GPU-accelerated Cellular Automata based Finite-Difference Model for Seismic Wave Propagation with OpenCL
Mark Garrett, Trinh Le, Vinil Nelaturi, Liwen Shih · 2017
Models based on Cellular Automata (CA) paradigm are inherently adaptable to distributed computing and provide opportunity to reach high efficiency when they are implemented on GPUs. Previous efforts have been devoted to accelerate seismic earthquake analysis applications on models based on the centered-difference finite difference (CDFD) paradigm. In this paper, these two paradigms are tested on CPU and GPU processors in the study of earthquake wave propagation. The preliminary results of performance evaluations are presented and discussed.