Parallelization of Tsunami Simulation on CPU, GPU and FPGAs
Authors: Fumiya Kono (University of Aizu), Naohito Nakasato (University of Aizu), Kensaku Hayashi (University of Aizu), Alexander Vazhenin (University of Aizu), Stanislav Sedukhin (University of Aizu), Kohei Nagasu (Tohoku University), Kentaro Sano (Tohoku University), Vasily Titov (National Oceanic and Atmospheric Administration)
Abstract: Tsunami is known as one of serious disasters. MOST (Method of Splitting Tsunami) is one of numerical solvers for modeling tsunami waves. Prediction of the arrival time of Tsunami is critical to evacuate people from coastal area. Therefore, fast computation of MOST enabled by parallel processing is important. We have developed a tsunami propagation code based on MOST and implemented different algorithms for parallelization by OpenMP and OpenACC. We have conducted benchmarking of these parallelized codes on various architectures such as multi-core CPU systems, Many Integrated Core architecture, GPU. In this poster, we compare the performance of various parallel implementations. We found that a parallelized code applied spatial blocking on OpenACC was the best performance at present.
Concurrently, we are developing an accelerator of the MOST computation with Field-Programmable Gate Arrays (FPGAs) for high-performance and power-efficient simulation. We also present preliminary evaluation of its prototype implementation with a 28nm FPGA.
Two-page extended abstract: pdf