Contribute
Register

[Success] AMD RX6000 Series working in macOS

There was obviously a reason the card was returned, at least with Amazon you will get your money back, if not a similar card for the same price…

Also they won’t ask too many questions, simply telling them it has a coil whine issue should be sufficient for a return.
 
There was obviously a reason the card was returned, at least with Amazon you will get your money back, if not a similar card for the same price…

Also they won’t ask too many questions, simply telling them it has a coil whine issue should be sufficient for a return.

Yes Amazon customer service is great. I just requested the return of the card.
I will order either the same, or an RX 6600 XT, or an RX 6650 XT. I'll report back when I get my new card.
Thanks for all guys.

Have a good evening
 
Thanks guys,

I made several changes, in my EFI folder, and in my config.plist, based on your posts and your configs: Kexts update (Lilu, AppleALC, WhateverGreen).
I still have the coil whine, but otherwise my config seems to work very well.
Some results with Geekbench, and Sensei. I do not know what you think about it?
I put my EFI attached

View attachment 562206


View attachment 562207
The rx6600 cannot give such a result in a metal test.How did you get a result comparable to rx6900xt?
 
@aka80 @RUSSKIYVOLK
Scores are good for a 6900 XT.
My mistake, it’s a 6600, not 6900, so the scores are GB 4 and not GB 5 as @Edhawk says bellow.
I think you have done the right thing returning the card.
You know that 6650 needs device-id spoofing with SSDT-BRG0.aml, right? 6600 works OOTB. All 6000 series need agdpmod=pikera in boot args but you already know this.
 
Last edited:
There is a huge difference between the scores from Geekbench 4 and Geekbench 5! The two should never be compared, as it is like comparing Apples and Oranges.

Only about 100,000 points difference between the two scores for the same card using the older version of Geekbench compared to the score for GB5!

Does GB4 test the card for Apple's Metal framework/API?
 
There is a huge difference between the scores from Geekbench 4 and Geekbench 5! The two should never be compared, as it is like comparing Apples and Oranges.

Only about 100,000 points difference between the two scores for the same card using the older version of Geekbench compared to the score for GB5!

Does GB4 test the card for Apple's Metal framework/API?
Indeed, I didn't see that there was a fourth version
 
There a five major versions of Geekbench. None should be used to compare against another version. As each version uses different criteria when evaluating a CPU or GPU’s performance.
 
Please clarify: What is output from ./gfxutil -f gfx0 ?
Assuming this is in response to my trouble getting the XTXH variant working in MacOS...
For me my card doesn't appear as gfx0. I get "DevicePath not found!"

Instead my RX 6900XT appears as:
Code:
19:00.0 1002:73af /PC01@0/BR1A@0/SL01@0/pci-bridge@0/display@0 = PciRoot(0x1)/Pci(0x0,0x0)/Pci(0x0,0x0)/Pci(0x0,0x0)/Pci(0x0,0x0)
19:00.1 1002:ab28 /PC01@0/BR1A@0/SL01@0/pci-bridge@0/pci1002,ab28@0,1 = PciRoot(0x1)/Pci(0x0,0x0)/Pci(0x0,0x0)/Pci(0x0,0x0)/Pci(0x0,0x1)
 
I made some progress working off this SSDT which was also made for X299 systems. It has a lot of extra stuff in it I don't understand, but I can now name the pci-bridge and display devices, as well as assign a device-id. However, whenever I assign the ID I want to spoof (versus some random id), the system instantly restarts when acceleration kicks in at boot. Others have gotten this card working in MacOS... I don't understand what I'm doing wrong.
 

Attachments

  • Screenshot 2023-01-23 at 7.00.32 PM.png
    Screenshot 2023-01-23 at 7.00.32 PM.png
    203.4 KB · Views: 36
  • SSDT-RX6900.dsl
    6.8 KB · Views: 33
Back
Top