12970

cufftShift: High Performance CUDA-accelerated FFT-shift Library

Marwan Abdellah
Ecole Polytechnique Federale de Lausanne (EPFL), Switzerland
22nd High Performance Computing Symposium (HPC ’14), 2014

@article{abdellah2014cufftshift,

   title={cufftShift: High Performance CUDA-accelerated FFT-shift Library},

   author={Abdellah, Marwan},

   year={2014}

}

For embarrassingly parallel algorithms, a Graphics Processing Unit (GPU) outperforms a traditional CPU on price-per-flop and price-per-watt by at least one order of magnitude. This had led to the mapping of signal and image processing algorithms, and consequently their applications, to run entirely on GPUs. This paper presents CUFFTSHIFT, a ready-to-use GPU-accelerated library, that implements a high performance parallel version of the FFT-shift operation on CUDA-enabled GPUs. Compared to Octave, CUFFTSHIFT can achieve up to 250x, 115x, and 155x speedups for one-, two- and three dimensional single precision data arrays of size 33554432, 81922 and 5123 elements, respectively. The library is designed to be compatible with the CUFFT library, which lacks a native support for GPU-accelerated FFT-shift operations.
Rating: 3.3/5. From 3 votes.
Please wait...

* * *

* * *

HGPU group © 2010-2024 hgpu.org

All rights belong to the respective authors

Contact us: