Purdue Analysis Facility¶
The Purdue Analysis Facility (Purdue AF) provides an interactive environment for fast and scalable CMS physics analyses using dedicated computing resources at Purdue.
Log in with a Purdue, CERN, or FNAL account — see Login methods.
What you get¶
- A personal JupyterLab session with the CPU cores and memory you choose, and optional Nvidia A100 or T4 GPUs — see Getting started.
- A modern HEP software stack managed via Pixi environments,
including
coffea,ROOT,RDataFrame, and popular machine learning libraries such aspytorch,tensorflow, andxgboost. - Scalable computing via Dask Gateway and, for Purdue users, Slurm batch jobs.
- Multiple data access methods — XRootD, XCache, Rucio — and a variety of private and shared storage volumes.
- Flexible access options: web browser (JupyterLab or VS Code), SSH, your local VSCode-based IDE, or any MCP-capable AI agent.
The software and functionality are regularly updated to provide state-of-the-art tools and features for fast, efficient, collaborative HEP research.
Where to start¶
| If you want to... | Go to... |
|---|---|
| Create your first session | Getting started |
| Learn the interface | How to use Purdue AF |
| Understand where to store your files | Storage volumes |
| Set up an analysis environment | Pixi environments |
| Read CMS datasets | Data access |
| Scale your analysis to hundreds of cores | Scaling out |
| Manage your session with an AI agent | Agentic interface |
| Fix a problem | Troubleshooting |
| Ask a question | Support |
