Comsol

Comsol is only available to licensed groups through the shared research license pool - see https://me.ucsb.edu/comsol for more information.

To use Comsol, save your file as a '.mph' file on the computer you regularly use Comsol from. Transfer this file to Knot7 or Pod.

You will need a .job file to run your job in the queue, an example one is at the end of this page - note that you really only need to adjust the 'ppn' (processors/node) and the job name lines.

Once you set the filenames and 'ppn', you submit with a

sbatch myfile.job (assuming you've named the file below myfile.job). To run on a regular knot7 node (48GB of RAM and up to 12 cores) use

sbatch myfile.job

To run on the 'fat' (high memory, up to 1TB) nodes (where you can set ppn as high as 32) use

sbatch -p largemem myfile.job

In this example, progress info goes to batch.log and when finished your output is in the file myfile-out.mph that you can transfer back to your desktop to look at.

#!/bin/bash  -l
#SBATCH --nodes=1 --ntasks-per-node 4
### Leave 'nodes' at 1
### change 'tasks-per-node' up to the number of cores you want
#SBATCH --time=24:00:00
### you can leave the time line out 

module load comsol/56
cd $SLURM_SUBMIT_DIR
/bin/hostname

FILENAME="curved3d"
### Change line above to prefix of your .mph file
###    in this example, it is curved3d.mph

comsol -np $SLURM_NTASKS  batch -inputfile ./${FILENAME}.mph -outputfile ./${FILENAME}-out.mph -batchlog ${FILENAME}.log
### don't need to mess with this line