So after uninstalling the SD card driver, the non-audio affected high CPU usage still happens maybe after every 2nd or 3rd wake from sleep from a fresh boot. Rebooting the SP2 clears the high CPU usage.
Given the conversation on the previous page about _L6D and GPE, I think I'll try disabling things in my device manager one by one (specifically audio) to see if any of them remove the issue. Already tried disabling some USB-related devices but that didn't affect it.
Just to let you guys know. My problem happened again.
All I did was watch YouTubes videos, and play Civ 5 (all plugged in).
And I close the Type Cover 2, and went to bed.
I have my Surface Pro 2 set to go to sleep after 60min when plugged in. Hybrid sleep is disabled. (just checked).
Hibernation is set by windows default at 240min. Hmm could it be that?
Unlike last time, I have no wireless problem this time. And I don't have a micro SD card in, nor do I have one anywhere in my house.
I do have my 2.4GHz wireless mouse receiver connected on USB.
System and System interrupts are (combined) at 23-30% fluctuating.
Removing my USB wireless receiver, and all is fixed. But I don't know if it's because Windows executed the audio sound, or a problem with the USB controller.
Putting it back, and the problem is like it never existed.
I dont' recall if last time I have the USB wireless receiver, thought. I don't think so, as it's usually details I mention.
Please note that my mouse is about 5 years old. I used it on my laptop, and PC at times, running Vista, Windows 7 and now Windows 8, and never noticed any problem.
Nor had noticed any issue before the latest firmware.
So could it be just the USB controller that is buggy?
It's quite clear it's most probable a DSDT bug (trivially fixable for whoever designed the DSDT for MS, if it wasn't MS itself). I can think of a few ways to workaround it but no idea what secondary effects it would have (e.g. why is the wakeup ISR enabled at all when the system is already in S0?). For now I'm trying to guess why I can't wakeup the SFPro2 under Linux using the keyboard, which seems to be related to the "affected" function.
Audio from Windows desktop, and XBox Music, doesn't solve the problem. Removing the USB mouse receiver didn't fix the problem this. No wireless issue. I turn off Wireless card and back on, still there on both instances.
Bluetooth was already off, but I turned on on, thinking maybe a driver bug, nope still the CPU consuming 30%.
uninstall the USB controller, lead no improvement, detaching the keyboard didn't help.
Putting the device to sleep and wake up, didn't fix it.
Removing audio (although windows asks me to reboot), didn't help either.
I tried to remove the most I can without reboot in Device manager, and nothing helped. Reboot it is.
Serendipitously, I just discovered something new about this behaviour: I was on the bus and checking my task manager after waking it from sleep, found it using 24-30% CPU, and was about to start disabling devices in Device Manager while navigating with my stylus when I noticed that, suddenly, it went down to 10-15%. Shocked, I moved the stylus away and the CPU usage almost immediately jumped back up to 24-30%. This phenomenon also happened when just touching the display constantly, such as by dragging Device Manager around my desktop with my finger. It wasn't completely linked in that there were a few occasions where merely hovering over a small part of the screen wouldn't trigger the drop in CPU usage but, after moving the stylus cursor over certain selectable elements, it would then retrigger the usage drop.
So this high CPU usage might be related to the touch sensing somehow.
I have the new Wacom Feel drivers (installed yesterday), and all day I have trying to reproduce the problem I was getting. I was thinking it's because of hibernation.
Well, the only thing I have discovered is that the system doesn't hibernate automatically, but I can do it manually, although sometimes it works, but others time it crashes the system. When it works, it works fine on wake up. So I have auto hibernation disabled (which I really don't care, never intended to use, default was like 260min or something like that). It seams to happen when I use my Surface Pro 2 for long period of time, so we will see.
I do know it makes no sense, that I solved it before by removing my USB mouse receiver, but I thinking right now, what triggers the issue, not how to solve it.
I had an issue with the system process running at about 20%-30% upon waking (unrelated to the SD card issue). I "solved" it by disabling the two memory diagnostic tasks and I haven't had a recurrence since (it has been over a month) - of course, it doesn't solve the SD card issue. I have no doubt that there is an underlying problem that caused the task to spin its wheels, but at least that very annoying symptom has abated.
I finally had this happen again and, unfortunately, using touch or the stylus didn't reduce the CPU usage so perhaps that was just a coincidental finding given the situation at the time... or perhaps there are even more variants of this high CPU usage.
Just to provide an update. I no longer have the problem. I simply can't reproduce it. I think the simply fact that I have uninstalled (but did not check the box to delete the drivers) in Device Manager all devices one by one, and restarted a few times until everything is installed back in Windows, solved the problem.
I'm seeing the same issue on my SP3. The best "solution" I've found so far is I just leave AIMP, my chosen music player, open and always playing through my playlist... I just hit the mute icon on it and forget about it. It's a totally ridiculous hack, but it does keep the CPU < 10% pretty consistently. Being as my SP3 is actually my desktop machine it's not really a problem... but I shouldn't have to resort to such measures obviously.