A Hardware Modeling of High Precision Floating Point Arcsine/Arccosine Function Based on CORDIC Algorithm

Fast and accurate calculation of floating-point arcsine/arccosine function has important applications in modern engineering. In order to improve the precision and calculation speed of the arcsine/arccosine function, this paper proposes an improved four-step rotation prediction iteration method based on the CORDIC algorithm, which completes the floating-point operation only through addition, subtraction and shift operations, and solves the problem of excessive iterations of the traditional CORDIC algorithm, at the same time, meets the floating point rules of the IEEE754 standard. This design is written in Verilog, and the Modelsim platform is used to verify and simulate the design. Only 32 iterations are required to complete a quadruple-precision arcsine/arccosine function calculation. Using TSMC's 65nm technology and 500MHz clock, the total design area is about 0.8mm
²
, only 32 clock cycles are required to complete the quadruple-precision arcsine/arccosine function. The results show that the algorithm has the advantages of fast data processing speed and higher precision.