CUDA-Based GPU Acceleration of Warpxm Continuum Kinetic Modeling

High-fidelity simulations of plasma dynamics using the continuum kinetic model are challenging due to the high dimensionality of the Vlasov equation, involving up to three dimensions in both space and velocity. However, the model is useful for understanding the plasma physics associated with sheared-flow-stabilized Z pinches that cannot be easily deduced from reduced multi-fluid models. Therefore, efforts to reduce computational time and resources required for simulations involving the continuum kinetic model are important to help gain a full understanding of various aspects of these Z pinches. This research involves work to accelerate an existing high order discontinuous Galerkin implementation of the continuum kinetic model in the WARPXM framework developed at the University of Washington and Zap Energy, harnessing massively parallel GPU architectures using the CUDA programming language. Preliminary results comparing CPU and GPU timings will be presented as well as lessons learned.