Skip to content
This repository has been archived by the owner on Dec 5, 2021. It is now read-only.

A collection of useful scripts for executing code on the high-performance computing clusters of the Technical University of Denmark (DTU). The scripts cover the set-up of the environment and the job submission to the load handler.

Notifications You must be signed in to change notification settings

Algebrazebra/DTU-HPC-Scripts

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

7 Commits
 
 
 
 
 
 

Repository files navigation

DTU-HPC-Scripts

This repository collects useful scripts for executing code on the high-performance computing clusters of the Technical University of Denmark (DTU). Since the available modules are dependent on the login node, I have to add the disclaimer that I personally only used and tested them on the second login node: login2.hpc.dtu.dk. At the time of writing, the default system Python (version 2.7.5) was used.

Setting your Python environment with Tensorflow and Keras

First, connect with the DTU HPC servers using ssh. As stated, I recommend the second login node. After logging in, navigate to your desired project directory and execute the following to set up your Python environment including Tensorflow and Keras based on the CUDA 9.0 install. After exiting, you can, for example, login to an interactive GPU node to run your code.

linuxsh
wget https://github.com/Algebrazebra/DTU-HPC-Scripts/raw/master/setup.sh
sh setup.sh
rm -f setup.sh
exit

Submitting to the Job Queue

Jobs are submitted using a submission shell script via

bsub < submit.sh

The shell script contains the instructions for the load handler as well as necessary commands to execute your code. A sample job script is provided with the submit.sh file in this repo. Simply alter the file to your needs and liking. For more information on the job script please refer to the official documentation given here: Batch Jobs.

Upon successful submission you can check the status of your current submissions with

bstat

About

A collection of useful scripts for executing code on the high-performance computing clusters of the Technical University of Denmark (DTU). The scripts cover the set-up of the environment and the job submission to the load handler.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages