How do exponentiation algorithms work? [duplicate]

Question

Exponentiation in most modern languages is easy .. I use the common operator for that in my language of choice or whatever function that compensates that to get the desired functionality.
I want to know, how does this exactly work ?

The following algorithm in C is often used to demonstrate this effect ..

double exp(val, pow) {
    for(int i = 0; i < pow; ++i)
        val *= val;
    return val;
} // exp(2, 3) -> 8

However, there is a serious bug here .. What if pow is 2.6 ? That would return 8 also ..
That's simply because the loop condition only compares the two numbers ..
But when I do something like this, it works well ..

#include <math.h>
int main() {
    printf("The result of 2 to the power of 2.6 is %.2f", pow(2, 2.6));
}

How can the latter behavior be achieved ?

Edit:

According to the answers, it seems the taylor expansion algorithm is the key to exponentiation, so .. what about multiplication ? How can decimal multiplication be achieved ?

Answer 1

Exponentiation is usually implemented as (lots of special cases plus) a reduction to exp. If you have an exp function and its inverse ln handy, you can compute x^y as

exp(y*ln(x))

But you might wonder how exp is implemented. For small arguments, the series expansion works well:

exp(x) = 1 + x + x^2/2 + x^3/6 + x^4/24 + ...

Edit: This is the Taylor expansion referred to in the other answers.

For larger values there are argument reduction techniques that can be used to compute the value.