.. _cluster-software: ******** Software ******** The CoSTAR cluster supports multiple software workflows. Whether you need a modern containerised pipeline, a quick Python environment, or a graphical interface, you have total control over your software stack. Choose your preferred workflow below to learn more: .. raw:: html
.. grid:: 1 1 3 3 :gutter: 3 :margin: 1 .. grid-item-card:: :octicon:`package` Containers :link: software-container :link-type: ref The recommended HPC standard. Build once, run anywhere, and ensure perfect reproducibility. * **Apptainer:** The natively supported platform on CoSTAR. * **Docker:** Easily convert Docker workflows to Apptainer. * **Registries:** Pull directly from Docker Hub or University registries. .. grid-item-card:: :octicon:`browser` Interactive Apps :link: software-ood :link-type: ref Graphical user interfaces and real-time interactivity directly in your web browser. Launch interactive desktop sessions and graphical tools including **JupyterLab**, **VS Code**, and **MATLAB** without any command-line setup. .. grid-item-card:: :octicon:`terminal` Virtual Environments (Conda) :link: software-conda :link-type: ref Manage Python and R dependencies locally within your home directory. Read our full guide on setting up **Miniconda**, which is our recommended tool for handling complex scientific packages and GPU dependencies. Note: Standard tools like Python's built-in ``venv`` or high-speed managers like ``uv`` also work well on the cluster. Requesting New Software ======================= CoSTAR is designed for **users to manage their own software stacks** using :ref:`Containers