Skip to content

Latest commit

 

History

History
89 lines (74 loc) · 2.69 KB

HPC_Matlab.md

File metadata and controls

89 lines (74 loc) · 2.69 KB
Raw

Running Matlab code at High Performance Computing (HPC) cluster at NYU.

Step-1: Use Git-Bash in Windows (or terminal in Linux) and login to hpc as follow

  • From NYU
$ ssh <your-net-ID>@prince.hpc.nyu.edu
<your-net-ID>@prince.hpc.nyu.edu's password:
  • Outside NYU
$ ssh <your-net-ID>@gw.hpc.nyu.edu
<your-net-ID>@prince.hpc.nyu.edu's password:

$ ssh prince.hpc.nyu.edu
<your-net-ID>@prince.hpc.nyu.edu's password:

(Optional): Check your Quota

$ myquota

Filesystem   Environment   Backed up?   Allocation       Current Usage
Space        Variable      /Flushed?    Space / Files    Space(%) / Files(%)

/scratch    $SCRATCH       No/Yes        5.0TB/1.0M       0.00GB(0.00%)/4(0.00%)
/beegfs     $BEEGFS        No/Yes        2.0TB/3.0M        0.00GB(0.00%)/0(0.00%)

Step-2: Use $SCRATCH folder to run your commands from. It offers huge space 4Tb/user but get deleted in 60 days, so move all the content to $ARCHIVE after running the code. More details here.

$ cd $SCRATCH
$ pwd

Step-3: Clone the code from my repository

$ git clone https://github.com/ishjain/mmWave.git
$ cd mmWave

(Optional): Edit the batch file. The content of mybatch.sbatch is pasted here.

#!/bin/bash
#SBATCH --nodes=1
#SBATCH --ntasks-per-node=1
#SBATCH --cpus-per-task=1
#SBATCH --time=1:00:00
#SBATCH --mem=2GB
#SBATCH --job-name=myTest
#SBATCH --mail-type=END
#SBATCH [email protected]
#SBATCH --output=ish_%j.out
  
module purge
module load matlab/2017b

RUNDIR=$SCRATCH/mmWave/
cd $RUNDIR

matlab -nodisplay -nodesktop -r "run SimulationLOS.m" //Change the filename here

(Optional): Check the modules. You will find matlab/2017b

$ module avail

Step-4: Submit the job

$ sbatch mybatch.sbatch
Submitted batch job 4605025

(Optional): Check the job queue

$ squeue -u <your-net-ID>
JOBID   PARTITION     NAME     USER ST       TIME  NODES NODELIST(REASON)
4605025    c32_38   myTest   ikj211  R       0:14      1 c32-01

Step-5: Check the results in local directory

$ ls -lastr

Step-6: When running the same code many times (atmost 400). In the main file SimulationLOS.m, we first get the environment variable aID = getenv('SLURM_ARRAY_TASK_ID'). The aID is different for different parallel computations. So, you can use these id to seed the random generator and use them to name the output file. Note: aID is an array, so first convert to int for seed.

sbatch --array=1-100 mybatch.sbatch

This above command runs the main code on 100 parallel machines with aID ranging from 1 to 100.