Computational Physics and Methods

Enabling innovative large-scale simulations of physics phenomena on tomorrow's scientific computing
platforms

Innovating in methods research and physics software development to model complex physical systems at scale

We develop the scientific methods and software that power discovery, from open-source research tools to production-scale codes operating on the world’s fastest supercomputers. Through advanced computational techniques and multi-scale, multi-physics models, we uncover the hidden dynamics of complex systems, from turbulent fluid flows to the life cycles of stars. Our expertise includes radiation transport, shock hydrodynamics, turbulent mixing, earth systems, astrophysics, plasma physics, and physics-aware machine learning. We advance computational physics methods and software to expand scientific frontiers in service of the nation. 

R&D Focus Areas

Astrophysics

Astrophysics spans extreme scales and processes, from the origins of our solar system to the extreme ends of stars’ lives. We focus on nuclear astrophysics, transient astrophysics, and planetary astrophysics, using advanced computational methods to model phenomena from plasma microphysics to cosmic evolution. Our staff are also affiliated with LANL’s Center for Theoretical Astrophysics, strengthening connections across the broader astrophysics community. 

Software

Artemis: An astrophysical multifluid radiation hydrodynamics code built on Parthenon, with particular attention paid to methods for accuracy and robustness when studying planet formation

AthenaPK: Performance portable version of Athena++ built on Parthenon and Kokkos FleCSPH Smoothed Particle Hydrodynamics code with solid material modeling Nublight General Relativistic Neutrino Radiation Magnetohydrodynamics for Neutron Star Merger Disks

Parthenon: Performance Portable Block-Structured AMR Framework

Phoebus: A general purpose general relativistic radiation magnetohydrodynamics solver Singularity-EOS Performance portable equations of state, both terrestrial and astrophysical

Singularity-Opac: Opacity physics for both thermal (photon) and neutrino transport

Computational Fluid and Plasma Dynamics

We develop advanced simulation tools to model complex physical processes, including metal casting, fluid dynamics, plasma kinetics, shock hydrodynamics, and radiation hydrodynamics. Our research delivers high-performance software that expands scientific understanding and supports applications of national importance.

We created MASS-APP, a powerful exascale-ready framework for modeling plasmas across scales. Built on LANL’s FleCSI framework, it opens new frontiers in understanding space weather, global security, astrophysics, and fusion.

We develop codes and methods in support of the Inertial Confinement Fusion science campaigns for simulating and analyzing capsule implosions and other high-energy-density physics (HEDP) experiments.

Software

Quinoa: Adaptive computational fluid dynamics.
Enabling research and numerical analysis in fluid dynamics.

Truchas: Physics-based modeling and simulation of manufacturing processes
Core capabilities for metal casting, with developing capabilities for metal additive manufacturing and microwave casting.

Earth Systems Modeling

We contribute to large, multi-institutional DOE research and production code projects funded by the Office of Science Biological and Environmental Research (BER). Our team specializes in ocean and sea-ice dynamics, polar science and AI-enabled real-time natural disaster prediction.

Projects

InteRFACE: Improving our fundamental understanding of changes in Arctic coastal systems

HiLATRASM: Answering Arctic and Antarctic climate change questions through targeted application of global modeling and analysis capabilities

E3SM: State-of-the-science Earth system modeling, simulation, and prediction

Physics-Aware AI/ML

We develop AI and machine learning methods that integrate physics knowledge to make scientific computing faster, more reliable, and physically consistent.

• Mission-Driven AI for Physics – Inverse problems, inference, and other critical challenges
• Model Integration – Quickly moving new approaches into production through strong partnerships with modelers
• Physics + Experiments – Combining expertise and data to improve analysis with state-of-the-art AI/ML techniques
• Reduced-Order Models – Building efficient data-driven tools for complex dynamical systems
• Multiscale AI – Designing methods that connect physics across different length and time scales

Radiation Transport

We develop cutting-edge methods to solve the radiation transport equation for a wide range of mission-driven and astrophysics applications.

Software

PARTISN: PARallel TIme-dependent SN
Jaybenne: A Monte Carlo thermal radiation transport library built on top of the Parthenon framework for easy incorporation into multi-physics software
SuperNu: Monte Carlo radiative transfer for modeling supernova and kilonova spectra

University of Michigan Collaboration - SPARC

In 2024, we became the first LANL group to assign staff to LANL’s UM duty station! Our team works with UM faculty and students on LANL-relevant challenges, including advancing software and methods for the RIOT multiphysics code and developing AI/ML approaches for modeling and designing HEDP experiments. To learn more about this exciting collaboration, vist the LANL Michigan SPARC webpage.