Re: Absolute Value |x - y|
Re: Absolute Value |x - y|
- Subject: Re: Absolute Value |x - y|
- From: John Stiles <email@hidden>
- Date: Sun, 5 Dec 2004 09:22:46 -0800
On Dec 4, 2004, at 7:11 PM, Frederick Cheung wrote:
On 4 Dec 2004, at 20:56, Gregory Weston wrote:
On Dec 4, 2004, at 2:18 PM, Malte wrote:
i've been looking around trying to find a method that calculates the
|absolute value| of a number.
Without success.
Does anyone know if such a function exists?
They do, but you're probably better off just writing
T abs = ((x<y) ? (y-x) : (x-y));
where T is whatever type you need the result to be.
The difference is probably that fabs has well defined behaviour in
case one (or both) of x and y is something like NaN or +/- \infty,
whereas this might not be the case for the expression you give.
If it were me, I'd use __fabs(x-y). Two instructions, branchless. :)
That being said, the result of the - operator and the < operator when
given +/-Inf and/or NaN is perfectly well defined. It's part of the
IEEE standard and it's really something any serious programmer should
know. (Honestly, it surprises me how many talented programmers never
bothered to learn the basics of IEEE floating point.) Unless I'm
missing something, the ternary expression above would generate the
correct result.
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Cocoa-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden