- Indico style
- Indico style - inline minutes
- Indico style - numbered
- Indico style - numbered + minutes
- Indico Weeks View
Title: “Multiphysics Simulations on GPUs using high-order finite elements”
Abstract: Lawrence Livermore National Laboratory (LLNL) is developing next generation simulation capabilities for exascale computing systems as part of the Multiphysics on Advanced Platforms Project (MAPP). One of the key outcomes of this project is MARBL, an arbitrary Lagrangian-Eulerian (ALE) multiphysics code that utilizes high-order finite element discretizations. In this presentation, we will discuss the finite element methods created for ALE multiphysics, with a focus on “matrix-free” partial assembly techniques that allow for scalable and GPU-efficient computation of high-order finite element operators. We will also outline the multiphysics capabilities of MARBL and its applications in modeling inertial confinement fusion (ICF) and high-energy density physics (HEDP) experiments, including large-scale GPU simulations on the El Capitan supercomputer. Additionally, we will cover recent work applying MARBL to astrophysics simulations and conclude with our vision for the future centered on the development of differentiable multiphysics codes.
Bio: Rob Rieben is a computational physicist and code project leader at the Lawrence Livermore National Laboratory (LLNL) where he has spent the past 25 years developing and implementing numerical methods for massively parallel, high-performance, multi-physics simulation codes for the National Nuclear Security Administration (NNSA). He is currently the lead for LLNL’s next-generation multi-physics code as part of the Multi-Physics on Advanced Platforms Project (MAPP), where he and his team are developing algorithms and simulation capabilities to run on the world’s most powerful supercomputers to enable scientific discovery and tackle some of the most complex challenges in advancing national security and exploring the frontiers of high energy density physics.