Name
Abstract | ||
|---|---|---|
This paper will show an alternative method to compute the two-dimensional Discrete Fourier Transform. While current GPU Fourier transform libraries need a large buffer for storing intermediate results, our method can compute the same output with far less memory. This will function by exploiting the separability of the Fourier transform. Using this scheme, it is possible to transform rows and columns independently. As multiple lines can be transformed at once, the available memory on the device can be used to reduce the number of necessary kernel calls drastically. We will also prove that our approach can compete with the timings of the two-dimensional transform provided by NVIDIAs CUFFT library but consumes at the same time far less memory by enabling the transformation of much bigger data sets on the device. |
Year | DOI | Venue |
|---|---|---|
2010 | 10.1109/CIT.2010.209 | Computer and Information Technology |
Keywords | Field | DocType |
nvidias cufft library,multiple line,current gpu fourier,necessary kernel,bigger data set,large buffer,alternative method,available memory,intermediate result,two-dimensional discrete fourier transform,memory management,gpgpu,discrete fourier transform,coprocessors,planning,graphics,fourier transform | Cyclotomic fast Fourier transform,Computer science,Parallel computing,Fourier transform,Real-time computing,Computational science,Fast Fourier transform,Memory management,Discrete Hartley transform,Discrete Fourier transform,S transform,Pseudo-spectral method | Conference |
ISBN | Citations | PageRank |
978-1-4244-7547-6 | 1 | 0.39 |
References | Authors | |
6 | 4 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Daniel Kauker | 1 | 14 | 2.46 |
| Harald Sanftmann | 2 | 20 | 3.12 |
| Steffen Frey | 3 | 116 | 18.93 |
| Thomas Ertl | 4 | 4417 | 401.52 |