Skip to content

Modules Environment

NERSC uses the module utility to manage nearly all software. There are two advantages of the module approach:

  1. NERSC can provide many different versions and/or installations of a single software package on a given machine, including a default version as well as several older and newer version.
  2. Users can easily switch to different versions or installations without having to explicitly specify different paths. With modules, the MANPATH and related environment variables are automatically managed.

What is module

module is a shell function that modifies user shell upon load of a modulefile. The module function is defined as follows

$ type module
module is a function
module () 
{ 
    eval `/opt/cray/pe/modules/3.2.11.4/bin/modulecmd bash $*`
}

Note

module is not a program

Module Commands

Note

Cray default module is modules/3.2.11.4 and documentation is based on this version

modules/4.1.3.1 is also available to support non-Cray modulefiles that require Modules 4.x. While most modulefiles at NERSC should work with Modules 4.x, there are some compatibility issue with PrgEnv- modules, and we don't recommend modules/4.1.3.1 for "normal" use.

General usage:

nersc$ module [ switches ] [ subcommand ] [subcommand-args ]

Further reading:

  • module help
  • man module
  • man modulefile
  • Online manual (note: some features may only be available in later versions than what is installed on NERSC systems)

Module Usage

List currently loaded modules:

module list

List all available modules:

module avail 
module av

Show availability of specific module:

module avail <module-name>

Show availability of all modules containing a substring:

module avail -S <substring>

Display what changes are made when a module is loaded:

module display <module-name>
module show <module-name>

Add a module to your current environment:

module load <module-name>
module add <module-name>

Note

This command is silent unless there are problems with the module.

Tip

If you load then generic name of a module, you will get the default version.

module load gcc

To load a specific version use the full name

module load gcc/8.1.0

Remove module from the current environment:

module unload <module-name>
module rm <module-name>

Note

This command will fail silently if the specified module is not loaded.

Switch currently loaded module with a new module:

module swap <old-module> <new-module>
module switch <old-module> <new-module>

To purge all modules::

Note

this will remove all your modules from active environment including startup modules loaded in your shell. To restore your environment with startup modules, its best to relogin

module purge

To view help for a particular module::

module help <module-name>

To see a condensed list of module you can use module -t and use this with list or avail

$ module -t list
Currently Loaded Modulefiles:
modules/3.2.11.4
nsg/1.2.0
altd/2.0
darshan/3.1.7
intel/19.0.3.199
craype-network-aries
craype/2.6.2
cray-libsci/19.06.1
udreg/2.3.2-7.0.1.1_3.29__g8175d3d.ari
ugni/6.0.14.0-7.0.1.1_7.32__ge78e5b0.ari
pmi/5.0.14
dmapp/7.1.1-7.0.1.1_4.43__g38cf134.ari
gni-headers/5.0.12.0-7.0.1.1_6.27__g3b1768f.ari
xpmem/2.2.20-7.0.1.1_4.8__g0475745.ari
job/2.2.4-7.0.1.1_3.34__g36b56f4.ari
dvs/2.12_2.2.156-7.0.1.1_8.6__g5aab709e
alps/6.6.57-7.0.1.1_5.10__g1b735148.ari
rca/2.2.20-7.0.1.1_4.42__g8e3fb5b.ari
atp/2.1.3
PrgEnv-intel/6.0.5
craype-haswell
cray-mpich/7.7.10
craype-hugepages2M

At NERSC three Cray PrgEnv- modules are available: PrgEnv-cray, PrgEnv-gnu and PrgEnv-intel (the default). These modules provide respectively the Cray, GNU or Intel compiler via the compiler wrappers cc (C) CC (C++) and ftn (Fortran). You can only have one PrgEnv module loaded at a time: for instance if we load PrgEnv-gnu without first unloading PrgEnv-intel we get the following message.

    $ module load PrgEnv-gnu
    PrgEnv-gnu/6.0.5(77):ERROR:150: Module 'PrgEnv-gnu/6.0.5' conflicts with the currently loaded module(s) 'PrgEnv-intel/6.0.5'
    PrgEnv-gnu/6.0.5(77):ERROR:102: Tcl command execution failed: conflict PrgEnv-intel

To circumvent this you must swap or unload modules as follows

    # swap modules
    module swap PrgEnv-intel PrgEnv-gnu

    # unload + load
    module unload PrgEnv-intel
    module load PrgEnv-gnu

Creating a Custom Module Environment

You can modify your environment so that certain modules are loaded whenever you log in.

The first option is to use shell commands.

bash

In ~/.bash_profile

module swap PrgEnv-${PE_ENV,,} PrgEnv-gnu

csh

In ~/.login

set pe = ` echo $PE_ENV | tr "[:upper:]" "[:lower:]" `
module swap PrgEnv-${pe} PrgEnv-gnu

snapshots

The second option is to use the "snapshot" feature of modules.

  1. swap and load modules to your desired configuration
  2. save a "snapshot" with module snapshot <snapshot-filename>

Then at any time later restore the environment with module restore <snapshot-filename>.

Install Your Own Customized Modules

You can create and install your own modules for your convenience or for sharing software among collaborators. See the man modulefile or the modulefile documentation for details of the required format and available commands. These custom modulefiles can be made visible to the module command by module use /path/to/the/custom/modulefiles.

Tip

Global Common is the recommended location to install software.

Note

Make sure the UNIX file permissions grant access to all users who want to use the software.

Warning

Do not give write permissions to your home directory to anyone else.

Note

The module use command adds new directories before other module search paths (defined as $MODULEPATH), so modules defined in a custom directory will have precedence if there are other modules with the same name in the module search paths. If you prefer to have the new directory added at the end of $MODULEPATH, use module use -a instead of module use.

Known issues with modules

Zero exit code for invalid modules

$ module load X
ModuleCmd_Load.c(244):ERROR:105: Unable to locate a modulefile for 'X'
$ echo $?
0

This means that module commands often return a "success" code (0) even if the command failed, which can lead to surprising errors in eg job scripts.

Incompatibilities with modules/4.1.3.1

modules/4.1.3.1 is available but not recommended. If you do need Modules 4.x, you can access it with module swap modules modules/4.1.3.1 (note that module load modules/4.1.3.1 will abort with an error). If using Modules 4.x, we recommend carefully checking that your script or usage has the correct outcome.

Module FAQ

  1. Is there an environment variable that captures loaded modules?

Yes, active modules can be retrieved via $LOADEDMODULES, this environment variable is automatically changed to reflect active loaded modules that is reflected via module list. If you want to access modulefile path for loaded modules you can retrieve via $_LM_FILES

  1. How to restore MODULEPATH in user session?

If you run into an error such as following::

$ module avail
ModuleCmd_Avail.c(217):ERROR:107: 'MODULEPATH' not set

You should try a new login shell and see if it fixes the issue. Check to see if your startup scripts (~/.bashrc, ~/.bash_profile) or ~/.cshrc for tcsh/csh