Contribute
Register

macOS 12.3 Update Causes Problems for 5700/6800/6900 Graphics Cards

FYI: If I only add

Code:
                <key>@0,name</key>
                <string>ATY,Adder</string>
                <key>@1,name</key>
                <string>ATY,Adder</string>
                <key>@2,name</key>
                <string>ATY,Adder</string>
                <key>@3,name</key>
                <string>ATY,Adder</string>

My computer will crash after waking from sleep. I needed to add the full device property from this post:

 
Last edited:
Hi @iCpryde

Can you use Hackintool to geth the PciRoot path like CaseySJ describes in this post here:


Just copy the path and post it here and we can help you out.

Thanks

Ted

Mine is PciRoot(0x2)/Pci(0x0,0x0)/Pci(0x0,0x0)/Pci(0x0,0x0)/Pci(0x0,0x0)

I have the same card. I'm still not quite sure where I would add the info to my config.plist. Would this be in the Device Properties?

Thanks!
 
Mine is PciRoot(0x2)/Pci(0x0,0x0)/Pci(0x0,0x0)/Pci(0x0,0x0)/Pci(0x0,0x0)

I have the same card. I'm still not quite sure where I would add the info to my config.plist. Would this be in the Device Properties?

Thanks!
YES!
 
Test Python(5500) and Adder(5700) for 5600XT, both work and fix the performance issue.
Adder has better performance vs Python (25%+ OpenGL score).

Thank you all! No need to downgrade now.
Thanks @WaTeZHUANGJIDINGZHI

For those who are curious about the file that is linked, it is a set of OpenCore config.plist files that have the framebuffer injections for different Radeons that may have experienced graphics slowdowns.

Just for convenience, I extracted the relevant DeviceProperties sections from each plist file and will put them here.

NOTE: You will have to confirm the PciRoot path using Hackintool, gfxutil, or a similar utility in order to properly inject the framebuffer for your card. If your path doesn't match, then the framebuffer won't work.

Edit: Please use the framebuffers with caution! See this post here from @blaatschaap :


These might not be the correct framebuffers, particularly for the 6900, which might be Belknap. However, a few of us have used Carswell successfully. With the Vega, it was possible to test different framebuffers so maybe it's possible with the Navi's as well. Make sure to test these on an EFI that isn't on your primary drive!


Radeon 5500:

Code:
    <key>DeviceProperties</key>
    <dict>
        <key>Add</key>
        <dict>
            <key>PciRoot(0x0)/Pci(0x1,0x0)/Pci(0x0,0x0)/Pci(0x0,0x0)/Pci(0x0,0x0)</key>
            <dict>
                <key>@0,name</key>
                <string>ATY,Python</string>
                <key>@1,name</key>
                <string>ATY,Python</string>
                <key>@2,name</key>
                <string>ATY,Python</string>
                <key>@3,name</key>
                <string>ATY,Python</string>
            </dict>
        </dict>
        <key>Delete</key>
        <dict/>
    </dict>

Radeon 5700:

Code:
    <key>DeviceProperties</key>
    <dict>
        <key>Add</key>
        <dict>
            <key>PciRoot(0x0)/Pci(0x1,0x0)/Pci(0x0,0x0)/Pci(0x0,0x0)/Pci(0x0,0x0)</key>
            <dict>
                <key>@0,name</key>
                <string>ATY,Adder</string>
                <key>@1,name</key>
                <string>ATY,Adder</string>
                <key>@2,name</key>
                <string>ATY,Adder</string>
                <key>@3,name</key>
                <string>ATY,Adder</string>
            </dict>
        </dict>
        <key>Delete</key>
        <dict/>
    </dict>

Radeon 6600:

Code:
    <key>DeviceProperties</key>
    <dict>
        <key>Add</key>
        <dict>
            <key>PciRoot(0x0)/Pci(0x1,0x0)/Pci(0x0,0x0)/Pci(0x0,0x0)/Pci(0x0,0x0)</key>
            <dict>
                <key>@0,name</key>
                <string>ATY,Deepbay</string>
                <key>@1,name</key>
                <string>ATY,Deepbay</string>
                <key>@2,name</key>
                <string>ATY,Deepbay</string>
                <key>@3,name</key>
                <string>ATY,Deepbay</string>
            </dict>
        </dict>
        <key>Delete</key>
        <dict/>
    </dict>

Radeon 6800:

Code:
    <key>DeviceProperties</key>
    <dict>
        <key>Add</key>
        <dict>
            <key>PciRoot(0x0)/Pci(0x1,0x0)/Pci(0x0,0x0)/Pci(0x0,0x0)/Pci(0x0,0x0)</key>
            <dict>
                <key>@0,name</key>
                <string>ATY,Belknap</string>
                <key>@1,name</key>
                <string>ATY,Belknap</string>
                <key>@2,name</key>
                <string>ATY,Belknap</string>
                <key>@3,name</key>
                <string>ATY,Belknap</string>
            </dict>
        </dict>
        <key>Delete</key>
        <dict/>
    </dict>

Radeon 6900:

Code:
    <key>DeviceProperties</key>
    <dict>
        <key>Add</key>
        <dict>
            <key>PciRoot(0x0)/Pci(0x1,0x0)/Pci(0x0,0x0)/Pci(0x0,0x0)/Pci(0x0,0x0)</key>
            <dict>
                <key>@0,name</key>
                <string>ATY,Carswell</string>
                <key>@1,name</key>
                <string>ATY,Carswell</string>
                <key>@2,name</key>
                <string>ATY,Carswell</string>
                <key>@3,name</key>
                <string>ATY,Carswell</string>
            </dict>
        </dict>
        <key>Delete</key>
        <dict/>
    </dict>
 
Last edited:
Doing some testing on my 6900XT using the Carswell injection and performance is tolerable but not where it should be. I'm probably still down at least 20%, but not down 80% anymore.. i can only guess Carswell is still only achieving 6800W performance and not true 6900XT that is also overclocked.

Also note I'm actually testing gaming performance and not just geekbench.

Either way it's a workable solution for now so at least machine isn't a brick, it just isn't 100% either.
 
What shall be the value for 5600XT? I mean Python for 5500, Adder for 5700, but what for 5600? Thanks.

Edit: still looking for the value for 5600XT. I tried the values used for 5500, it is much better now (OpenGL score doubled after fixing).

I'm not sure. mattystonnie's said that his 5700 SSDT worked for the 5600, so maybe try Adder.

 
I'm not sure. mattystonnie's said that his 5700 SSDT worked for the 5600, so maybe try Adder.

Thanks!!! Tried Adder and it is better than Python. Another 25% increase in OpenGL test.
 
@maxxx I'm using a Sapphire Radeon RX 5700 XT 8 GB NITRO+ SE so I think I would need to use the ATY,Adder instead of Carswell .... but I'm not sure where I would add those 4 lines? I know you said VGA path so would that be under the first PciRoot part right after no-hda-gfx?

View attachment 544228
Get Hackintool application and see PCIe tab.
highlighted in green is your device path, you should put another node under <Add> with the value on the device path.

1647573660017.png
 
Back
Top