3D FFTs on a Single FPGA

The 3D FFT is critical in many physical simulations and image processing applications. On FPGAs, however, the 3D FFT was thought to be inefficient relative to other methods such as convolution-based implementations of multigrid. We find the opposite: a simple design, operating at a conservative frequency, takes 4 mu s for 16(3), 21 mu s for 32(3), and 215 mu s for 64(3) single precision data points. The first two of these compare favorably with the 25 mu s and 29 mu s obtained running on a current Nvidia GPU. Some broader significance is that this is a critical piece in implementing a large scale FPGA-based MD engine: even a single FPGA is capable of keeping the FFT off of the critical path for a large fraction of possible MD simulations.