Search code examples
mathdirectxdirectx-10

Performance difference between built in magnitude function and this code

Is there a big difference between using the DirectX vector magnitude math function as opposed to the following code? 

float hyp = sqrt(pow(globalVector.x, 2) + pow(globalVector.y, 2))
Solution

  • Not enough for you to care.

    There are improvements that I'd recommend to the function you posted.

    No need to call the power function; x*x and y*y will be just as good and cheaper.

    I'd protect against roundoff by scaling, like this pseudo code:

    if (abs(x) > abs(y)) {
   r = abs(y/x);
   hyp = x*sqrt(1 + r*r);
} else {
    r = abs(x/y);
    hyp = y*sqrt(1 + r*r);
}