摘要
快速傅立叶变换 (FFT)运算速度相比离散傅立叶变换 (DFT)提高了 1至 2个数量级 ,但当离散信号的有限序列太长时 ,FFT算法同样也无法满足实时系统的需要。本文提出一种基于共享存储的多处理机计算FFT的并行算法 ,并给出该算法的性能分析。计算机仿真实验表明 ,该FFT并行算法有效、高速 ,适合于实时处理系统。
Compared to Discrete Fourier Transform(DFT),Fast Fourier Transform(FFT)is improved by 1 to 2 steps in speed,but when the sequence length of discrete signal is too large,FFT can not satisfy the requirements of the real-time system.A parallel algorithm for FFT on shared-memory multi-computer system is detailed in this paper,and its performance is analyzed.Computer emputer emulation shows that this FFT parallel algorithm is efficient,high speed,and also adapted for real-time process system.
出处
《航空计算技术》
2000年第3期5-7,共3页
Aeronautical Computing Technique
基金
北京大学视觉与听觉信息处理国家重点实验室开放课题 !(N990 1Z51 )资助
