Facebook has written a Fast Fourier Transform  (fbfft) that is 1.5x faster than the NVIDIA CUFFT implementation at sizes 8-64. The paper "Fast Convolutional Nets with fbfft: A GPU Performance Evaluation" discusses the performance increases by changing to a non-zero padded FFT layout (potentially eliminating data copies), the use of autotuning, and clipping to conditionally load … [Read more...]