GPU Acceleration of Riemann Solvers

GPU Acceleration of Riemann Solvers

Internship Description

Riemann solvers are the core algorithms in numerical methods for wave propagation. Because they can exhibit high arithmetic intensity, they are well suited to implementation on GPUs. In this project, the student will further develop an existing experimental GPU Riemann
solver (CUDAClaw) into a software tool that is accessible to application scientists incomputational fluid dynamics and water waves. This will include interfacing the code with the PyClaw software and implementing additional GPU Riemann solver kernels.
​​​

Deliverables/Expectations

A GPU-enabled solver and a report.

Faculty Name

Field of Study

Mathematics, computer science, or engineering