Approximate calculation device for sigmoid function

Abstract

The invention discloses an approximate calculation device for a sigmoid function. The approximate calculation device comprises a negative number input conversion unit, a constant multiplication unit,a shift unit, a special value generation unit, a carry-save adder and a negative number result conversion unit. The negative number input conversion unit is used for taking an absolute value of an input negative value x and outputting a binary source code of the absolute value; the constant multiplication unit calculates a value of 1.4375 * x and outputs an integral part k and a decimal part p ofan operation result; the shifting unit is used for shifting the input p; the special value generation unit is used for generating an approximate value lambda = sigmoid (kln2); the carry-save adder isused for realizing addition of three numbers; and the negative result conversion unit realizes a conversion function of a corresponding result when the input x is a negative number. According to the device, approximate calculation of the sigmoid function can be achieved, and while high approximate precision is kept, operation delay and power consumption are greatly reduced, and area expenditure isreduced.

Description

technical field [0001] The invention belongs to the field of VLSI design, in particular to an approximate calculation device for sigmoid function. Background technique [0002] The Sigmoid function is a nonlinear function widely used in digital signal processing and deep neural networks. Due to the non-linear characteristics of the Sigmoid function, it is difficult to implement it quickly on hardware and maintain high fitting accuracy. There have been many sigmoid function fitting methods in the past, but these methods require more hardware resource overhead, which is not conducive to hardware implementation. [0003] Currently, approximation methods for piecewise linear fitting are widely used in FPGA or VLSI implementations. Although this method can have high fitting accuracy, it still needs to use a lookup table to store the parameters of the piecewise function, and needs to use a multiplier with high delay and large area. Therefore, this method not only requires more ...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to overcome the shortcomings of large hardware resource overhead and long delay required by the existing sigmoid function approximation method, and provide a high-speed, low power consumption and area-saving computing device for the sigmoid function

Images