Parallel Implementation and Branch Optimization of EBE-FEM Based on CUDA Platform

Authors

  • Yan Zhang School of Electrical Engineering Shenyang University of Technology, Shenyang, 110870, China
  • Xiuke Yan School of Electrical Engineering Shenyang University of Technology, Shenyang, 110870, China
  • Xudong Ren School of Electrical Engineering Shenyang University of Technology, Shenyang, 110870, China
  • Sheng Wang School of Electrical Engineering Shenyang University of Technology, Shenyang, 110870, China
  • Dongyang Wu School of Electrical Engineering Shenyang University of Technology, Shenyang, 110870, China
  • Baodong Bai School of Electrical Engineering Shenyang University of Technology, Shenyang, 110870, China

Keywords:

Branch optimization, CUDA, EBEFEM, parallel computation

Abstract

The finite element analysis of large complex structures makes higher demand on memory capacity and computation speed, which leads to the inefficiency of traditional serial finite element method (FEM) for such large-scale problems. In this paper, the element-byelement finite element method (EBE-FEM) has been implemented parallelly on CUDA (Compute Unified Device Architecture) platform, and been programmed using C++ language. The thread branches that exist in parallel reduction program have been researched and optimized to improve parallel efficiency. The correctness of algorithms and programs are verified by the analysis of an open slot of motor. The optimized parallel program is applied to analyze the main magnetic field of a singlephase transformer. The results show that the EBE-FEM implemented on CUDA platform is more effective than serial EBE-FEM, and branch optimization can improve the speedup further.

Downloads

Download data is not yet available.

Author Biographies

Yan Zhang, School of Electrical Engineering Shenyang University of Technology, Shenyang, 110870, China

Yan Zhang He received his B.S. degree in Electrical Engineering and Automation from Langfang Normal University in 2017. He is currently working towards M.S degree in Shenyang University of Technology. His research interests include numerical analysis and parallel calculation of electromagnetic fields on CUDA platform.

Xiuke Yan, School of Electrical Engineering Shenyang University of Technology, Shenyang, 110870, China

Xiuke Yan received her B.S. degree, M.S. degree and Ph.D. degree in Electrical Engineering from Shenyang University of Technology, China, in 1996, 1999 and 2005, respectively. She is currently a Professor in Shenyang University of Technology. Her research interests include numerical analysis of coupled field and optimization design of electrical equipment, parallel algorithm research of finite element method.

References

D. M. Fernandez, M. M. Dehnavi, W. J. Gross, and D. Giannacopoulos, “Alternate parallel processing approach for FEM,” IEEE Trans. Magn., vol. 48, no. 2, pp. 399-402, Feb. 2012.

E. Barragy, G. Raham, and F. Carey, “Parallel element-by-element solution scheme,” Int. J. Numer. Methods Eng., vol. 26, no. 11, pp. 2367-2382, 1988.

X. Yan, X. Han, D. Wu, D. Xie, B. Bai, and Z. Ren, “Research on preconditioned conjugate gradient method based on EBE-FEM and the application in electromagnetic field analysis,” IEEE Trans. Magn., vol. 53, no. 6, pp. 1-4, June 2017.

A. Capozzoli, O. Kilic, C. Curcio, and A. Liseno, “The success of GPU computing in applied electromagnetics,” Applied Comp. Electromag. Soc. Journal, vol. 33, no. 2, pp. 148-151, Feb. 2018.

J. C. K. Wake and S. Watanabe, “Scalable GPUparallelized FDTD method for analysis of largescale electromagnetic dosimetry problems,” Applied Comp. Electromag. Soc. Journal, vol. 31, no. 6, pp. 661-668, June 2016.

D. Wu, X. Yan, R. Tang, and D. Xie, “Parallel realization of element by element analysis of eddy current field based on graphic processing unit,” Applied Comp. Electromag. Soc. Journal, vol. 33, no. 2, pp. 168-171, Feb. 2018.

J. Nickolls, “GPU parallel computing architecture and CUDA programming model,” IEEE Hot Chips Symp., HCS, pp. 1-12, May 2007.

I. Kiss, S. Gyimothy, Z. Badics and J. Pavo, “Parallel realization of the element-by-element FEM technique by CUDA,” IEEE Trans. Magn., vol. 48, no. 2, pp. 507-510, Feb. 2012.

S. Wang, X. Yan, Y. Zhang, and D. Wu, “Research on EBE-FEM realized by CUDA applying to electromagnetic field analysis,” IEEE Stud. Conf. Electric Mach. Syst., SCEMS, 2018.

E. Zhang, Y. Jiang, and Z. Guo, “Streamlining GPU applications on the fly: Thread branch elimination through runtime thread-data remapping,” Proc. Int. Conf. Supercomputing, pp. 115-125, 2010.

Z. Hu, Analysis and Calculation of Motor Electromagnetic Fields. Beijing: China Machine Press, 1989.

Downloads

Published

2020-06-01

How to Cite

[1]
Yan Zhang, Xiuke Yan, Xudong Ren, Sheng Wang, Dongyang Wu, and Baodong Bai, “Parallel Implementation and Branch Optimization of EBE-FEM Based on CUDA Platform”, ACES Journal, vol. 35, no. 6, pp. 595–600, Jun. 2020.

Issue

2020: Vol 35 Iss 6

Section

Articles