Research on Heterogeneous Parallel Algorithms for Radar Signal Processing in OpenCL Platform

An efficient radar signal processing parallel algorithm for Open Computing Language (OpenCL) is presented in this paper, which reveals a rational allocation schema of tasks between the central processing unit (CPU) and the graphics processing unit (GPU). The powerful parallel processing ability of GPU is used to replace the traditional serial processing method by the way of pipeline acceleration and algorithm parallel acceleration, and the real-time performance of the proposed algorithm is improved by the design idea of intra-module parallelism and inter-module flow processing. The algorithm is applied to CPU-GPU heterogeneous computing platform and has a good portability. The effectiveness and real-time performance of the algorithm are verified through simulation data.