Gigabyte Z490 Vision D (Thunderbolt 3) + i5-10400 + AMD RX 580

[CaseySJ]

Here's an extra fun lilu kext. Who says you need a M1

View attachment 529484

Where there's a will...

 

[CaseySJ]

Haha, well we can't let you have all the fun...

I am not certain if AdviseFeatures need to be enabled. The update may work without this flag, but I didn't test with it off.

Downloading and installing Beta 7 now (incremental update, not full installer).

 

[jimbojingle]

When I see posts in this thread, my mind automatically thinks "Z490 Vision D" or "Z490 Vision G". This thread is primarily for these two boards.

Anyway, to adapt the SSDT to GA-Z77X-UP5-TH, please post these two files:

• System DSDT (from MaciASL)
• Screenshot of IORegistryExplorer
  • Simply scroll down to the section containing the Nvidia GPU (or "IONDRVFramebuffer")
  • If you use the 'search' field, click the 'x' to the right of that field after the search is done

These errors indicate a mismatch between config.plist and the version of OpenCore you're using. If OpenCore Configurator was used to edit config.plist, please look at the title bar of the app. It will tell you which version of OpenCore will be supported by the config.plis

Good -- no change needed to the SSDT from my previous reply!

On your board, Nvidia GPU is on PCI0.PEG0.PEGP, which is the same as Designare Z390 (slot 1).

Thanks for confirming.

I'm not having any luck tracing the cause for this error.

 

[CaseySJ]

Thanks for confirming.

I'm not having any luck tracing the cause for this error.

What errors are you seeing? Do they begin with OCS: ?

 

[CaseySJ]

Here's an extra fun lilu kext. Who says you need a M1

Two can play that game!

Update: AdvancedMap.kext is not compatible with AMD Ryzen.

 

[teosoft]

Hi @XStylus, do you mind sharing your EFI folder? I have nearly the same base configuration; only going to use W480 Vision W and AMD W5700 Pro. Thanks in advance!

Ladies and gentlemen, may I present the first (to my knowledge) W480 Vision D running MacOS.

View attachment 487894


All the installation steps went smooth as butter except for a minor catch with Step 12, specifically regarding QuickSync HEVC and H.264 hardware acceleration. VideoProc gave me a thumbs down, which I somewhat expected for reasons that I'll get into.

--Edit: FIXED! See link:

Gigabyte Z490 Vision D (Thunderbolt 3) + i5-10400 + AMD RX 580

NOTE 5 has been added to the top of Post 1. USBInjectAll 0.7.6 can be downloaded from here.

www.tonymacx86.com

According to Intel ARK, the iGPU for the I#-10#00 processors is the UHD Graphics 630 and has a device ID of 0x9BC5. On the other hand, the iGPU for the Xeon W-1290P is the UHD Graphics P630 and has a device ID of 0x9BC6.

My thought is that the graphics adapters are likely extremely similar and the P630 would likely work with the regular 630 driver, but it likely doesn't recognize the device ID and therefore does not load. (Edit: A quick tweak in OpenCore as referenced at the link above will fix this.)



I dunno if Sleep functionality is supposed to work on Hackintoshes, but it does not in my case at present.

--EDIT: FIXED! Disabling memory XMP in the BIOS was the solution in my case. Also, I've since switched to four sticks of ECC memory since that's the main benefit of using a Xeon chip anyway.



That issue aside, everything else appears to run okay. Both on-board ethernet ports work (though I have also added an X550 X540 card for 10GbE), every USB port works, and Thunderbolt works (though I haven't done the firmware flash hack to enable extended mode). I have not installed a WiFi adapter so I can't verify Handoff, Bluetooth, or otherwise just yet.

That aside, I haven't done any deep testing to really bang on this thing, nor do I even know where to start to really stress test it. However, I did do a Cinebench R20 run, which came up with 541pts single core, 6336pts multi-core. That seems a bit on the low side, so I'm open to additional tuning suggestions.


In any case though, I see this as an excellent start. Would love to see others join me on this one, once the quirks are ironed out. The W480 Vision D and the Xeon W-1290P are going for $379 and $560 respectively, and given how heavily marked up the i9-10900K is at some stores, you just might come out ahead, and without having to sit in backorder hell.


UPDATE: Five months in, this machine has been working like an absolute champ. I have not tested Handoff since I don't have an iOS mobile device, but I have experienced NO issues of any kind.

 

[santitf]

Open your config in OpenCore Configurator, "Device Properties".
Add new device: PciRoot(0x0)/Pci(0x1,0x0)/Pci(0x0,0x0)
Set properties:

name	Data	23646973706C6179
IOName	string	#display
class-code	Data	FFFFFFFF

A few days and all is working fine.
Only a question about saving electricity: my system uses 75W in idle. If I disconnect the GPU pcie cable attached to my 3060ti, the consume down to 38W in idle (almost double the consumption).
Is there any alternative to don't unplug the pcie cable using OC settings/SSDT,...?

 

[emistral]

Some questions:

• Is the Windows boot problem happening with OpenCore 0.6.1 or 0.6.2?
• Are you sure that the computer is booting from the correct disk? We can check the boot disk priority in BIOS --> Boot
• The solution to ACPI_BIOSerror was to ensure that all SSDTs are conditionally injected only into macOS.
  • We do this by adding this to each SSDT: If (_OSI ("Darwin")) { then do Mac stuff }
• I can check your SSDTs if you compress and post the OC/ACPI folder.

@CaseySJ to answer your questions:
1. Windows problem started with 0.6.2
2. In the BIOS I am booting from opencore. Sequence is Opencore, then windows 10 then UEFI Samsung (which is the main NVME where the OC EFI is (same as first in boot sequence). How does OC pick up the disk to boot? is it dynamic or static? I mean if I have a windows 10 disk, then swap it for another windows 10 disk, does Oc pick the right disk or does it maintain a static entry of the previous disk and would eventually fails if it tries to boot from that disk?
3. See attached. Note that those are aml from 0.7.2 release that you posted a while ago. I have not transitioned to 0.7.3 yet. About to do it today.
I had a quick look at some of the SSDT but the use of if (_OSI("Darwin")... is a bit inconsistent. some of those SSDT have the if statement for instance inside a method, some have the if statement encapsulating methods and code, some have no if statement, etc.
I would expect the if statement to be the first thing at the beginning of the SSDT?
The ones that looked weird to me are:
- SSDT-EC
- SSDT-DTPG
- SSDD-DMAC
- SSDT-AWAC

Also I compared some SSDT in 0.6.1 to 0.7.2 and the following ones did not have the if statement in 0.6.1:
- GPRW
- PLUG
- MCHC
- TB3HP
- UIAC V2

 

[Ayvan]

A few days and all is working fine.
Only a question about saving electricity: my system uses 75W in idle. If I disconnect the GPU pcie cable attached to my 3060ti, the consume down to 38W in idle (almost double the consumption).
Is there any alternative to don't unplug the pcie cable using OC settings/SSDT,...?

I think, it's normal idle power consumption for GPU and GDDR memory on minimal frequncies... I don't know how to reduce it using software methods like SSDT or settings/kexts.

 

[jimbojingle]

What errors are you seeing? Do they begin with OCS: ?

Yes they do.

I attach a screenshot.