Department of Computational and Applied Mathematics

In this talk I present a GPU accelerated solver that implements a high order splitting scheme for a spectral element solution of the incompressible Navier Stokes (INS) equations. While others have implemented this scheme on clusters of CPU-like processors, to my knowledge this work is the first to explore its performance on the GPU. This solver divides the splitting scheme into OpenCL kernels that take advantage of the GPU memory architecture, allowing for massively parallel computations. These rapid computations have the potential to significantly enhance computational fluid dynamics (CFD) simulations that arise in areas such as weather modeling or aircraft design procedures. I present convergence results for several canonical test cases.