If your 'slice' is 20 ms... it would appear that you are inferring a frame rate of 50 fps? Is that correct?
If it is, this seems a bit excessive for a Flex application unless you're doing some serious animation. Keep in mind that 1000 ms = 1 second = [ frame rate ] x [ ms slice ]... anything less / more than this will often lead to the perception that something that should be happening isn't ( e.g. a self-imposed race condition ). What I'm trying to say is that I would recommend aiming for an 'ideal' test scenario where a majority would agree that the given example should perform without issues - then proceed to extremes ( render heavy / script heavy frames ) in order to determine where failures begin to appear. Be prepared for questions like, "is it possible that the routine(s) in question weren't performing optimally before... and that FP changes no longer hide my flabby code-stuffs?" So... I would think that it would be in Adobe's best interest to share any benchmarking tools they use internally for Flash Player with this project. More eyes = more fun! Just sayin' Rick Winscot On Thursday, April 12, 2012 at 9:46 AM, Michael A. Labriola wrote: > > I did some more testing and found that it seems to be a case of GC never > > collecting any garbage. I set up an app to create a new object and assign > > it to a > > property every 20 milliseconds. On 10.3 and 11.1 the memory footprint > > remained steady. On 11.2, it grew and grew. > > > > > We definitely see it collecting garbage for us, but it is leaking like crazy > too and it certainly uses a lot more memory for the same Flex app. This has > huge impact to us as well so we are also working on it. > > Mike
