为提高网络环境下细粒度并行FDTD计算的性能,引入局域网两层并行能力的概念,发展了一种高性能的混合并行FDTD算法。在该算法中,通过利用MPI与OpenMP多线程技术,在传统域分解FDTD算法基础上,实现了数据与任务的两层并行化。作为算法实现的应用,对一种常见的车载隐藏式印刷天线进行了模拟研究。计算在不同数量的PC机上执行,并与传统的网络并行FDTD算法进行了比较。数值结果表明,当域分解粒度较小时,该混合并行方法能够有效地提高局域网并行FDTD的算法性能。
A high performance hybrid parallel PDTD algorithm in developed by introducing the concept of two layers of parallel capacity based on local area network(LAN). In the hybrid algorithm, MPI is used in conjunction with OpenMP multithreading to achieve two level parallelism of the data and tasks at the basis of the domain decomposition FDTD method. By using this hybrid algorithm, a common hidden printed antenna for automobile is simulated. Computation results show when the grain size of the domain decomposition is on the small side, this novel hybrid parallelization scheme is an effective technique for improving the efficiency of the parallel FDTD based on LAN.