A Data-flow Methodology for Accelerating FFT

The N-point digital Fourier transform involves hard the real number of the sample buffer with N separate basis vectors. Since every real number involves N multiplications and N additions, the whole time is proportional to N2, in different words, it’s an O(N2) algorithmic rule. However, it turns out that by smartly re arranging these operations, one will optimize the algorithmic rule all the way down to O(Nlog2(N)), that for large N makes a large difference. The optimized version of the algorithmic rule is named the fast Fourier transform, or the FFT. During this paper, we tend to discuss regarding an efficient way to obtain Fast Fourier transform algorithm (FFT). According to our study, we are able to eliminate some operations in hard the FFT algorithmic rule due to property of complex numbers and that we are able to do the FFT during a higher execution time because of a major reduction of the required twiddle factors and to extra factorizations