Porting WarpX to GPU-accelerated platforms
Andrew T. Myers, Ann Almgren, L. D. Amorim, John B. Bell, Luca Fedeli, Lixin Ge, Kevin Gott, Dave P. Grote, Mark Hogan, A. Huebl, Revathi Jambunathan, Remi Lehé, C. Ng, Michael Rowan, Olga V. Shapoval, Maxence Thévenet, Jean-Luc Vay, Henri Vincenti, Eloise Yang, N. Zaïm · Parallel Computing · 2021
WarpX is a general purpose electromagnetic particle-in-cell code that was originally designed to run on many-core CPU architectures. We describe the strategy, based on the AMReX library, followed to allow WarpX to use the GPU-accelerated nodes on OLCF’s Summit supercomputer, a strategy we believe will extend to the upcoming machines Frontier and Aurora. We summarize the challenges encountered, lessons learned, and give current performance results on a series of relevant benchmark problems.