REAL FOURIER-TRANSFORMS ON A MASSIVELY PARALLEL MACHINE

In this paper we compare two implementations for computing Fourier transform of real data on a massively parallel SIMD machine (DAP-510). The first implementation is based on the Real Fast Fourier Transform (RFFT) algorithm, and the second is based on the Fast Hartley Transform (FHT) algorithm. On a sequential computer it has been shown that both the RFFT and FHT algorithms are faster than the Fast Fourier Transform (FHT) algorithm for computing Fourier transform of real data. However, it is not obvious that the same is true on a parallel machine. The communication requirements of the RFFT and the FHT algorithms, which are critical to the cost of any parallel implementation, are different from those of the FFT algorithm. We show that both the implementations have identical performances and are superior to the FFT implementation.