Presented By: Nuclear Engineering & Radiological Sciences
NERS Colloquia Series: Exascale Computing Project
Steven Hamilton, Oak Ridge National Laboratory
Abstract:
The ExaSMR project is developing methods and software for coupled Monte Carlo neutron transport and computational fluid dynamics simulations for analysis of nuclear reactors. A part of the larger DOE Exascale Computing Project, ExaSMR is preparing to perform simulations on the nation’s two first exascale computing platforms: the Frontier supercomputer at Oak Ridge National Laboratory and the Aurora machine at Argonne National Laboratory. This talk will provide a brief overview of the Exascale Computing Project, followed by a deep dive into several research topics being investigated by the ExaSMR team. Challenges associated with maximizing the computational performance of both continuous-energy Monte Carlo neutron transport and CFD solvers on GPU-based computing platforms will be addressed, as well as a discussion of strategies for efficiently integrating isotopic depletion into Monte Carlo simulations. Finally, a perspective will be provided on the value of collaboration in computational science applications. Multiple examples will be provided where open engagement beyond typical research teams has resulted in advancements both to individual simulation codes and to the community as a whole.
Bio:
Steven Hamilton is a Senior R&D Staff member in the HPC Methods for Nuclear Applications Group at Oak Ridge National Laboratory, where he has worked since 2011. He received his BS and MS degrees from Georgia Tech in Nuclear Engineering and his PhD from Emory University in Computational Mathematics. Steven currently leads the ExaSMR project, a multiphysics simulation effort targeting advances in nuclear reactor modeling as part of the DOE Exascale Computing Project. His research includes methods for both deterministic and Monte Carlo radiation transport, with a focus on high performance computing and GPU-based computing architectures. He has also worked on nonlinear algorithms for multiphysics simulations as well as linear solvers for computational fluid dynamics solvers. He is an active developer on the Denovo deterministic and Shift Monte Carlo radiation transport codes, both part of ORNL’s SCALE suite of analysis tools.
The ExaSMR project is developing methods and software for coupled Monte Carlo neutron transport and computational fluid dynamics simulations for analysis of nuclear reactors. A part of the larger DOE Exascale Computing Project, ExaSMR is preparing to perform simulations on the nation’s two first exascale computing platforms: the Frontier supercomputer at Oak Ridge National Laboratory and the Aurora machine at Argonne National Laboratory. This talk will provide a brief overview of the Exascale Computing Project, followed by a deep dive into several research topics being investigated by the ExaSMR team. Challenges associated with maximizing the computational performance of both continuous-energy Monte Carlo neutron transport and CFD solvers on GPU-based computing platforms will be addressed, as well as a discussion of strategies for efficiently integrating isotopic depletion into Monte Carlo simulations. Finally, a perspective will be provided on the value of collaboration in computational science applications. Multiple examples will be provided where open engagement beyond typical research teams has resulted in advancements both to individual simulation codes and to the community as a whole.
Bio:
Steven Hamilton is a Senior R&D Staff member in the HPC Methods for Nuclear Applications Group at Oak Ridge National Laboratory, where he has worked since 2011. He received his BS and MS degrees from Georgia Tech in Nuclear Engineering and his PhD from Emory University in Computational Mathematics. Steven currently leads the ExaSMR project, a multiphysics simulation effort targeting advances in nuclear reactor modeling as part of the DOE Exascale Computing Project. His research includes methods for both deterministic and Monte Carlo radiation transport, with a focus on high performance computing and GPU-based computing architectures. He has also worked on nonlinear algorithms for multiphysics simulations as well as linear solvers for computational fluid dynamics solvers. He is an active developer on the Denovo deterministic and Shift Monte Carlo radiation transport codes, both part of ORNL’s SCALE suite of analysis tools.
Explore Similar Events
-
Loading Similar Events...
