MathF-functions are faster than the Math-counterpart and it looks cleaner, so use MathF when we cast to float or int anyway.