Acceleration of FDTD Method Using Deep Computing Unit
Parallel computing finds extensive applications in numerical modeling of large-scale electromagnetic problems. We present a method for accelerating finite-difference time-domain (FDTD) method using deep computing unit (DCU). Simulation results show that the proposed method provides a significant improvement in efficiency compared to serial FDTD and parallel FDTD using multicore CPU. It offers a potential improvement for FDTD method in solving large-scale electromagnetic (EM) problems combining with high-performance computing (HPC) technology.
Authors
- Qingtao Sun (ORCID: https://orcid.org/0000-0002-7022-2616)
- Yan Tian (ORCID: https://orcid.org/0000-0002-3770-0604)
- Wen Chen (ORCID: https://orcid.org/0000-0002-3258-7278)
- Lixiao Wang (ORCID: https://orcid.org/0009-0000-8274-4940)
- Qing Liu (ORCID: https://orcid.org/0000-0001-5286-4423)
Institutions
- Hong Kong Polytechnic University (HK)
Publication Details
- Published
- 2024-10-23
- DOI
- https://doi.org/10.1109/isape62431.2024.10841235
- Primary Topic
- Electromagnetic Simulation and Numerical Methods
- Type
- article
- Field-Weighted Citation Impact
- 0.00