Managing Software Modules on Clipper

Tags Clipper

Clipper uses the Modules package to present software as modules to cluster users.

Modules allow software to be loaded/unloaded as necessary. The Modules package manages dependencies between modules and can set/unset environment variables so that applications run in the correct environment.

This document describes basic Modules usage.

Modules usage

Listing Available Modules

Use the module avail command to list available modules. This will display a list of all available modules.

[hpcuser1@clipper ~]$ module avail
------------------------------------------------ /cm/local/modulefiles -------------------------------------------------
boost/1.77.0              cm-setup/9.2       freeipmi/1.6.8   module-git                   python39
cluster-tools-dell/9.2    cmd                gcc/11.2.0       module-info                  R/4.3.2
cluster-tools/9.2         cmjob              ipmitool/1.8.18  null                         shared
cm-bios-tools             cmsh               lua/5.4.4        openldap                     slurm/slurm/21.08.8
cm-image/9.2              cuda-dcgm/3.1.3.1  luajit           openmpi/mlnx/gcc/64/4.1.5a1
cm-scale/cm-scale.module  dot                mariadb-libs     python3

------------------------------------------------ /cm/shared/modulefiles ------------------------------------------------
aws-cli/2.13.32                  gcc9/9.5.0                       lapack/gcc/64/3.10.0        ucx/1.10.1
blacs/openmpi/gcc/64/1.1patch03  gcc10/10.4.0                     mpich/ge/gcc/64/3.4.2
blas/gcc/64/3.10.0               gcc11/11.3.0                     mvapich2/gcc/64/2.3.7
bonnie++/2.00a                   gcc12/12.3.0                     netcdf/gcc/64/gcc/64/4.8.1
cm-pmix3/3.1.4                   gcc13/13.2.0                     netperf/2.7.0
cm-pmix4/4.1.1                   gdb/10.2                         nvhpc-byo-compiler/23.3
cuda12.0/blas/12.0.1             globalarrays/openmpi/gcc/64/5.8  nvhpc-hpcx/23.3
cuda12.0/fft/12.0.1              hdf5/1.12.1                      nvhpc-nompi/23.3
cuda12.0/nsight/12.0.1           hdf5_18/1.8.21                   nvhpc/23.3
cuda12.0/profiler/12.0.1         hwloc/1.11.11                    openblas/dynamic/(default)
cuda12.0/toolkit/12.0.1          hwloc2/2.7.1                     openblas/dynamic/0.3.18
default-environment              intel-tbb-oss/ia32/2021.4.0      openmpi/gcc/64/4.1.2
fftw3/openmpi/gcc/64/3.3.10      intel-tbb-oss/intel64/2021.4.0   openmpi4/gcc/4.1.2
gcc8/8.5.0                       iozone/3_492                     py-tensorflow/2.14.0_py39

Listing Loaded Modules

Use the module list command to list loaded modules. This will display a list of all loaded modules.

[hpcuser1@clipper ~]$ module list
Currently Loaded Modulefiles:
 1) dot   2) R/4.3.2   3) aws-cli/2.13.32   4) default-environment   5) gcc/11.2.0   6) slurm/slurm/21.08.8

Loading Modules

Use the module load command followed by the name of the module to load the module. For example, to load the GNU Compiler Collection (GCC) version 12 module, run the following command:

[hpcuser1@clipper ~]$ module load gcc12

Unloading Modules

Use the module unload command followed by the name of the module to unload the module. For example, to unload the GCC version 12 module, run the following command:

[hpcuser1@clipper ~]$ module unload gcc12

Purging All Loaded Modules

Use the module purge command to unload all modules, including sticky modules. This will completely reset your module environment.

[hpcuser1@clipper ~]$ module purge

Default Modules

Clipper loads a minimal number of modules by default. As of November 7, 2023, the default environment includes:

  • The Slurm job scheduler
  • Version 11.2.0 of the GCC compiler
  • The R programming language
  • The AWS CLI for interacting with Amazon Web Services
  • Dot, a special module which appends the current directory to your PATH environment variable.

Python is available through the default system PATH and is not part of the modules system currently.

Additional Resources

For more information on using Modules, please refer to the official documentation: https://modules.readthedocs.io/

Was this helpful?
0 reviews
Print Article

Details

Article ID: 17168
Created
Tue 11/7/23 10:04 AM
Modified
Tue 11/7/23 3:57 PM