Vector unit acceleration technique for DGTD method

In this paper, we present a hardware acceleration technique using the SSE (streaming SIMD (single instruction multiple data) extensions) instruction set to improve discontinuous Garlekin method Time Domain (DGTD) for solving Maxwell's equations. The improvement performance of the DGTD methods is investigated via some typical examples. The numerical results of the SSE accelerated DGTD method on the modern 4-core CPU demonstrates that the speedup ratio can reach up to 4 times compared to the traditional DGTD method.