Evaluation of libraries for the one-dimensional sparse Fourier transform

Computing Sparse Fourier transforms is becoming popular in many applications working with large signals that are sparse in the spectral domain. This poster compares the performance of several C++ libraries computing the Sparse Fourier transform including MSFFT, AAFFT, DMSFT and GFFT. The proposed benchmark consists of random signals with various length and sparsity as well as signals received from one dimensional ultrasound simulation.  As the reference, the FFTW library, widely used library for computing the Fast Fourier transform, is used.