out

date

# Set up the environment
. /opt/modules/default/init/bash
module load quantumespresso/5.4.0

unset OMP_NUM_THREADS



EXE=/prod/QuantumESPRESSO/5.4.0/bin/pw.x

# Change to the work directory and copy all the necessary files to this folder

cd $TMPDIR
cp $INDIR/$INPUT $INDIR/*.UPF $TMPDIR
	
# Run the job

mpijob $EXE < $INPUT > $OUTPUT

# Create a new directory in the $SCRATCH directory, each new directory for each job specifying the JobID, and copy the output to this directory

mkdir -p $SCRATCH/output_$LSB_JOBID
cp -r * $SCRATCH/output_$LSB_JOBID

# copy in HOME and delete in scratch
mkdir -p $HOME/output_$LSB_JOBID
cp -r * $HOME/output_$INDIR_$LSB_JOBID

rm -r $SCRATCH/output_$LSB_JOBID
date

To submit to pirineus, usually for very big jobs that requires a lot of nodes and therefore longer waiting queue

#!/bin/ksh
# specifies the name of your job
# BSUB -J jobname
#
# send standard output here
# BSUB -o quantum_espresso.log  
#
# send standard error here
# BSUB -e quantum_espresso.err 
#
# pick a queue
# BSUB -q parallel16 
#
# pick the machine
# BSUB -R pirineus
#
# Indicate the number of cores
# BSUB -n 16
# All process should be executed in the same node
# BSUB -R span[hosts=1] 
#
# Send an email notice once the job is finished
# BSUB -N -u MYEMAIL@iciq.es

INDIR=$HOME/Co3O4/GEOM_OPT # directory with your input and the .UPF files
INPUT=INPUT.in
OUTPUT=OUTPUT.out

date

# Set up the environment
. /opt/modules/default/init/bash
module load quantumespresso/5.4.0

unset OMP_NUM_THREADS
pseudo_dir= $INDIR/
export ESPRESSO_PSEUDO=$INDIR/
EXE=/prod/QuantumESPRESSO/5.4.0/bin/pw.x

# Change to the work directory and copy all the necessary files to this folder

cd $TMPDIR
cp $INDIR/$INPUT $INDIR/*.UPF $TMPDIR
	
# Run the job

mpijob $EXE < $INPUT > $OUTPUT

# Create a new directory in the $SCRATCH directory, each new directory for each job specifying the JobID, and copy the output to this directory

mkdir -p $SCRATCH/output_$LSB_JOBID
cp -r * $SCRATCH/output_$LSB_JOBID

# copy in HOME and delete in scratch
mkdir -p $HOME/output_$LSB_JOBID
cp -r * $HOME/output_$INDIR_$LSB_JOBID

rm -r $SCRATCH/output_$LSB_JOBID
date