HPC cluster with exceed on demand visualization tool

Hi all,

I am working with fitrgp function in matlab and doing some computations for large datasets. The computation time is really too much(parfor doesnot seems to be helping much). So i have installed HPC cluster with exceed on demand as visualization tool. But the question is how to do job submission in HPC cluster for the scripts written in matlab, in order to reduce the computational time? It would be great, if you guys could give some suggestions.

Thanks a lot in advance!!!!

Answer 1

[Sahithi Kanumarlapudi]

If your matlab code is in the file ‘myScript.m’, you might use a job submissions script like:

#!/bin/bash 
 
#SBATCH -t 2:00 
 
#SBATCH -N 1 
 
#SBATCH -n 12 
 
#SBATCH -mem-per-cpu 1024 
 
#SBATCH -L matlab 
 
 
. ~/.profile 
 
module load matlab 
 
 
matlab -nodisplay -nosplash -r "run('myScript.m'); exit" > myScript.out

Answer 2

[Saurabh Chaudhary]

you can make the slurm file like this, and save the file xyx_slurm.sl and in order to run it enter sbatch xyz_slurm.sl in terminal

#!/bin/bash

#SBATCH --job-name=matlab_test

#SBATCH --time=0-03:00 # adjust this to match the walltime of your job

#SBATCH --nodes=1

#SBATCH --ntasks=1

#SBATCH --cpus-per-task=1 # adjust this if you are using parallel commands

#SBATCH --mem=4000 # adjust this according to the memory requirement per node you need

#SBATCH --partition=small

# Choose a version of MATLAB by loading a module:

module load matlab/R2022a

matlab -nodisplay -nosplash -softwareopengl < /pathofyourdirectory/cosplot.m