Task: check if x (double) is very close to zero using a threshold of let's say 1e-8. Which led me to two performance questions.
Methods I currently use:
1: The save method for when I'm unsure if the data is completely positive or not.
double threshold = 0.00000001;
if ( abs(x) < threshold );
//remove x
2: If I'm pretty sure that the data is completely positive.
if ( x < threshold )
//remove x
So that leads me to my first question, is the second method faster than the first method or not? I just always assumed the second method to be faster...
Another way I've been thinking about as an alternative to method 1 would be to use
if ( x < threshold && x > -threshold )
//remove x
Which leads me to my second question, is this method faster or slower than method 1?
Aucun commentaire:
Enregistrer un commentaire