基于CORDIC算法的高精度浮点超越函数的FPGA实现

The implementation of high-precision floating transcendental functions on FPGA based on the CORDIC algorithm

提出了一种新的输入输出浮点处理单元硬件架构,将数据从CORDIC算法内部格式转换为处理器能够支持的IEEE754标准浮点数据格式。输入数据支持2种不同的角度单位浮点数据直接输入,同时,硬件模块还直接支持超过360°的大角度数据输入。在Altera公司NiosⅡ处理器系统中以用户自定义指令的形式实现了该浮点硬件计算模块,并通过C语言程序验证了该模块的正确性。

This paper presents a novel architecture of input and output floating-point execution unit,which can converts data from the format of CORDIC algorithm system to the format accepted by processors.Two different units of angle can be directly supported by the floating-point CORDIC module, the wide angle even above 360° also can be directly supported by the hardware module. Finally,we accomplish the hardware module by adding custom instruction on the Nios II processor system of Altera corpora- tion.The floating-point CORDIC module＇s correctness is proved by C program running in the Nios II processor.