Institutional Repository of Vilnius University
Lygiagretizuota greitoji Furje transformacija naudojant CUDA /
Abstract
dc:descriptionAlgorithms that allow calculating discrete Fourier transform faster, also known as Fast Fourier transform algorithms, are explorer. Mathematical properties selected radix-2 Cooley-Tukey and radix-2, radix-4, radix-8, radix-16 Stockham algorithms are described. This is accompanied by examination of CUDA parallel programming model and its practices with regard to implementation optimization. Lastly, CUDA enabled parallel algorithm versions and their implementations are described. These implementations are compared both with one another, as well as with FFTW and CuFFT libraries. Out of the created implementations radix-16 Stockham stands out as the most performant algorithm, outperforming FFTW and being close in terms of speed to CuFFT.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2022
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Prievelis, Dominykas,
Rights
dc:rights- Statement dc:rights
-
- info:eu-repo/semantics/openAccess
- Language dc:language
- lit
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.vu.lt/VU:ELABAETD146242270&prefLang=en_US