Square-root computer capable of reducing error

Abstract

A SQRT calculator capable of calculation with a minimal error is provided. The integer calculation unit selects a largest integer from a set of integers with a square of each of the integers smaller than an input datum. The transformation unit transforms the selected integer from the integer calculation unit by multiplying it by 2 and shifts a decimal point of the resulting number to the right by 1 place, thereby adding a certain number less than 10 to the decimal point shifted number to calculate a transformation value. The calculation unit shifts a decimal point of the number less than 10 to the left by 2 places and multiplies the transformation value by the resulting value, thereby subtracting the multiplied value from the input datum and choosing a largest number less than 10 with the subtracted value being in a desired range as a second decimal number of the square root. Thus, the SQRT calculator is capable of calculation with minimal error and, furthermore, has a minimum size of hardware.

Description

technical field [0001] The present invention relates to a SQRT (square root) calculator, in particular to a SQRT calculator capable of reducing calculation amount and having a simple hardware structure. [0002] The present invention is based on Korean Patent Application No. 2002-7753, which is incorporated herein by reference. Background technique [0003] figure 1 It is a schematic diagram of the hardware structure of a traditional seed-type SQRT calculator. The SQRT calculator 10 includes a selection signal output unit 11 , a slope (slope) storage unit 13 , a Y-axis intercept value (y-intercept) storage unit 15 , a multiplier 17 and an adder 19 . [0004] The selection signal output unit 11 transmits the slope selection signal and the Y-axis intercept value selection signal to the slope storage unit 13 and the Y-axis intercept value storage unit 15 respectively to select the slope and the Y-axis intercept value, which are required for calculating the square root of the ...

AI Technical Summary

Problems solved by technology

[0020] To minimize these errors, a conventional SQRT calculator needs to reduce the spacing of its subregions, which will result in an increase in the number of coefficients that need to be stored

When the square root is calculated using parallel processing for each sub-area divided according to the conventional SQRT calculator, the calculation speed will be accelerated, but the size of the hardware will increase because the processing of the calculation of each sub-area requires its own unit, such as a multiplier and an addition device

[0021] On the other hand, the method using the Newton-Raphson algorithm has disadvantages in that many repeated calculations are required to obtain accuracy, and additional dividers and differentiators are required in addition to the multipliers, and the size of the hardware will increase.

Images