Title :
Parallel and distributed solution for 3-dimensional TDFEM analysis in computational electromagnetics
Author :
Wu, Xia ; Jin, Yaqiu ; Zhou, Lezhu
Author_Institution :
Key Lab. of Wave Scattering & Remote Sensing Inf. (Minist. of Educ.), Fudan Univ., Shanghai, China
Abstract :
In this work, a parallel and distributed solution for 3-Dimensional time-domain finite element method (TDFEM) analysis system was presented. The MPI (Message Passing Interface) was the programming model of choice for scientific applications running on large-scale clusters. Therefore, SuperLU of an MPI version so called SuperLU_DIST, to be used on distributed memory machines, was employed for speeding up the computation of the electric field and radiation patterns of 3-Dimensional electromagnetic radiation problems. Second order Lagrange interpolation polynomials was utilized in this TDFEM approach for temporal discretization, and the open-region was truncated by Sacks PML (perfectly matched layer). Two dipole array antennas and a diamond antenna are computed in numerical results section. We analyze the performance of the parallel and distributed TDFEM using MPI library while executing SuperLU, and compared with those from the original TDFEM by using UMFPACK solvers.
Keywords :
antennas; computational electromagnetics; distributed memory systems; message passing; 3-dimensional TDFEM analysis; Lagrange interpolation polynomials; MPI library; computational electromagnetics; distributed memory machines; distributed solution; electromagnetic radiation problems; large scale clusters; message passing interface; parallel solution; temporal discretization; Antenna arrays; Antenna feeds; Antenna radiation patterns; Arrays; Diamond-like carbon; Dipole antennas; Time domain analysis; Antennas; electromagnetic radiation problems; parallel algorithms; time domain finite element methods;
Conference_Titel :
Microwave Conference Proceedings (APMC), 2010 Asia-Pacific
Conference_Location :
Yokohama
Print_ISBN :
978-1-4244-7590-2
Electronic_ISBN :
978-1-902339-22-2