Write Pow Function Without math.h in C

Question

Hi I want to write a code for a pow function without using math.h libary.

How is this code?how can i fix it when b<0

    int MyPow(int a,int b){
      if(b<0)      
        return 1 / MyPow (a,-b)
      else if(b==0)
        return 1;
      else if(b==1)
        return a;
      else
        return a*MyPow(a,b-1)
    }

that's not the issue... what he means is 'What is this code doing that deviates from your expectations?', 'What does it return, and what should it return', that sort of thing. — Psychemaster, Aug 27 '14 at 11:09
@Omer math.h is a standard header, i wonder why it was not available ? — TerryG, Aug 27 '14 at 11:11
I think you should check `if (a==0) { return 0; }` and `if (b==0||a==1) { return 1; }` and `if (b==1) { return a; }` and finaly `return myPow(a*a, b/2) * ((b%2==0) ? 1 : a);` — Jayesh Bhoi, Aug 27 '14 at 11:12
@Psychemaster I just asked if its good function of pow and how can I upgrade this function — Omer, Aug 27 '14 at 11:12
As a suggestion try exponentiation by squaring algorithm, see http://en.wikipedia.org/wiki/Exponentiation_by_squaring. It's much faster `O(log2(n))` than "naive" method. This may teach you recursion as well. — Grzegorz Szpetkowski, Aug 27 '14 at 11:12
This question appears to be off-topic because it is about code review and should be migrated to http://codereview.stackexchange.com/ — bitmask, Aug 27 '14 at 11:20
@Omer-kindly re-visit my code! I have updated it as it contained a minor flaw! Please re-edit your code! — Am_I_Helpful, Aug 27 '14 at 11:27
If `b < 0`, then the result isn't `int`. Pow(1, -2) is 0.25. So without changing the return type it isn't really possible to return anything useful for `b<0`. — Klas Lindbäck, Aug 27 '14 at 11:28
Possible duplicate of [How can I write a power function myself?](http://stackoverflow.com/questions/2882706/how-can-i-write-a-power-function-myself) — jww, Aug 27 '14 at 11:59
@KlasLindbäck The original code had no branch for `b<0` so it would most likely have caused a stack overflow (tail recursion optimisation is not possible here!) — Erich Kitzmueller, Aug 27 '14 at 12:34

score 3 · Answer 1 · edited May 23 '17 at 10:25

3

Probably best way I think, found here

int pow(int base, int exp)
    {
      if(exp < 0)
        return -1;

        int result = 1;
        while (exp)
        {
            if (exp & 1)
                result *= base;
            exp >>= 1;
            base *= base;
        }

        return result;
    }

edited May 23 '17 at 10:25

Community

1
1

answered Aug 27 '14 at 11:21

Suryakant Sharma

3,852
1
25
47

1

Good performance, since complexity is `O(log(exp))`. Note that it doesn't handle cases where `exp < 0`. – Klas Lindbäck Aug 27 '14 at 11:24
Yes you are right @KlasLindbäck. Any suggestion to make it work for `exp < 0` without compromising complexity? – Suryakant Sharma Aug 27 '14 at 11:32
2

As long as the return type is `int`, using a negative exponent is meaningless since the result will be truncated to 0. Otherwise, the normal solution to a negative exponent is: `return 1.0/pow(base, -exp)` – Klas Lindbäck Aug 27 '14 at 11:37
actually I was asking for return type double...thanks that helped – Suryakant Sharma Aug 27 '14 at 11:47

Am_I_Helpful · Accepted Answer · 2014-08-27T12:29:56.517

1

Everything seems perfect except for one condition :- when b<0.

For b<0,simply return

return (1.0/a)*MyPow(a,abs(b)-1);   //where  abs(b) is  absolute value of b.

OR

return (1.0/a)*(MyPow(a,b+1));

Also,your definition of function is not valid for performing negative exponentiation,you should change it to

float MyPow(int a,int b)

edited Aug 27 '14 at 12:29

answered Aug 27 '14 at 11:11

Am_I_Helpful

18,735
7
49
73

2

Why not just 1 / MyPow (a, -b) ? – gnasher729 Aug 27 '14 at 11:13
3

`1/MyPow(a,-b)` will be `0` almost always (unless `b==1`, and then it would be `1`). – bitmask Aug 27 '14 at 11:18
2

@shekharsuman: Because all involved values have type `int`. Integer division of `1` by `something` will return `0` if `something` is anything but `1` or `0`. – bitmask Aug 27 '14 at 11:23
Ahh!!! @bitmask--THANKS,I couldn't believe I have committed this mistake! THANKS again for reminding... – Am_I_Helpful Aug 27 '14 at 11:25
1

Still his function returns int, so it still will be 0 for negative b... – zoska Aug 27 '14 at 11:46
1

Yeah,he obviously needs to change the return type of his function from int to float---it's obvious! He'll do/manage that. – Am_I_Helpful Aug 27 '14 at 11:54
@gnasher729-I guess your suggestion couldn't be recursively implemented to obtain negative power in this case.If I am wrong,please clarify... – Am_I_Helpful Aug 27 '14 at 12:31

score 0 · Answer 3 · answered Aug 27 '14 at 11:13

A solution with less complexity taken from http://www.geeksforgeeks.org/write-a-c-program-to-calculate-powxn/

float power(float x, int y)
{
    float temp;
    if( y == 0)
       return 1;
    temp = power(x, y/2);       
    if (y%2 == 0)
        return temp*temp;
    else
    {
        if(y > 0)
            return x*temp*temp;
        else
            return (temp*temp)/x;
    }
}

Write Pow Function Without math.h in C

3 Answers3

Linked