Skip to content

Module System

The module system is introduced to meet the different requirements set by the courses. This system enables users to choose from the various versions of software available(eg. Python 2, Python 3).

The module system automatically loads required libraries.

(eg. When the user issues module load cuDNN. The appropriate CUDA library will be loaded. This is due to cuDNN's dependency on CUDA)

By defult, Python 3 is automatically loaded by when you log into the cluster

Recommend build configs

Refer to the build configs for instructions on loading modules related to

  • Tensorflow
  • Pytorch

Available commands

List currently loaded modules

module list

List all available modules

module available

Search for a package by name (case-sensitive)

module avail <package name>

List all available versions of a module by name (case-sensitive)

module spider <module name>

Load a module by name (case-sensitive)

module load <module name>

Remove a module from your current working environment (case-sensitive)

module unload <module name>

Remove all modules from your current working environment

module purge

Advance usage

Save the list of modules currently loaded into a collection for later use (case-sensitive)

module save <collectionName>

Load the modules from a collection (case-sensitive)

module restore <collectionName>

List current collections

module savelist

List modules in a collection (case-sensitive)

module describe <collectionname>

Disable a collection (case-sensitive)

module disable <collectionname>