12,943,367 members (79,682 online)
Technical Blog
alternative version

#### Stats

40.6K views
16 bookmarked
Posted 1 Apr 2013

# Square Root algorithm for C

, 21 Aug 2014 CPOL
 Rate this:
Square Root algorithm for C

## Introduction

For those who are looking for how square root algorithm works and “YES !, Welcome to my article”, but for those who are simply looking for performing square root on your work  and I have to tell you, what you have to do is some Googling on “math.h” header file to look for square root function that prepared inside it. Well, let’s begin our lesson “How square root algorithm works on C”. It’s not very difficult once you understand how the following equation is calculated.

The left hand side of the diagram shows the square root of 152.2756 and the right hand side shows the square root of 2. For those who understand how this calculation has been done, please skip to the code as I am going to explain how to solve this equation.

### Basic Principle of Square Root

Based on the question given above, we understand that square root of 152.2756 is 12.34 and the square root of 2 is 1.4142. If you do not believe, try calculator to find the answer.
Suppose we are finding the square root of 2 and the answer is 1.4142. We can express it such that:

If we expressed it in an algebra expression as follows:

Suppose a is continued n times, it can be expressed as follows:

Why I am treating it continue as n times but not 5 times? Because we cannot ensure that all positive numbers will give us a 5 digits answer after square root it. So we have to suppose there are n digits after square root. Right-hand-side term can be expanded as follows:

From the equation above, we can understand square root equation can be solved by following equation:

The most important rule of this equation is we pick a number from 1,2,3,4,5,6,7,8,9 and multiple it with 10^m , where m=Integer (exp. 3,2,1,0,-1,-2,...) and substitute into:

that must be smaller than the number we are going to minus or equal to it. It’s the most difficult part on square-root equation so let's understand it by solving square root of 2.

So, from the solution above, we understand that:

And we sum up all the values, we get the answer 1.414 approximate to 1.4142.

### Algorithm

Based on the method above, what kind of work should we apply on algorithm as follows:

• Create equation
• Create function of power of ten
• Find suitable number that is bigger than 0 for variable a from loop
• Find suitable number that is smaller than 0 but positive for variable a from loop
• Sum all variable a and return it as `double`

First, we are going to create the equation as we mentioned above:

We can express it as code as follows:

`(( 2 * rst ) + ( j * powerOfTen(i)))*( j * powerOfTen(i))`

where rst and j is declared as `double `and i is `int`. `powerOfTen `is a function that return a value of `i` times multiple of `10`.

It’s code of function `powerOfTen`:

```1:
2: double powerOfTen(int num){
3:     double rst = 1.0;
4:     if(num >= 0){
5:         for(int i = 0; i < num ; i++){
6:             rst *= 10.0;
7:         }
8:     }else{
9:         for(int i = 0; i < (0 - num ); i++){
10:            rst *= 0.1;
11:        }
12:    }
13:
14:    return rst;
15: }```

We have to judge that variable a must be smaller than upper number. It’s how it looks like:

`if(z - (( 2 * rst ) + ( j * powerOfTen(i)))*( j * powerOfTen(i)) >= 0)`

where z  is declared as `double `and treated as upper number.

There are two loops, one loop for finding the suitable number from 10000,1000,100,10… and another one loop for find suitable number from 1,2,3,4,5…

```1: for(i = max ; i > 0 ; i--){
2: // value must be bigger then 0

3: if(z - (( 2 * rst ) + ( j * powerOfTen(i)))*( j * powerOfTen(i)) >= 0)
4: {
5:     while( z - (( 2 * rst ) + ( j * powerOfTen(i)))*( j * powerOfTen(i)) >= 0)
6:     {
7:         j++;
8:         if(j >= 10) break;
9:
10:    }
11:    j--; //correct the extra value by minus one to j

12:    z -= (( 2 * rst ) + ( j * powerOfTen(i)))*( j * powerOfTen(i)); //find value of z

13:
14:    rst += j * powerOfTen(i);     // find sum of a

15:
16:    j = 1.0;
17:
18:
19: }
20:
21: }```

This is the same loop as above but this time, we are going to find the decimal number.

```1: for(i = 0 ; i >= 0 - max ; i--){
2: if(z - (( 2 * rst ) + ( j * powerOfTen(i)))*( j * powerOfTen(i)) >= 0)
3: {
4:     while( z - (( 2 * rst ) + ( j * powerOfTen(i)))*( j * powerOfTen(i)) >= 0)
5:     {
6:         j++;
7:         if(j >= 10) break;
8:     }
9:     j--;
10:    z -= (( 2 * rst ) + ( j * powerOfTen(i)))*( j * powerOfTen(i)); //find value of z

11:
12:    rst += j * powerOfTen(i);     // find sum of a

13:    j = 1.0;
14:    }
15: }```

This is how the whole `squareRoot `algorithm looks like:

```1: double squareRoot(double a)
2: {
3:     /*
4:           find more detail of this method on wiki methods_of_computing_square_roots
5:
6:           *** Babylonian method cannot get exact zero but approximately value of the square_root
7:      */
8:      double z = a;
9:      double rst = 0.0;
10:     int max = 8;     // to define maximum digit

11:     int i;
12:     double j = 1.0;
13:     for(i = max ; i > 0 ; i--){
14:         // value must be bigger then 0

15:         if(z - (( 2 * rst ) + ( j * powerOfTen(i)))*( j * powerOfTen(i)) >= 0)
16:         {
17:             while( z - (( 2 * rst ) + ( j * powerOfTen(i)))*( j * powerOfTen(i)) >= 0)
18:             {
19:                 j++;
20:                 if(j >= 10) break;
21:
22:             }
23:             j--; //correct the extra value by minus one to j

24:             z -= (( 2 * rst ) + ( j * powerOfTen(i)))*( j * powerOfTen(i)); //find value of z

25:
26:             rst += j * powerOfTen(i);     // find sum of a

27:
28:             j = 1.0;
29:
30:
31:           }
32:
33:      }
34:
35:      for(i = 0 ; i >= 0 - max ; i--){
36:          if(z - (( 2 * rst ) + ( j * powerOfTen(i)))*( j * powerOfTen(i)) >= 0)
37:          {
38:              while( z - (( 2 * rst ) + ( j * powerOfTen(i)))*( j * powerOfTen(i)) >= 0)
39:              {
40:                  j++;
41:
42:              }
43:              j--;
44:              z -= (( 2 * rst ) + ( j * powerOfTen(i)))*( j * powerOfTen(i)); //find value of z

45:
46:              rst += j * powerOfTen(i);     // find sum of a

47:              j = 1.0;
48:           }
49:      }
50:      // find the number on each digit

51:      return rst;
52: }```

Thank you.

## Share

 Japan
Hi! Thank you everyone who reading my article. My major is electronic and programming. Right now I am doing foreign study at Japan. I will like sharing to everyone with my works and if you do interesting with my works, please leave you comment on my blog. Any comments is welcoming.

## You may also be interested in...

 First Prev Next
 Nice kutynko2-Apr-13 5:24 kutynko 2-Apr-13 5:24
 Nice article. By the way, you have a typo in eq6 figure. In second block it says that (2*a1+a2)*a1=(2*1+0.4)*0.4 but should be (2*a1+a2)*a2=....
 Re: Nice Edison Heng14-Apr-13 20:12 Edison Heng 14-Apr-13 20:12
 Last Visit: 31-Dec-99 18:00     Last Update: 22-May-17 10:45 Refresh 1