I was writing my library and decided to test if there was a faster way of doing abs().

return a < 0 ? -a : a;

The above tests twice as fast as Math.abs(). I'd thought I'd check the PApplet code before bragging but you'd already beat me to it and implemented this.

http://dev.processing.org/source/index.cgi/tags/processing-0118/core/src/processing/core/PApplet.java?rev=2709&view=markup

But why haven't you done the same with max()?

Code:

int timer = 0;
float a = 3.0;
float b = 4.0;

void setup(){
  timer = millis();
  for(int i = 0; i < 10000000; i++){
    // function 1
    float x = (float)Math.max(a, b);
  }
  println(millis() - timer);
  timer = millis();
  for(int i = 0; i < 10000000; i++){
    // function 2
    float x = nMax(a, b);
  }
  println(millis() - timer);
}

float nMax(float a, float b){
  return a > b ? a : b;
}
 

The latter method seems more than twice as fast even as a function. I wouldn't normally quibble, but since you've already optimised abs(), what's the deal with max() or min()?