Available software

From Alliance Doc
Revision as of 03:27, 19 July 2017 by Phillips (talk | contribs) (Inserted reference to Python wheels under /cvmfs/soft.computecanada.ca/custom/python/wheelhouse/.)
Jump to navigation Jump to search
Other languages:

A current list of the software available on Compute Canada national systems is below. Note that this list changes frequently as new software is added and that users can request the installation or updating of a particular program or library by writing to support@computecanada.ca. A list of software installed on legacy regional systems can be found here,

Note that some prerequisites are loaded by default. See Using modules for details on how to use the module system.

Here are a few things to know about the software stack:

  • Most Python modules are not installed as modules. They are instead provided as binary wheels, stored on the Compute Canada systems under /cvmfs/soft.computecanada.ca/custom/python/wheelhouse/. One such package is Tensorflow. For instructions on how to install Python packages, see the Python page.
  • Similarly, most R or Perl packages are not installed either. We recommend installing them in your personal or group file space. See the R and Perl pages for instructions on how to do so.
  • Note that Docker is not available on Compute Canada clusters but Singularity is available in the directory /opt/software. Docker containers can be converted to Singularity as discussed here.
  • Some of the software packages listed below are not accessible to all users because they require you to have a license. You may need to be granted access to them by us. Loading the module for one of these will give you instructions on what to do to obtain access to it.
  • While the vast majority of the software packages below are accessible on all Compute Canada servers, there may be exceptions due to licensing restrictions.
  • Many packages related to the operating system, such as Autotools, Make, Git, and others, are not installed as modules but are part of the default environment. These are not listed below.
Other languages:

<translate>

In computing, a module is a unit of software that is designed to be independent, interchangeable, and contains everything necessary to provide the desired functionality. [1] The term "module" may sometimes have a more specific meaning depending on the context. This page describes a few types of modules and suggests links to further documentation content.

Disambiguation

Lmod modules

Also called "environment modules", Lmod modules are used to alter your (shell) environment so as to enable you to use a particular software package, or to use a non-default version of certain common software packages such as compilers. See Using modules.

Python modules

In Python, a module is a file of code (usually Python code) which can be loaded with the import ... or from ... import ... statements to provide functionality. A Python package is a collection of Python modules; the terms "package" and "module" are frequently interchanged in casual use. [2]

Certain frequently used Python modules such as Numpy can be imported if you first load the scipy-stack Lmod module at the shell level. See SciPy stack for details.

We maintain a large collection of Python "wheels." These are modules which are pre-compiled to be compatible with the Standard software environments. Before importing modules from our wheels, you should create a virtual environment.

Python modules which are not in the scipy-stack Lmod module or in our wheels collection can be installed from the internet as described in the Installing packages section.

Other related topics

The main Available software page is a good starting point. Other related pages are:

Footnotes

</translate>