A software for evaluating local accuracy in the fourier transform

With the evolution of computers toward ever increasing computing speed, the numerical algorithms implemented on such computers are more and more complex, and require a very great number of computations. The numerical errors inherent in floating-point arithmetic of the computer generate errors in the results which may be fairly great. In this paper we propose a software based on the Perturbation Method for estimating the local accuracy in the Fast Fourier Transform, both in the case of exact data (computing errors alone) and in the case of experimental data (data errors and computing errors).