TY - GEN
T1 - Parallel subdomain level DGTD method with automatic load balancing
AU - Ren, Qiang
AU - Mi, Jiamei
N1 - Publisher Copyright:
© 2019 IEEE.
PY - 2019/12
Y1 - 2019/12
N2 - In this paper, a parallel subdomain-level discontinuous Galerkin time domain (DGTD) method based on the Message Passing Interface (MPI) library has been proposed for simulating complex structures or multiscale electromagnetic problems. The efficiency of parallel algorithm is greatly affected by load distribution, so an automatic load balancing strategy is proposed to reduce the load difference between processes in order to show the advantages of parallel algorithms. First, the relationship between the time required to solve system matrix and the degree of freedom (DoF) of the subdomains applying tetrahedral or hexahedral elements is obtained by some numerical experiments. Then the partition is adjusted to achieve load balancing and reduce the data exchange between processes by relationship above, so that the parallel algorithm approaches the linear speedup ratio. Finally, some numerical cases have been simulated to demonstrate the reliability and efficiency of the algorithm.
AB - In this paper, a parallel subdomain-level discontinuous Galerkin time domain (DGTD) method based on the Message Passing Interface (MPI) library has been proposed for simulating complex structures or multiscale electromagnetic problems. The efficiency of parallel algorithm is greatly affected by load distribution, so an automatic load balancing strategy is proposed to reduce the load difference between processes in order to show the advantages of parallel algorithms. First, the relationship between the time required to solve system matrix and the degree of freedom (DoF) of the subdomains applying tetrahedral or hexahedral elements is obtained by some numerical experiments. Then the partition is adjusted to achieve load balancing and reduce the data exchange between processes by relationship above, so that the parallel algorithm approaches the linear speedup ratio. Finally, some numerical cases have been simulated to demonstrate the reliability and efficiency of the algorithm.
UR - https://www.scopus.com/pages/publications/85082439279
U2 - 10.1109/PIERS-Fall48861.2019.9021794
DO - 10.1109/PIERS-Fall48861.2019.9021794
M3 - 会议稿件
AN - SCOPUS:85082439279
T3 - 2019 Photonics and Electromagnetics Research Symposium - Fall, PIERS - Fall 2019 - Proceedings
SP - 2330
EP - 2336
BT - 2019 Photonics and Electromagnetics Research Symposium - Fall, PIERS - Fall 2019 - Proceedings
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2019 Photonics and Electromagnetics Research Symposium - Fall, PIERS - Fall 2019
Y2 - 17 December 2019 through 20 December 2019
ER -