Slurm script generator for Nova

* indicates a required field

 
Parameters
* Compute node type:
compute for regular compute nodes
gpu for research GPU nodes (not class)
huge for large memory nodes
amd for AMD regular compute nodes
class for class nodes (GPU and regular)
* Number of compute nodes:
Use 1 unless running a distributed-memory parallel application
* Number of processor cores per node:
regular compute nodes have 36 cores,
community, amd, and huge nodes have 64 cores,
GPU nodes have up to 64 cores
Is this a GPU job?Yes:
* Walltime limit:
Maximum time the job may run. Shorter jobs may start running sooner.		 hours mins secs
Max memory per compute node:
Default memory is 5100 MB per processor core requested for the regular compute nodes (1 GB == 1024 MB)
One can request up to 369 GB per regular compute node if needed
Job name:
Default is the job script name. Displayed by the squeue command.
Receive email for job events:
BEGIN: when job starts, END: when job ends, FAIL: if job fails		 BEGIN END FAIL EMAIL
Job standard output file:
Default is "slurm-%j.out", where %j is the job ID.
Job standard error file:
Default is same file as the job standard output file.

Job Script

#!/bin/bash

# Copy/paste this job script into a text file and submit with the command:
#    sbatch thefilename
# job standard output will go to the file slurm-%j.out (where %j is the job ID)

#SBATCH --time=00:15:00   # walltime limit (HH:MM:SS)
#SBATCH --nodes=1   # number of nodes
#SBATCH --ntasks-per-node=36   # 36 processor core(s) per node 

# LOAD MODULES, INSERT CODE, AND RUN YOUR PROGRAMS HERE

 For additional information about the sbatch command, see:

Slurm Official Sbatch Command Documentation
ISU Sample Job Scripts