Title :
Parallel implementation of 1-D fast Fourier transform without inter-processor communications
Author :
Al Na´mneh, R. ; Pan, D.W. ; Adhami, R.
Author_Institution :
Dept. of Electr. & Comput. Eng., Alabama Univ., Huntsville, AL, USA
Abstract :
Computing 1-D fast Fourier transform (FFT) using the classical 4-step FFT on parallel computers requires intensive all-to-all communication. This all-to-all communication significantly reduces the performance of FFT. In this paper, we present the no-communication algorithm that is a parallel algorithm for 1-D FFT without inter-processors communication. The advantage of this algorithm is the absence of all-to-all communication between processors. The disadvantage of this algorithm is the extra computation compared to the classical 4-step FFT. The no-communication algorithm has been implemented and tested in 8-node symmetric multiprocessors (SMP). The results show that the no-communication algorithm performs better than the 4-step FFT for relatively small data sizes. However, 4-step FFT algorithm performs better than the no-communication for relatively large data sizes.
Keywords :
fast Fourier transforms; multiprocessing systems; parallel machines; 1D fast Fourier transform parallel implementation; 8-node symmetric multiprocessors; classical 4-step FFT; no-communication algorithm; parallel computers; Central Processing Unit; Concurrent computing; FETs; Fast Fourier transforms; Image analysis; Image processing; Parallel algorithms; Scalability; Signal processing algorithms; Testing;
Conference_Titel :
System Theory, 2005. SSST '05. Proceedings of the Thirty-Seventh Southeastern Symposium on
Print_ISBN :
0-7803-8808-9
DOI :
10.1109/SSST.2005.1460927