ScholarLens Back to search
2011•Jisuanji yingyong yanjiuRequires access

High performance FFT computation based on CUDA

Tao Yao

Open publisher page 7 citations

Abstract

The Fourier transform is essential for many image processing and scientific computing techniques.This paper presented an implementation to accelerate FFT computation based on CUDA.Based on the analysis of the GPU architecture and algorithm parallelism feature,brought a mapping strategy used multithread,and explored the optimization in memory hierarchy.The results on CUDA show an improvement,the average speedup reaches 2~6X compared with CUFFT supplied by NVIDIA library.

About this research paper

What this paper is about

The Fourier transform is essential for many image processing and scientific computing techniques.This paper presented an implementation to accelerate FFT computation based on CUDA.Based on the analysis of the GPU architecture and algorithm parallelism feature,brought a mapping strategy used multithread,and explored the optimization in memory hierarchy.The results on CUDA show an improvement,the average speedup reaches 2~6X compared with CUFFT supplied by NVIDIA library.

Why it matters

OpenAlex reports 7 citations for this work. Citation counts describe recorded attention and do not establish research quality.

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

The Fourier transform is essential for many image processing and scientific computing techniques.This paper presented an implementation to accelerate FFT computation based on CUDA.Based on the analysis of the GPU architecture and algorithm parallelism feature,brought a mapping strategy used multithread,and explored the optimization in memory hierarchy.The results on CUDA show an improvement,the average speedup reaches 2~6X compared with CUFFT supplied by NVIDIA library.

Key concepts: CUDA, Computer science, Parallel computing, Speedup, Fast Fourier transform, Memory hierarchy, Computation, Computational science

Related papers

Related paper 1Related paper 2Related paper 3
Back to paper searchBrowse research topicsOriginal source
High performance FFT computation based on CUDA — Research Paper | ScholarLens