Efficient parallelization of a TSW-MLFMA scheme for large scale problems

Abstract

The Multilevel Fast Multipole Algorithm (MLFMA) has made a great contribution to accelerate the matrix-vector product in the iterative solving of electromagnetic scattering problems. A sophisticated parallel MLFMA for use on distributed-memory computers and the use of large-patch basis functions are good methods to enhance the capability to simulate large scale problems. The article introduces the approach used for the travelling standing wave basis functions (TSW) parallelization based on MLFMA. The method makes it easy to solve bigger scattering problems compared with other methods under the same computing platform.