研究目的
To propose a novel and efficient algorithm to parallelize the finite-difference time-domain method, reducing computational time significantly.
研究成果
The proposed time-division parallel FDTD algorithm is simple and efficient, allowing the observation period to be divided into an arbitrary number of subsections for distributed computation without data exchange among nodes. It notably reduces computational time and is suitable for studying a wide variety of time-domain electromagnetic, optic, and plasmonic problems.
研究不足
The computational cost of obtaining the initial response becomes more expensive than that for FDTD computation when the number of nodes increases beyond a certain point for short observation periods.
1：Experimental Design and Method Selection:
The proposed algorithm divides the observation period into subsections, distributing computation to corresponding computer nodes without requiring data exchange among nodes during parallel computation.
2：Sample Selection and Data Sources:
A two-dimensional silver cylinder with diameter d = 1 nm was used to verify the performance and accuracy of the proposed parallel FDTD algorithm.
3：List of Experimental Equipment and Materials:
Computational model using the conventional FDTD method, computer nodes for parallel computing.
4：Experimental Procedures and Operational Workflow:
The observation period is divided into subsections, initial responses are computed using the FDCFD–FILT technique, and the conventional FDTD method is applied to update the electromagnetic field estimates.
5：Data Analysis Methods:
The time-domain response is obtained by merging the partial responses of all subsections, and the accuracy is compared with the conventional FDTD method.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
