ScholarLens Back to search
2011•Jisuanji fangzhenRequires access

FDTD Simulation Using Graphic Processing Units Based on CUDA

Fanmin Kong

Open publisher page 0 citations

Abstract

It takes lots of time to simulate electromagnetic problems with Finite Difference Time Domain(FDTD) method because of numerical dispersive and time domain iteration,but now General Purpose Graphics Processing Units(GPGPU) can be used to solve it.It can be found that FDTD is an inherently data parallel algorithm by analyzing its formulation.Parallel method for FDTD algorithm was put forward using GPU based-on Compute Unified Device Architecture(CUDA) model.Comparing with the calculation with traditional CPU,the CUDA parallel FDTD has enough precision and the calculation speed is differently accelerated in different sizes of Yee cells.The acceleration can come up to tens of times by optimizing parallel method.

About this research paper

What this paper is about

It takes lots of time to simulate electromagnetic problems with Finite Difference Time Domain(FDTD) method because of numerical dispersive and time domain iteration,but now General Purpose Graphics Processing Units(GPGPU) can be used to solve it.It can be found that FDTD is an inherently data parallel algorithm by analyzing its formulation.Parallel method for FDTD algorithm was put forward using GPU based-on Compute Unified Device Architecture(CUDA) model.Comparing with the calculation with traditional CPU,the CUDA parallel FDTD has enough precision and the calculation speed is differently accelerated in different sizes of Yee cells.The acceleration can come up to tens of times by optimizing parallel method.

Why it matters

A significance statement is not available in the OpenAlex record.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

It takes lots of time to simulate electromagnetic problems with Finite Difference Time Domain(FDTD) method because of numerical dispersive and time domain iteration,but now General Purpose Graphics Processing Units(GPGPU) can be used to solve it.It can be found that FDTD is an inherently data parallel algorithm by analyzing its formulation.Parallel method for FDTD algorithm was put forward using GPU based-on Compute Unified Device Architecture(CUDA) model.Comparing with the calculation with traditional CPU,the CUDA parallel FDTD has enough precision and the calculation speed is differently accelerated in different sizes of Yee cells.The acceleration can come up to tens of times by optimizing parallel method.

Key concepts: Finite-difference time-domain method, CUDA, Parallel computing, Computer science, Computational science, General-purpose computing on graphics processing units, Acceleration, Graphics

Related papers

Related paper 1Related paper 2Related paper 3
Back to paper searchBrowse research topicsOriginal source
FDTD Simulation Using Graphic Processing Units Based on CUDA — Research Paper | ScholarLens