Intel UHD 630 7MB issue

[GCWShadow]

Hi guys
I have finished the installation of MacOS 11.4 (Big Sur) on my build:
CPU: i7-10700k
M/B: Gigabyte Z590 Vision G

Currently everything looks fine except that the display is quite laggy.
After doing some research I noticed that it has to do with the VRAM setup.

I have walked through several guides about the graphic setup, but it turns out that people's config does not work for me.

The ig-platform-id I have used in my config.plist is <07009B3E> and I have also applied device-id <9BC50000>.
I have tested several platform-id provided in WEG intelHD guide, but nothing works.
It is worth notice that this device-id is necessary for now because without it I'll get signal lost during boot.

Attached is the config.plist I'm using, and any kind of guidance is appreciated.

 

[Edhawk]

Other than a number of disabled/unnecessary placeholder entries from the Sample.plist provided by the OC Developers, I couldn't see anything wrong with your config.plist, not with regards the graphics setup.

• I assume you have already tried using the alternative framebuffer <00009B3E>
• I assume you have disabled CFG_Lock in your bios, so you have no need to enable Kernel > Quirks > AppleXcpmCfgLock.
• Is there a reason you haven't enabled Misc > Debug > AppleDebug and ApplePanic, as they are normally required.
• The shikigva=80boot argument won't do anything in Big Sur, as the Shiki fixes don't work in Big Sur plus you are only using the IGPU.
  • Shikigva=80 is meant to be used when you have an AMD and IGPU working in Catalina and lower.
  • See the screenshot below containing the changes for DRM in Big Sur, specifically the third and fifth items in the list:
  • 

 

[GCWShadow]

Yes
I've tried several platform-id including <00009B3E>.
And for the AppleDebug and ApplePanic, I turned it off when I try to disable the verbose during boot.

And here are some update after some research.
I found out that the signal lost is probably due to the iGPU is being loaded as the framebuffer patching might not done correctly.
The reason for saying that is because I can access the "signal lost" build from the ScreenSharing app on another Mac.
After log into it remotely, I found out that Hackintool can actually detect the iGPU device.

And I have went through the Framebuffer patching guides from tonymacx86 and OpenCore.
But nothing work after I try to correct the BusID and connector-type.

 

[Edhawk]

OK, so I assume you have two cables connected to your display(s). One from your GTX 980Ti (presumably for Windows gaming etc.) and another from one of the motherboard IGPU outputs.

Can I ask how they are connected, i.e. which connectors and which cables you are using for both OS's.

Have you tried booting macOS with the GTX 980Ti removed from the case, not just disabled with the boot argument.

Have you looked at disabling the GTX 980Ti using the Device Properties from this OC troubleshooting guide - https://dortania.github.io/OpenCore-Install-Guide/extras/spoof.html This would be used in place of the -wegnoegpu boot argument.

 

[GCWShadow]

Actually not, I've set in the BIOS that IGFX is the primary display output.
So in both OS I use a same cable (HDMI) to push the display directly from MOBO.

I've setup the app preference in Windows10 to run game on my 980ti.
Anyway, my current build is running with the -wegnoegpu flag so now the display is going well.

I can give it a try later with the removal of the discrete graphic card, but I really doubt that work though.

 

[UtterDisbelief]

Actually not, I've set in the BIOS that IGFX is the primary display output.
So in both OS I use a same cable (HDMI) to push the display directly from MOBO.

I've setup the app preference in Windows10 to run game on my 980ti.
Anyway, my current build is running with the -wegnoegpu flag so now the display is going well.

I can give it a try later with the removal of the discrete graphic card, but I really doubt that work though.

Not sure if I'm missing the point here, so forgive me if so, but ...
You will not get iGPU display output with hardware acceleration and full memory, on a Z590 motherboard with a Comet Lake iGPU.

You may get headless operation but that's not a display.

Yes, you will get the output you see but menu and dock will be solid colour and memory limited. Movement will be laggy.

 

[GCWShadow]

Not sure if I'm missing the point here, so forgive me if so, but ...
You will not get iGPU display output with hardware acceleration and full memory, on a Z590 motherboard with a Comet Lake iGPU.

You may get headless operation but that's not a display.

Yes, you will get the output you see but menu and dock will be solid colour and memory limited. Movement will be laggy.

Just trying to understand how things work.
So now I have a booted build with no signal on the monitor.
And from remote login I'm able to see that the OS got the iGPU running. (including transparent dock, intel power gadget indicate iGPU's frequency no longer 0, etc.)

That's why I thought I just did something wrong with the patching.
Why is that not gonna work with the Z590 + CML build? Thanks.

 

[UtterDisbelief]

Just trying to understand how things work.
So now I have a booted build with no signal on the monitor.
And from remote login I'm able to see that the OS got the iGPU running. (including transparent dock, intel power gadget indicate iGPU's frequency no longer 0, etc.)

View attachment 522439

That's why I thought I just did something wrong with the patching.
Why is that not gonna work with the Z590 + CML build? Thanks.

Okay, maybe I'm still misunderstanding. If you remote-in to your desktop, what physical display are you viewing?

The problem seems to be that although the UHD630 is outwardly the same in the Comet Lake CPU as it is in the Coffee Lake, it actually has a different Device ID. macOS will run this properly in a 400-series motherboard, with full acceleration, but not a 500-series. I'm not sure what dictates this in macOS.

Most builders of 500-series/Comet Lake builds use a discrete GPU. The iGPU can be active for headless use, but so far I have seen no-one get display output fully-accelerated from the iGPU. If you've done this, then great. It will be something to use in future.

I would like to be proved wrong, because I researched and built a similar system, but haven't been so far.

 

[GCWShadow]

Okay, maybe I'm still misunderstanding. If you remote-in to your desktop, what physical display are you viewing?

The problem seems to be that although the UHD630 is outwardly the same in the Comet Lake CPU as it is in the Coffee Lake, it actually has a different Device ID. macOS will run this properly in a 400-series motherboard, with full acceleration, but not a 500-series. I'm not sure what dictates this in macOS.

Most builders of 500-series/Comet Lake builds use a discrete GPU. The iGPU can be active for headless use, but so far I have seen no-one get display output fully-accelerated from the iGPU. If you've done this, then great. It will be something to use in future.

I would like to be proved wrong, because I researched and built a similar system, but haven't been so far.

I'm using a MacBook Pro to do the remote log in.

And yeah, I saw you have done several discussions somewhere else on the same topic.
It turns out that it is more relevant to the support from Apple.

Now my option left is to sit and wait for some solution coming out in the future, or I should turn back to High Sierra and just use my 980ti to drive the display.

 

[UtterDisbelief]

I'm using a MacBook Pro to do the remote log in.

And yeah, I saw you have done several discussions somewhere else on the same topic.
It turns out that it is more relevant to the support from Apple.

Now my option left is to sit and wait for some solution coming out in the future, or I should turn back to High Sierra and just use my 980ti to drive the display.

There won't be any future support from macOS - in my opinion. Apple simply won't make a 500-series computer, unless they surprise us with another Intel Mac.

What to do depends on your use-case. If you need Big Sur or Catalina or full access to the Apple App Store then use another, native or supported, GPU. In theory you can still disable your GTX980 using the techniques you have already used and run another GPU along side it. Many do this. If cost is a problem, do as I do and use a cheap, natively supported GT710 for macOS output.

Although I haven't posted anywhere, I have installed and run High Sierra on a 500-series/Comet Lake mix. There is a specific issue with USB3 support though so if that's important it makes for another sticking-point.