支持超越函数的浮点运算单元的设计与实现  

Design and implementation of floating-point operation unit with transcendental function

作　　者：付江平[1] 张盛兵[1] 高德远[1] 郭亚鹏[1] 

机构地区：[1]西北工业大学 计算机学院,西安710065

出　　处：《计算机应用研究》2007年第9期179-181,共3页Application Research of Computers

基　　金：国家自然科学基金(60573101)

摘　　要：介绍自主设计的龙腾C2微处理器中浮点运算单元的设计与实现。该处理器与Intel 80486DX4指令系统兼容,支持IEEE 754标准扩展精度的浮点基本函数和超越函数运算。介绍了浮点运算单元的结构,分析了实现超越函数的高精度CORDIC算法的流程,讨论了实现浮点超越函数运算的数据通路和控制通路结构,并给出了仿真结果和精度评估结果。仿真和分析的结果表明,浮点运算单元的设计满足龙腾C2微处理器的设计要求。Longtium C2 is compatible with Intcl 80486DX4 instruction set, and includes floating-point operation on extendedprecision format which belongs to 1EEE standard 754, and also includes the transcendental function. The paper introduced the structure of floating-point operation unit firstly, and then analyzed the flow of HP CORD1C algorithm used for implementing the transcendental function, and discussed the corresponding data path and control path, and also showed the simulation results and precision evaluation results at the end. The results suggest the precision meets the requirement of Longtium C2 microprocessor.

关 键 词：浮点运算单元 超越函数 坐标旋转数字计算机算法 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]