Floating point comparing with >= versus (> or ==)

Question

I noticed the following in a chunk of code I'm maintaining/extending:

float systemVersion = [[[UIDevice currentDevice] systemVersion] floatValue];
if (systemVersion > 3.2 || systemVersion == 3.2 ) {
    //Stuff
}

I know floating point can result in some odd comparison behaviors due to precision, but would the above behave any differently than the chunk of code below?

float systemVersion = [[[UIDevice currentDevice] systemVersion] floatValue];
if (systemVersion >= 3.2) {
    //Stuff
}

Seems to be the same even if you throw NaNs and infinities at it. :) — Mysticial, Sep 04 '12 at 18:30
Please tell me you don't have to deal with code that stores version numbers as floats... :-( — Mark Dickinson, Sep 04 '12 at 18:46
@MarkDickinson It looks like the version number is stored as an NSString, which someone before me decided to turn into a float (clobbering the minor version number)... As I go through the code I'm hoping to change it to match [this answer](http://stackoverflow.com/a/5337804/264775), but I'm hesitant to change this outside the segments I'm working in. — thegrinner, Sep 04 '12 at 19:14

score 2 · Accepted Answer · answered Sep 04 '12 at 18:25

2

It's the same thing. In many compilers, the resulting machine instructions are exactly the same (although I can't say for sure that's true for clang).

answered Sep 04 '12 at 18:25

Husker Jeff

857
1
5
6

score 1 · Answer 2 · answered Sep 04 '12 at 18:22

1

No, it wouldn't. I think this code, that turned out this way by accident (maybe after a couple of small adjustments).

answered Sep 04 '12 at 18:22

GolezTrol

114,394
18
182
210

Floating point comparing with >= versus (> or ==)

2 Answers2