General Script Examples

About

These examples demonstrate how to run jobs and request resources on HPC using the job scheduler SLURM. Each example comes with downloadable scripts that can be run with minor modifications.

Scripts

Serial Job

A basic example script that requests a single CPU on one node.

High Memory Node Job

A basic example demonstrating how to request a high memory node on either Ocelote or Puma.

Using a Pipefail

This script uses a pipefail to kill a SLURM job in the event of a failure at any point in the pipeline.

Disabling Core Dumps

Sometimes when applications crash, they will save a snapshot of the working memory at the time of the crash to a core dump file. Core dump files can be handy for debugging, but they can also be quite large. In some cases, it may be preferable to disable core dumps to prevent your storage allocations from filling up. This example shows how to disable core dumps.

Retrieving Scratch Files

Example of using a compute node’s local storage, or scratch space, during a job’s execution and retrieving the output before the job’s walltime limit is reached.

Using Job Dependencies

Sometimes projects need to be split up into multiple parts where each step is dependent on the step (or several steps) that came before. SLURM dependencies are a way to automate this process.

---