The CPU screen spike bug with Air 2

[mtingle]

are you sure its related to the scroll-window function?

The only reason I say this is because it seems to be triggered when you touch the screen.

Obviously most of the time people touch the screen to scroll so it triggers the CPU spike. but you only have to try to change a volume/ pan or knob of any sort and the cpu hits a spike. this also explains why plugin interaction like the reports people are giving related to fabfilter plugins are seeing the cpu spike.

A 'screen scroll window' function bug just doesn't explain why all these other touch related interactions are triggering the spike.

it would be interesting to see a list of the cpu's activities when the screen is touched and the spike occurs.

it could just as easily be a complex issue with ram allocation or multicore issues that are teething problems with the ipad air 2 and iOS 8.

note: I'm doing lots of heavy editing with large projects today and hitting this spike 'a lot'. I load these same projects onto an Air 1 running iOS 7 and have no problems so I know it's not the projects themselves.

Air 2: iOS 8.1: Auria 1.154: CPU ~60%: 400mb free RAM

[Rim]

If it happens with other graphics elements, I'm not seeing it here. So far I have only seen it happen with scrolling (and with Fabfilter plugin GUIs, which I'm still asking them about). For example, just moving a fader back and forth doesn't do it for me. But as soon as you start scrolling, spikes happen quite easily. And in the editor, pinch zooming doesn't do it, but scrolling does.

No need to speculate any more, as I've sent Apple a simple example project which clearly demonstrates the issue on an Air 2. It's in their hands at this point...

Rim

[mtingle]

sure, just trying give as much info to help track down the bug as possible.

the project I'm currently working on will play through when in the mixer screen but triggers the spike when in the edit screen. not touching a thing in either case.

Hope Apple can resolve this issue soon. Their showcase best ipad yet is still to be fully realised.

[mtingle]

Not sure if you've tested 8.1.1 yet Rim, but after a reboot the CPU spikes seem to have been cleared up.

What's happening your end?

[Yodasoldier]

don't speak too soon......Updated to 8.1.1 and still have spikes and cpu overload as before and as people have described with moving, scrolling etc.

[mtingle]

don't speak too soon......Updated to 8.1.1 and still have spikes and cpu overload as before and as people have described with moving, scrolling etc.

have you fully powered down and restarted? I had to do that to my air 2 before it behaved.

can't get any spikes on a previous project that was playing up with the new iOS update.

Still, early days regarding testing.

[Rim]

iOS8.1.1 still has the same CPU spikes on an Air 2 in my tests here. Even after rebooting. Out of curiosity, is your iPad WiFi or WiFi + cellular?

Rim

[Yodasoldier]

64gb wifi model here.

[Dylan]

I don't know much about iOS coding, but if 8.1.1 is behaving okay for some and not for others, might it be that something else somewhere in the system is interfering with the performance causing those CPU spikes? If so, I guess there are so many different variables, it makes it quite tricky to identify exactly where the problem lies. Someone else mentioned battery level - not sure where I read it now - but claimed that the CPU was much lower on a full battery. Just trying to help identify the culprit, but if it is code related I'll zip up now!

[mtingle]

64gb wi-fi

[Rim]

I've got a WiFi model myself here. The best way I've found to produce these spikes is this:

1. Load The Approach demo
2. Add two Saturns, choosing HQ mode and some preset. The CPU should be around 60-70% now.
3. Press play (important, because the CPU overflow message won't show unless you're playing).
4. From the mixer screen, take your finger and move it back and forth on the screen very fast (making sure the mixer scrolls while you do it). It usually only takes a few tries to get the spikes to happen.

Rim

[mtingle]

I'll be making full tests later this week on some big projects. Sounds like nothing has really changed though.

[mtingle]

(posted in wrong thread)

[mtingle]

[double post]