Tom, After a little more poking around I found this issue in the Adobe bug-base for the Flash Player.
http://bugs.adobe.com/jira/browse/FP-1149 ...and... https://bugs.adobe.com/jira/browse/FP-2009 It looks like the root cause isn¹t with skinning but rather frame rate handling which could be triggered in many different ways; one of them being stateful skins. Nice. I downloaded the example at Effective UI for reference and witnessed the problem first hand. I was able to get it to behave (kind of) using a technique outlined by the masterful Grant Skinner. http://www.gskinner.com/blog/archives/2009/05/idle_cpu_usage.html The problem is that while the application has focus cpu usage will be higher... So, I guess the real question is if cpu consumption while the application is active is justifiable / manageable. Cheers, Rick Winscot On 5/28/09 1:57 PM, "Tom McNeer" <[email protected]> wrote: > > > > > > Rick, > > On Thu, May 28, 2009 at 9:54 AM, Rick Winscot <[email protected]> wrote: >> >> >> >> Your observation that your application footprint increased when touching >> creationPolicy is what one should expect. > > > > > > > Actually, I wasn't very clear: it wasn't that the application footprint > increased (though it did, minimally); rather it was that by having a > TabNavigator with a creationPolicy of "all" (which I have now managed to work > around), five components were created immediately, each of which added to the > CPU load because of the skinning issue that was common to all of them. > >> an increased footprint when adding skins into the mix is also something you >> will probably see. > > > > > > > Sure. But this is CPU usage caused by the stateful skins. > >> The part that has me scratching my head is that you are seeing high cpu usage >> - which is something that is typically linked to refresh/redraw rates. Do you >> have any source we can take a look at? > > > > > > > > > The source wouldn't help you, Rick. It's very typical forms-in-a-tab navigator > stuff. Yes, the issue must be redrawing -- pretty constant redrawing I'd say. > > If you look at the blog post I referenced, you'll get a better view of the > problem. Effective UI did a good deal of testing when they ran into the same > issue, and concluded that it had to be a result of the use of stateful skins > -- although they didn't seem to be able to determine the underlying cause. > They documented some pretty impressive CPU usage by simply adding more skinned > elements. > > I'm just kinda surprised more people haven't run into it. > >
