Tuesday, July 23
Transport equations are important in many different applications such as nuclear design, radiation therapy in medical science, and radiation effects in global weather models. They are also a fundamental part of many algorithms used to model more complicated applications such as shielding of the electronics satellites.
New numerical methods for solving these equations are being designed. This is specially true when considering parallel architectures. Parallel algorithms are important since transport equations involve many variables, such as angle of scaterring, flux and energy and time. Standard approaches to solving transport discretizations have to be modified to get good results on high performance computers. The speakers will discuss several approaches on different computer architectures.
Parallel Methods in Transport Theory
Organizer: Suely B. Oliveira
Texas A&M University
- 8:30 A Time-Optimal Parallel Ray Tracing Algorithm for One-Dimensional Discrete-Ordinate Computations
- Robert D. Jarvis and Paul Nelson, Texas A&M University
- 9:00 Parallel Algorithms for Linear Boltzmann Equation Based on Complete Phase Space Decomposition
- Ali Haghighat, G. Sjoden, and M. Hunter, Pennsylvania State University
- 9:30 3-D Self-Adjoint Unstructured Grid Radiation Transport Methods
- Jim E. Morel and John M. McGhee, Los Alamos National Laboratory
- 10:00 Phase Space Decomposition for a Massively Parallel Neutron Transport Model
- Milo R. Dorr, Lawrence Berkeley Laboratory