引言DSP结构可以分为定点和浮点型两种。其中,定点型DSP可以实现整数、小数和特定的指数运算,它具有运算速度快、占用资源少、成本低等特点;灵活地使用定点型DSP进行浮点运算能够提高运算的效率。目前对定点DSP结构支持下的浮点需求也在不断增长,主要原因是:实现算法的代码往往 Introduction DSP structure can be divided into fixed-point and floating-point two. Among them, the fixed-point DSP can achieve integer, decimal and specific index operations, it has the advantages of fast operation, less resource consumption, low cost; flexible use of fixed-point DSP floating-point operations can improve the efficiency of the operation. The current demand for floating-point support for fixed-point DSP architectures is also growing, mainly due to the fact that the code that implements the algorithm tends to