Gaussian Job Script
Here is a simple job script g16.sb
for running Gaussian job
g16.com
:
g16.sb
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 |
|
where the Gaussian input file g16.com
can be found from the previous
section Running Gaussian by Command Lines.
For the resource request (#SBATCH
lines) above, since Gaussian
can only run in parallel with
shared memory
in HPCC system, only 1 task (with 1
node) is requested in line 2. The number of CPUs requested in line 3 is
the same as the setting of "%NProcShared
" (=4
) in the Gaussian input
file. The memory request in line 4 should be larger than the setting of
"%Mem
" (=5GB
) in the Gaussian input file in case the job runs out of
memory. Please also make sure the walltime request in line 5 is longer
enough to finish the job.
In the command line, you need to make sure Gaussian/g16 is loaded
as in line 10. If you would like to use scratch directory other than
/mnt/scratch/$USER
for the Gaussian scratch files, you could set up a
different one with line 11 and 12. The calculation of the Gaussian job
is executed in line 14 with input file g16.com
and output file
g16.log
. Once the calculation is done, line 17 and 20 will be executed
to print out the job information and resource usage respectively to the
SLURM output file ( with file name: slurm-<JobID>.out
).