Design of Floating Point Division Unit Based on X87 Instruction Set
Based on the X87 instruction set architecture and TSMC 65nm process library,this paper uses SRT-16 algorithm to design the floating-point division unit.SRT-16 algorithm is mainly optimized by using the structure of overlapping quotient and remainder calculation part,so that the circuit area increased on the basis of the original traditional algorithm SRT-4 is less,and the number of cycles is doubled.The fly conversion method of quotient is introduced into the result of processing quotient.Through the construction of the automatic comparison platform based on System Verilog and the C comparison model,the auto-matic comparison is completed,the verification speed is accelerated,the function coverage statistics are carried out,and the coverage rate is 100％.The timing of the circuit is comprehensively detected by DC,and the timing of the circuit is optimized.It meets the timing requirements of 1.08 ns and the working frequency is 900 MHz.
万方数据
维普
