Thread: Recursive Power function

Code:

#include <stdio.h>
#define BAD -1
int Power(int x, int n);    

main()
{
        Power(3,20); 
}

int Power(int x, int n)     
{
        int middle;

        if(n < 0) exit(BAD);
        else if(n == 1) return x;
        else if(n == 0) return 1;  

        else if( (n % 2) == 0) // even
        {
                middle = n/2;
                printf("In even middle = %d\n", middle);
                return Power(x, middle) * Power(x, middle);
        }

        else
        if( (n % 2) == 1) //odd
        {
                middle = n/2;
                printf("In odd middle = %d\n", middle);
                return Power(x, middle) * Power(x, middle) * x;
         }


}

I am trying to get it to do something like this, if it is (x^10) it will do (x^5)*(x^5), and if x^11 it will do (x^5)*(x^5)*x.

It works for Power(3, (0->4)), but once I make the power higher than 4 or 5 it will give me the incorrect answer. I dont know why it works up to a limit of numbers, and then starts to give me wrong answers.

Thanks.

It works fine for the most part. The reason why it's not working sometimes is because for example, some values are too large to fit in an integer. That's why your results will appear incorrect.

Originally posted by beege31337
It's working on my compiler.

Really, what does Power(3,10) return for you?

I never print out the result, on a unix server I use echo $? to see what the program returns.

I tried it on my home computer, on my Borland compiler and it gave the correct answer. Dont know what was happening before.

Thanks for the help!

Originally posted by MethodMan
I tried it on my home computer, on my Borland compiler and it gave the correct answer. Dont know what was happening before.

My Borland compiler (5.5) has an int which is the size of a long int, so the max number you fit in it is quite large. You can compare your 2 compilers by printing something like this to see if that's your problem:

Code:

#include <limits.h>
...
printf ("INT_MAX is %d\n", INT_MAX);

Originally posted by Shiro
Also return 0 is not always correct, since some OS'es use 0 as representation of false. Better would be to use the constants as defined in the ANSI C standard:

EXIT_SUCCESS and EXIT_FAILURE

ansi C says 0 is synonomous for EXIT_SUCCESS

Originally posted by Shiro
My copy of the ANSI C standard says (page 315):



Note the _or_. There are OS'es on which EXIT_SUCCESS is implemented as 1, therefore it is recommended to use EXIT_SUCCESS for portability.

or means or, no more, no less. 0 is synomous with success, period. if an implementation doesnt do this, it is flawed.