See that last line in my quoted text:
That line was removed moments after being posted. If you refresh the page you'll see what I mean.
I did see that right after I replied, but I wanted to set the record straight for others that may not of noticed regardless. I don't understood why MacRumors claims it's support as of 12.0.1. I left a comment over there asking them on whether they have evidence to back their claims.
As far as I understand this stuff, you can give any legal ACPI name you want. The trick is that all unnamed devices have to be defined one by one and in order. "Device" means that it has an address (@0, @1), and this address is to be reflected in the SSDT; IOReg entities without an address (IOPP) do not exist in ACPI, do not belong in the ACPI path and should not be in the SSDT. You can't jump to the GPU if there are three PCI bridges without ACPI names in the way; you have to define the first bridge @1, then the second bridge @0 within the first bridge, the third bridge @0 within the second, and finally you get to the eGPU.
This one should be fine. It’s likely that it has device ID 0x73BF, which is natively recognized. But on the off chance that it has the unsupported device ID 0x73AF, then we can still make it work by spoofing its device ID with a simple device property. Just install it in a PCIe slot (not eGPU).
Lol. I'm hoping to change that in the near feature. I spoke to Apple's engineers today to get them to add support for 0x73af... it could be as soon as next week (soonest Weds. or Thursday, if not more than likely the following Weds. or Thursday). I have a follow up call with the engineers on Thursday next week. I wasn't able to get them to add Navi22, unfortunately. Everyone may be out of luck for the 6700 XT. They're more likely to add Navi24 GPUs than the 6700 XT which aren't even released yet.
I just looked at the code for the SSDT-EGFX.aml you provided... I've actually tried that exact code a few days ago and it didn't work (it was one of the earliest things I tried), so I'm down to trying the SSDT-EGFX-SPOOF.aml you provided. I'll let you know how that goes tomorrow.
@mrt1976 : That should work out of the box, but you never know!
SSDT-EGFX-SPOOF is SSDT-EGFX with an additional _DSM method to inject the device-id. If the simple SSDT doesn't work, the more elaborate will not work either. @CaseySJ is probably right that we shouldn't make static changes to removable devices. And it just cannot work if the removable device is not fully initialised when the ACPI table is loaded.
Confirming PowerColor RED DEVIL Radeon RX 6600 XT 8GB GDDR6 PCI Express 4.0 ATX Video Card 6600XT 8GBD6-3DHE/OC working with 2020 MacBook Pro and Razor Core X via Thunderbolt and using Monterey Beta 12.1.
