Automatic Load-Balanced Partitioning Strategy in Parallel Multilevel Fast Multipole Algorithm

2012 ◽  
Vol 182-183 ◽  
pp. 893-897
Author(s):  
Zhi Xun Gong ◽  
Wei Qin Tong ◽  
Ying Li
Keyword(s):  
Load Balancing ◽  
Electromagnetic Scattering ◽  
Large Scale ◽  
Fast Multipole ◽  
Scattering Problems ◽  
Multilevel Fast Multipole Algorithm ◽  
Transition Level ◽  
Fast Multipole Algorithm ◽  
Partitioning Strategy ◽  
Load Balanced

In this paper, we propose an automatic load-balanced partitioning strategy for parallel multilevel fast multipole algorithm(MLFMA) based on distributed-memory architectures to solve the large scale electromagnetic scattering problems. We focus on the automatic load-balancing partitioning strategy because that our original scheme requires that users input the transition level, which is usually determined on the users’ experience and sometimes lead to a bad load-balancing partition. By introducing the automatic load-balancing algorithm to our pervious partitioning technique, our implementation can automatically achieve the best load-balancing and consequently attain better parallel efficiency. To present the effectiveness of the new strategy, we analyze results of the previous implementation according to different inputs of transition level and compare them with the result of implementation using the new algorithm.

scholarly journals Acceleration of the Multi-Level Fast Multipole Algorithm Using K-Means Clustering

Electronics ◽  
2020 ◽  
Vol 9 (11) ◽  
pp. 1926
Author(s):  
Dal-jae Yun ◽  
Haewon Jung ◽  
Hoon Kang ◽  
Woo-Yong Yang ◽  
Dong-Wook Seo
Keyword(s):  
Transfer Function ◽  
Electromagnetic Scattering ◽  
Computation Time ◽  
Addition Theorem ◽  
Iterative Solver ◽  
Fast Multipole ◽  
Multilevel Fast Multipole Algorithm ◽  
Multi Level ◽  
Simulation Results ◽  
Fast Multipole Algorithm

The multilevel fast multipole algorithm (MLFMA) using K-means clustering to accelerate electromagnetic scattering analysis for large complex targets is presented. By replacing the regular cube grouping with the K-means clustering, the addition theorem is more accurately approximated. The convergence rate of an iterative solver is thus improved significantly. However, irregular centroid locations as a result of the K-means clustering increase the amount of explicit transfer function calculations, compared with the regular cubes. In the MLFMA, a multilevel hierarchical structure is applied to the finite multipole method (FMM) to reduce transfer function calculations. Therefore, the MLFMA is suitable for applying K-means clustering. Simulation results with both canonical and realistic targets show an improvement in the computation time of the proposed algorithm.

Sign in / Sign up

Export Citation Format

Share Document