Search thousands of pre-installed software packages across all ACCESS resources — or load them directly using environment modules.
The Software Documentation Service
The ACCESS Software Documentation Service (SDS) is a searchable catalogue of software available across all ACCESS resources. It indexes thousands of packages — compilers, MPI libraries, Python environments, domain codes, and commercial tools. Search by software name or category to see which resources have it installed and what version is available.
Using Environment Modules
Once logged in to an ACCESS resource, software is loaded using the Environment Modules system. Modules set up your shell environment — paths, libraries, and compiler flags — so you do not need to manage these manually.
| Command | Description |
|---|---|
module avail | List all modules available on this system |
module avail python | List all Python-related modules |
module load python/3.11 | Load a specific module version |
module list | Show currently loaded modules |
module unload python | Unload a module |
module purge | Unload all loaded modules (start clean) |
module show python/3.11 | Show what a module sets up (paths, env vars) |
Put module commands in your job script
Always include module purge followed by your required module load commands at the top of your Slurm job script. This ensures reproducibility — your job will use exactly the versions you specified regardless of what you have loaded in your interactive session.
Software via Science Gateways
Some domain-specific software is best accessed through a Science Gateway rather than loading a module. Gateways provide validated, pre-configured software stacks with a domain-specific interface — no command line required.
Browse available science gateways
Containers (Singularity / Apptainer)
If the software you need is not available as a module and is not in a science gateway, containers are often the best option. Most ACCESS resources support Singularity / Apptainer containers, which let you bring your own software environment. See the full guide to using containers on ACCESS.
Docker images work too — via Singularity
You can convert any Docker image to a Singularity container and run it on ACCESS resources without root access. Ask.CI has guides on this.
Frequently Asked Questions
The software I need isn't available — what should I do?
First, check the Software Documentation Service (sds.access-ci.org) to confirm it's not on another resource. Then try:
- Install into your home directory (
pip install --user, conda, spack) - Use a Singularity container with the software pre-installed
- Open a support ticket asking the RP to install it system-wide (support.access-ci.org/open-a-ticket)
Can I install Python packages with pip or conda?
Yes. Use pip install --user PACKAGE to install into your home directory without root access. For more complex environments, conda is usually available via a module (module load anaconda or similar) — create a named environment with conda create -n myenv so packages are isolated per project. Avoid installing packages into the base conda environment, as it may conflict with system software. For GPU-accelerated packages (e.g., PyTorch, TensorFlow), load the appropriate CUDA module before installing.
Is commercial software like MATLAB or ANSYS available?
Some ACCESS resources have site licenses for commercial software such as MATLAB, ANSYS, Gaussian, and others. Availability varies by resource. Check the Software Documentation Service (sds.access-ci.org) for the most current list. If a package is not listed, contact the resource provider — they can confirm whether a license exists and how to access it. For MATLAB specifically, many users run it without a GUI using matlab -nodisplay -r 'script' inside a job script.
How do I compile my own code on ACCESS resources?
Load a compiler module before compiling: module load gcc (or intel, nvhpc, etc.). Use module avail to see what is installed. For MPI-parallel code, load an MPI module (module load openmpi) and compile with mpicc / mpif90. Build systems like CMake and Make are available on all resources. If your code depends on a library (HDF5, NetCDF, FFTW), check whether a module exists before building from source — pre-built modules are optimized for the hardware.