Optimization of FFT Parallel Algorithm on Multi-Core CPUs

Abstract

Fast Fourier Transformation is widely applied in fields of digital signal processing and communication.Multi-core CPU is increasingly common,and based on the parallelism of FFT algorithm,butterfly algorithm can be flexibly decomposited;we can improve the computational efficiency of FFT via investigating the allocation of parallel blocks and nested relations to optimize the algorithm,and reasonably allocating threads to realize parallel computing of multi-core CPU.