Sine, cosine and arc tangent function operation device and method based on CORDIC algorithm

Abstract

The invention discloses a sine and cosine and arc tangent function operation device and method based on a CORDIC algorithm, and the device comprises an exception processing module which detects whether an input radian value is abnormal, and outputs the input radian value to a preprocessing module if the input radian value is legal; a preprocessing module which is used for converting the radian value input by the exception processing module into a range which can be calculated by a CORDIC algorithm and inputting the radian value into the four-step parallel iteration module; a four-step parallel iteration module which is used for realizing four-step parallel iteration operation based on the CORDIC algorithm, specifically, carrying out X-channel, Y-channel and Z-channel iteration operation on the data obtained by the preprocessing module, and predicting symbol factors; and a standardized floating point normalization module which is used for carrying out leading zero detection on the data obtained by the four-step parallel iteration module, and then normalizing and combining the data into a standard floating point form for outputting. According to the method, the expressible precision of the sine and cosine and arc tangent function calculation results is improved, and meanwhile, the number of iterations needed by calculation is greatly reduced.

Description

technical field [0001] The invention relates to the technical field of sine-cosine and arctangent function calculations, in particular to a sine-cosine and arctangent function calculation device and method based on a CORDIC algorithm. Background technique [0002] Calculation of sine-cosine and arctangent functions are quite common transcendental function calculations in scientific computing. At present, the main hardware design schemes mainly include look-up table method, polynomial approximation method and CORDIC (COordinate RotationDIgital Computer, coordinate rotation digital computer) method; With the improvement of precision requirements, the hardware resource consumption of the look-up table method and the polynomial approximation method increases exponentially; the traditional CORDIC algorithm can calculate trigonometric functions by shifting and adding, but it will generate a huge calculation delay under high precision requirements. At present, the calculation preci...

AI Technical Summary

Problems solved by technology

[0002] Calculation of sine-cosine and arctangent functions are quite common transcendental function calculations in scientific computing. At present, the main hardware design schemes mainly include look-up table method, polynomial approximation method and CORDIC (COordinate RotationDIgital Computer, coordinate rotation digital computer) method; With the improvement of precision requirements, the hardware resource consumption of look-up table method and polynomial approximation method increases exponentially; the traditional CORDIC algorithm can calculate trigonometric functions by shifting and adding, but it will generate a huge calculation delay under high precision requirements

At present, the calculation accuracy of sine-cosine and arctangent functions mainly includes fixed-point, single-precision floating-point and double-precision floating-point, and there are few researches on the calculation hardware implementation of 128-bit floating-point

Images