Title :
The radix-r one stage FFT kernel computation
Author :
Jaber, Marwan A. ; Massicotte, Daniel
Author_Institution :
Electr. & Comput. Eng. Dept., Univ. du Quebec a Trois-Rivieres, Trois-Rivieres, QC
fDate :
March 31 2008-April 4 2008
Abstract :
The FFT process is an operation that could be performed through different stages. In each stage, the butterfly operation is computed in which the accessed data is multiplied by certain Walpha, added or subtracted and finally it is stored or held for further processing. This process is repeated to each stage until the final stage where the processed data is driven to the output. In this paper, an appropriate indexing or mapping schemes between the input data and the coefficient multipliers throughout the different stages are yield to a computation single stage by collapsing all stages into a computation single stage. The result is a reduction of communication load and arithmetic operations.
Keywords :
fast Fourier transforms; parallel processing; arithmetic operations; butterfly operation; coefficient multipliers; communication load; indexing schemes; mapping schemes; radix-r one stage FFT kernel computation; Algorithm design and analysis; Discrete Fourier transforms; Frequency domain analysis; Kernel; Laboratories; Mathematical model; Signal analysis; Signal processing; Signal processing algorithms; Speech analysis; Discrete Fourier transforms; Frequency domain analysis; Parallel processing;
Conference_Titel :
Acoustics, Speech and Signal Processing, 2008. ICASSP 2008. IEEE International Conference on
Conference_Location :
Las Vegas, NV
Print_ISBN :
978-1-4244-1483-3
Electronic_ISBN :
1520-6149
DOI :
10.1109/ICASSP.2008.4518427