High power usage after long sleep

[tienhuynh5312]

Hi,
I am currently experiencing an issue after a long sleep (overnight), the T480 CPU voltage is above 1.01V and energy consumption is above 10Watts, but Activity Monitor shows the CPU is idle and no task is hogging the resources.
What I tried, but the issue still persists:

1. include DGPU.aml to disable DGPU
2. remove CPUFriend
   Thanks

[EETagent]

Hi,
thanks for reporting yet another issue. That's really high value for idle and we should really focus on this one. I think that it could be caused because some kexts gets mad after long sleep? Or it is false report? You probably measured that with VoltageShift, can you give Intel® Power Gadget a try?

[tienhuynh5312]

@EETagent I actually used Intel Power Gadget. I still have the issue either VoltageShift enabled or disabled.

[EETagent]

Sadly, only thing that could be done about that now is to remove all kexts except Lilu, WEG, VirtualSMC, NVMeFix, USBPorts (?) and try it again.

I think it is for sure some kernel extension, question is which? VoodooRMI, itlwm, IntelBluetoothFirmware, all are still quite experimental. Also please try sudo powermetrics --show-process-energy, we should see there which process is guilty.

[tienhuynh5312]

I have Broadcom wifi card, so I dont really use Intel wireless kexts and Heli anyway. For now, I will try different kext configuration and see. However, it will take awhile because of long sleep time duration.

[tienhuynh5312]

The result looks promising for now with this configuration:

1. HibernationFixup disabled with sleepmode 0
2. RTCMemoryFixup disabled.

[EETagent]

I actually forgot to remove RTCMemoryFixup as it is useless when there is no boot argument.

Hibernation on Hackintosh is still a hit or miss. If disabling it fix this issue, it should be removed for now.

[tienhuynh5312]

Generally speaking, 99% hackintosh issues are sleep-wake related lol. I will keep monitoring this, and report when I am done.

[tienhuynh5312]

with both CPUFriend and CPUDataProvider disabled. The issue still remains.

[EETagent]

Please run sudo powermetrics --show-process-energy and show Energy tab. Thanks

[zombillano]

iGPU is causing that, happened to me before too. If you see carefully to the graph showed on the screenshot you can see that it keeps at full even when there's nothing making use of the GPU. May be an issue with WhateverGreen, there's already an issue on acindanthera's bugtracker repo. I'm on mobile so it's difficult to me to link it, but if you look at that repo you'll find it.

I think we should report what we find over that issue, because right now that issue doesn't have so much interest from acidanthera's group.

Edit: It does happen to me when I am using the iGPU (games, streaming video, etc) and while doing so I put the machine to sleep. Doesn't matter if it is a long or a short sleep.

[tienhuynh5312]

Let's me try to use this Intel Framebuffer patch. The patch that we got shows the UHD 620 has 2GB RAM but in fact it only has 1.5G

[EETagent]

@tienhuynh5312

In the Dortania guide ( Their quality and amount of detail improved a lot ) there is stated

For UHD620 users we recommend the following values:
device-id=59160000
AAPL,ig-platform-id=0000C087

As far as I remeber though, I did not have success with this config.

I am really sure that properties as model, device-type, hda-gfx, every con* ( For HD6** only ) and those other two AAPL are useless and should be removed

[MSzturc]

I am really sure that properties as model, device-type, hda-gfx, every con* ( For HD6** only ) and those other two AAPL are useless and should be removed

It depends. Mostly they are useful for external monitors. For example for my T460 i have a 1440p Monitor with 144hz connected via display port.

To be able to have audio via Displayport

• i have to activate audio over igpu (hda-gfx)
• simulate that my displayport is an hdmi port to get audio transfer protocol working ( con2-type = 08000 )

Settings like con2-pipe are hardware specific. My Monitor for example goes to sleep after Resolution changes between apple logo and desktop on boot process. I guess because from the monitors perspective it looks like an connection loss that is not happen with pipe = 18

[zombillano]

Let's me try to use this Intel Framebuffer patch. The patch that we got shows the UHD 620 has 2GB RAM but in fact it only has 1.5G

Did it work?

[tienhuynh5312]

@zombillano it does not work.
It will show 1.5G but the problem still remains as you stated. The opencore guide did mention this issue. Hope we can get the fix on their end soon.

[zombillano]

I can't find where they did mention it, can you point me where do they say that?

Also I believe they think this isn't an issue or at least don't want to track it down since they closed the issue on their bug repo:

acidanthera/bugtracker#800 (comment)

Maybe if we try to make them see that this is a real issue and not an one individual problem maybe they can try to fix it...

[tienhuynh5312]

@zombillano https://dortania.github.io/OpenCore-Post-Install/universal/sleep.html#main-culprits
in Fixing GPUs section.

Special notes for iGPU users on 10.15.4 and newer:

iGPU wake is partially broken due to numerous hacks apple uses in 
AppleGraphicsPowerManagement.kext with real Macs, to get around this you'll likely 
need igfxonln=1 to force all displays online. Obviously test first to make sure you have this issue.
AAPL,ig-platform-id 07009B3E may fail for desktop Coffee Lake (UHD 630) users, 
you can try 00009B3E instead

in Preparations section

CPUs
AVX often breaks iGPUs and hurt overall stability

[zombillano]

@tienhuynh5312 Can you try using WEG 14.0? I find WEG 14.1 kind of buggy, it gives me high usage on streaming, gaming, etc while 14.0 appears to don't have that issue.

[tienhuynh5312]

@zombillano I literally just upgraded to Big Sur today, and all my kexts are in lastest version with WEG 1.4.2. I am not sure I can use lower build on bigsur though.

[Mateo1234454545]

I'm having this issue too, on a Dell 5570 laptop with an Intel UHD620 and Big Sur.
As @zombillano stated , it doesn;t not happen all the time after waking from sleep but only if you previously watched videos etc.
Hope we find a solution soon because when it happens there is nothing you can do to keep use the laptop. Only reboot solves it.

[zombillano]

@zombillano I literally just upgraded to Big Sur today, and all my kexts are in lastest version with WEG 1.4.2. I am not sure I can use lower build on bigsur though.

Are you getting the same issue with Big Sur?

I'm having this issue too, on a Dell 5570 laptop with an Intel UHD620 and Big Sur.
As @zombillano stated , it doesn;t not happen all the time after waking from sleep but only if you previously watched videos etc.
Hope we find a solution soon because when it happens there is nothing you can do to keep use the laptop. Only reboot solves it.

I don't plan to update to Big Sur any time soon, basically because I need the most stable build possible and Big Sur won't be stable until we get a public release. I've been running WEG 14.0 and it looks pretty stable on macOS 10.15.6, I'm getting a normal behaviour before and after sleep as well as while streaming video, playing minecraft, etc, etc. The only bad thing is that I think 14.0 isn't compatible with Big Sur.

Most likely the changes made to adapt WEG to Big Sur introduced a bug that causes hig GPU usage even on idle. We must point it out to acidanthera's group. Sadly, I believe they aren't willing to solve it soon, based on the fact that they closed the issue about this problem on their bugtracker repo. Also my experience with them has not been that good, neither when I had problems with my NVMe SSD or this issue with GPU usage. They only give attention to issues that affect a large number of users.

[tienhuynh5312]

@zombillano you should also remember Apple tailors their SOFTWARE to their HARDWARE. As long as their software fix errors on their product lines; they will call it a day. I have zero issue with my Intel Pro 7600P NVME which comes preinstalled on T480, so your issue might be related to a specific NVME drive that you have used.
As they stated, I think they are aware of problems with IGPU stability (see my previous comment). Due to recent and rapid changes from Apple implementation of how things work, it is better for them to investigate and give a total solution to it.
Hackintosh is a community project, so we should not expect time-manner fixes from devs. Final note, hackintosh "always" considered not to be your main stable workstation that you can rely on. If you want a stable machine and good support, you can always buy a real mac.

[Mateo1234454545]

Take a look here:acidanthera/bugtracker#800 (comment)

[zombillano]

@zombillano you should also remember Apple tailors their SOFTWARE to their HARDWARE. As long as their software fix errors on their product lines; they will call it a day. I have zero issue with my Intel Pro 7600P NVME which comes preinstalled on T480, so your issue might be related to a specific NVME drive that you have used.

It was a problem with my NVMe indeed, I ended replacing it with a WD SN 720 which works flawlessly.

As they stated, I think they are aware of problems with IGPU stability (see my previous comment). Due to recent and rapid changes from Apple implementation of how things work, it is better for them to investigate and give a total solution to it.

They are aware, but for the original problem that was submitted they already said they won't investigate or fix it. As for this problem (which is indeed a different one because it isn't related to igfxfw boot arg) we are trying to see which commit started the problem.

Hackintosh is a community project, so we should not expect time-manner fixes from devs.

Oh, don't misunderstand me, please. I'm well aware of how hackintosh and its community works. What I was trying to say is that because of the size of OpenCore project they only focus on issues that affect a large number of users. My issue with my SSD was kind of exclusive to me. Although I found several issues closed because they don't want to investigate the problem (as they literaly say). Of course they are on their right to decide wether or not solve something, I just wanted to point out that if we want to find a solution to this issue we might get together so they see it is a general problem and not a specific one.

I know that many of them do this for free. They deserve our respect and understanding about how they deal with issues, at the end of the day they are giving so much to us and we are giving nothing back.

Final note, hackintosh "always" considered not to be your main stable workstation that you can rely on. If you want a stable machine and good support, you can always buy a real mac.

I must dissent with you on this one. Hackintosh has been more than a "hobbie" project since so long. You can get a perfectly working build (how much effort you will need to achieve it is another topic) and use it as your main workstation. I've done it in the past, and certainly will keep doing on the future.

I won't buy that argument of getting a real mac if I want a stable machine as I won't buy an overpriced hardware for what it offers. OpenCore project even relies on the fact of trying to make hackintosh as stable as possible with as many "real-mac" features as possible, too. If they aren't doing this to allow the community to use hackintosh machines as their main stations then I don't see why they would bother doing all the hard work they do. Again, please don't misunderstand me if somehow I sound rude. English isn't my native language and I'm still learning it, so please take this as a friendly conversation more than an agressive discution :)

PD: We are really near to get a "perfect build" for the T480, which makes me feel so happy

[tienhuynh5312]

@zombillano yeah I hear you, we all want a perfect Hackintosh. I do have a desktop Hackintosh with i7 9700 which also experience this issue in the IGPU, but not as bas as T480.
Sometimes, the problems come from upstream code from Apple. If the real Macs is not stable, there is a higher chance of Hackintosh is not stable as well. If Apple fails to produce a stable software, we might expect hiccups on our end, but highly doubt the Opencore devs will chase after issues that come from Apple in the first place. They probably wait for Apple to fix it, then they make it work on Hackintosh.
Just my conspiracy, I feel like Apple likes to mess up their code sometimes to affect Hackintosh community in a way which discourages people using it from all the frustration they have experienced.
Rule of thumb in Hackintosh community, when you find a stable build and config, live with it to death lol.

BTW, I am on Big Sur Beta (last kext build), and right now with a lots of sleep/wake during video playbacks on Edge and Safari, I don't experience any issues. My iGPU locking up usually happens after LONG sleep.

[zombillano]

@zombillano yeah I hear you, we all want a perfect Hackintosh. I do have a desktop Hackintosh with i7 9700 which also experience this issue in the IGPU, but not as bas as T480.
I think IGPU issues aren't as bad on desktop as it is on a laptop because desktops usually have better refrigeration systems, so it really doesn't matter if it consumes 5-10W more.

Sometimes, the problems come from upstream code from Apple. If the real Macs is not stable, there is a higher chance of Hackintosh is not stable as well.

That's what I was saying. I had a MBP just before my T480. Had to change because I really hated how much my MBP overheated even on low load tasks. Tried repasting, cleaning it, nothing helped. I decided that I prefered to take the time to get a hackintosh build working and have options to solve my problems rather than getting stuck with what Apple gives by its own in terms of solutions.

Just my conspiracy, I feel like Apple likes to mess up their code sometimes to affect Hackintosh community in a way which discourages people using it from all the frustration they have experienced.

I wouldn't give them so much credit. To me they make bad hardware/software decisions by their own since they decided to orientate their products to be "exclusive" and fashion in detriment of the features that made a mac "pro". I hope that they do it better with their ARM machines.

Rule of thumb in Hackintosh community, when you find a stable build and config, live with it to death lol.
So true hahahahaha

My iGPU locking up usually happens after LONG sleep.
Long sleep also triggers it to me, but I was having the issue just after a cold boot. Somehow I can't reproduce it right now.

@EETagent Are you compiling by your own the WEG kext? Especifically the one included on the 4.0.0 release and the one on the new ACPI tables repo

[tienhuynh5312]

kexts.zip
I have been testing with these new compiled kexts from Lilu and Friend. I slept and woke my laptop many times, but I have not got the IGPU lockup yet.

[EETagent]

Why you use EnableLidWake.kext? SSDT-EXT4 is there for internal screen issues after sleep.

I have been testing with these new compiled kexts from Lilu and Friend. I slept and woke my laptop many times, but I have not got the IGPU lockup yet.

That is weird as master branch is pretty updated ( Manually compiled, Lilu and Friends was broken for me ), there were not any important updates from what I have read in commit log.

[tienhuynh5312]

@EETagent I was just trying my luck with the latest updates since I have problems with the previous version anyway. I am running on Big Sur right now.

[EETagent]

@tienhuynh5312
Did you made full transition? I installed Beta 4 on a second partition to do some OS stability tests as I plan to install it on my cMP 5.1 ( This machine will have a long life 😄 ). Feel free to post config related problems to #13

BTW, Are there still audio output issues for you on macOS 11?

[tienhuynh5312]

@EETagent audio output issue has "disappeared" for now. I installed Big Sur on my main drive. I know I should have not done it lol. Seems really stable, but iGPU has been locked up from wake just like before (just appeared just now).

[EETagent]

@tienhuynh5312 Maybe this issue can be fixed using latest WEG commit
acidanthera/WhateverGreen@947f69b

[zombillano]

@tienhuynh5312 Maybe this issue can be fixed using latest WEG commit
acidanthera/WhateverGreen@947f69b

I'll try it and report back. BTW, you can get latest builds of all the commits over here https://dortania.github.io/builds/?product=WhateverGreen&viewall=true

Edit: I've alredy tested that functionality making use of the boot-arg igfxnorpsc=1 with the commit before the one linked and it has a bug which makes the GPU getting stuck at 300 Mhz every now and then without any clear reason. The bug remains on this commit.

For me WEG 14.0 is the most stable one. 14.2 also works fine but with the bug mentioned before. If you don't care about optimizing battery usage at its max then maybe 14.2 is a good option

[EETagent]

BTW, you can get latest builds of all the commmits over here https://dortania.github.io/builds/?product=WhateverGreen&viewall=true

I know about that one. Actually, I am preparing GitHub action which would automatically build latest OpenCore and important kexts and push them to special automatized branch, that is low priority though