High Performance Computing: The Computational Resources behind the Agency’s Environmental Modeling and Visualization
High Performance Computing (HPC) provides the infrastructure, software and technical support for EPA researchers and scientists to run their own compute jobs on EPA HPC machines. This enables quick and easy access to the resources needed for a variety of research, including work on the Community Multiscale Air Quality Model (CMAQ) -- a powerful computational tool used by EPA and states for air quality management -- the Coastal Generalized Ecosystem Model (CGEM) and the Comprehensive Air Quality Model with extensions (CAMx).
High-End Computing Resources
The main high-end scientific computing resource is the HPC system Atmos. Atmos consists of Dell PowerEdge servers configured with 164 compute nodes, 14 debug nodes, two single processor nodes, two large memory nodes, and four GPU nodes.
Compute time on Atmos is shared by EPA scientists and researchers to run models that study air quality, water quality, contaminant transport, and the effects of chemicals on human and ecosystem health. For example, the Center for Computational Toxicology and Exposure (CCTE) uses HPC to study the chemical and molecular properties of contaminants to investigate toxicity and the risk to people and the environment.
Projects are allocated time on the HPC system through an annual review process. For FY 2024, projects are allocated around 51 million CPU hours.
Visualization Resources
Sophisticated visualization hardware and software support high-end, R&D visualization computing, providing the ideal technology for projects such as modeling the human respiratory system or air pollutant dispersion through an urban landscape. These visualizations can be converted to animations for display or graphics suitable for publishing.