Elements: Toolkit for High-Order Neutrino-Radiation Hydrodynamics in General Relativity

University Of Tennessee Knoxville, Knoxville TN

Investigators

Abstract

The majority of elements in the Periodic Table of Elements are synthesized either in massive stars and their explosive deaths in core collapse supernovae or in the merger of neutron stars born in such supernovae. Supernovae and neutron star mergers are complex, explosive, turbulent, and energetic events. They involve strong-field gravity, turbulent fluid flow, magnetic fields, radiation in the form of subatomic particles known as neutrinos, thermonuclear reactions, and high-density nuclear physics and exotic particle physics not seen in terrestrial experiments. The solution of the complex equations governing the dynamics of these catastrophic events requires the development of high-fidelity methods and simulation frameworks to execute these solution methods on the world’s leading supercomputing platforms. These are arduous tasks requiring the multidisciplinary expertise of astrophysicists, applied mathematicians, and computer scientists. The completion of these tasks typically requires years of development. The development of high-fidelity methods and well-engineered and high-performance open-source software then enables the broader astrophysical modeling community and facilitates the reproducibility of results, critical when complex systems are involved. Such development provides the astrophysics community with a set of optimized tools to explore some of the Universe’s most important phenomena, utilizing the National Science Foundation’s major investments in the Laser Interferometer Gravitational Observatory, designed to detect gravitational waves from core collapse supernovae and neutron star mergers, which encode key information about the dynamics in such events. The project offers a multidisciplinary research environment that contributes to training the next generation of computational scientists, bridging physics, mathematics, and high-performance computing. This project implements novel software elements for relativistic, spectral two-moment neutrino kinetics in the open source toolkit for high-order neutrino-radiation hydrodynamics (thornado). The methods are based on discontinuous Galerkin (DG) discretization and implicit-explicit (IMEX) time stepping.The project designs robust methods that capture key continuum properties of the phase-space flow (i.e., respecting physical bounds and conservation laws). thornado kinetics solvers, together with existing thornado solvers for hydrodynamics and gravity, and recognized open-source computational science software technologies, are coupled for deployment on high-performance computing systems to enable large-scale multi-physics models of core-collapse supernovae and related problems in nuclear astrophysics. Distributed parallelism and adaptive mesh refinement is enabled by AMReX, and time accurate multi-physics coupling is enabled by the SUNDIALS library, which offers advanced multi-rate IMEX integration and adaptive step-size control. For deployment on heterogenous computing systems, the project instantiates implementations that target accelerators (GPUs). The project joins expertise in physics, mathematics, and computational science to develop novel computational tools to improve understanding of core collapse supernovae — among the most energetic events in the Universe — and provide better predictions of their neutrino and gravitational wave signatures. It leverages long-standing research collaborations to train next-generation scientists in a multidisciplinary research environment. This award by the Office of Advanced Cyberinfrastructure is jointly supported by the Division of Physics in the Mathematics and Physical Sciences Directorate. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.