论文部分内容阅读
基于消息传递平台(message passing interface,MPI),从多层快速多极子算法(multilevel fast multipole algorithm,MLFMA)实施过程出发,实现了飞行器雷达散射截面(radar cross section,RCS)的并行计算,提出一种新的并行设计方案。改进了基于立方体组的并行处理技术,减少了最细层计算的通信量;考虑到内存的主要组成部分,为提高算法并行效率,采用一种用于积分方程中近组作用矩阵的并行处理方法,对其采用按行平均分配到各计算结点的方法组建相关矩阵元素,使该部分通信量为零,各节点可独立计算并节省内存;同时优化了其他物理量的内存分配。算例说明,所提方法可明显提高计算效率,并具有精度好、负载均衡优良的特点。
Based on the message passing interface (MPI), the parallel computation of the radar cross section (RCS) of the aircraft is implemented based on the multilevel fast multipole algorithm (MLFMA) A new parallel design. In order to improve the parallel efficiency of the algorithm, a parallel processing method for the matrix of near-group function in the integral equation is adopted, considering the main components of memory. , And uses the method of equally distributing rows to each calculation node to form the correlation matrix elements so that the traffic of the part is zero and each node can independently calculate and save the memory; meanwhile, the memory allocation of other physical quantities is optimized. The example shows that the proposed method can obviously improve the computational efficiency and has the advantages of good precision and excellent load balance.