Performance analysis and acceleration of cross-correlation computation using FPGA implementation for digital signal processing
Rafał Selma · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · 2016
Paper describes comparison of cross-correlation computation speed of most commonly used computation platforms (CPU, GPU) with an FPGA-based design. It also describes the structure of cross-correlation unit implemented for testing purposes. Speedup of computations was achieved using FPGA-based design, varying between 16 and 5400 times compared to CPU computations and between 3 and 175 times compared to GPU computations.