Jump to content

AMD Polaris IDs on Sierra / High Sierra


Ciro82
 Share

870 posts in this topic

Recommended Posts

Finally i could report, that my Sapphire RX480 Nitro runs under macOS 10.13, too.

But don't ask, what kind os work this was. It did NOT work nearly "out-of-the-box" ! <--- it is really far away from working like this.

 

But here are my results:

 

• iGPU set to primary, RX480 in slot#1 as secondary, CSM enabled = verbose boot but than black screen on both monitors (each card has its own monitor, connected via Displayport)

• iGPU, CSM enabled and RX480 set to FakeID 67ef1002 via CLOVER (as for successfull boot, when RX card is the only one, but w/o accelleration) = black screens

• iGPU, CSM enabled, RX460 in Akitio Node Encloser connected via Gigabyte Alpine Ridge Thunderbolt 3 card = Login Screen comes up but: after a login immediatly Logout and back to Login Screen

• iGPU, CSM enabled, RX480 in Akitio Node Encloser connected via Gigabyte Alpine Ridge Thunderbolt 3 card = Login Screen comes up but: after a login immediatly Logout and back to Login Screen

 

so i was thinking: what else to try to make it work? Than i remember, that Apple's eGPU Developer Kit is a RX580 and Sonnet eGPU encloser. And i remember, that it is possible to flash Sapphire RX480 Nitro with VBIOS from Sapphire RX580 Nitro. Botted into WIN10, get VBIOS from TechPowerUp website, flashed onto the RX480, put that newly flashed card back into the Akitio Node and rebooted into HighSierra:

 

SystemProfiler.jpg

As you can see, full METAL support, but system remains only 0MB VRAM ?!? But it work and i have also testen HEAVEN Benchmark, which reports good framerates (havent made a full benchmark test with it so far).

My personal conclusion: WHAT A DISASTER TO MAKE THIS ONE WORK! It may be my personal hardware, which have made this report and experience so horrible. As i told earlier: N E V E R had such problems within macOS 10.12.

 

Regards...

 

EDIT1: now i know, why it is showing 0MB VRAM: it is CLOVER related, cause i don't set any value for VRAM in config.plist - exactly i have completely no entry to set this value in my config.plist and which already never was a problem so far, cause CLOVER detected VRAMsize automatically. Booted with "inject ATI = false", i will get RadeonFramebuffer, but correct VRAM size shown in system profiler. Not a problem for me and maybe it is related cause it is somewaht in the night and i am realy tired from this horrible experience.

  • Like 1
Link to comment
Share on other sites

 

Yep 

 

AMD Radeon™ R9 Fury Series 7300

 

So, forgive my ignorance, but do I only need to change fb in clover to maltese? or malteseS ??

I've got 10.13 going ok, just no acceleration on my R9 Nano. (X99)

 

post-1564334-0-95491500-1496975719_thumb.png

 

Flag that. I got no output at all and some HD4xxx video. Back to baffin and no acceleration.

Link to comment
Share on other sites

Finally i could report, that my Sapphire RX480 Nitro runs under macOS 10.13, too.

But don't ask, what kind os work this was. It did NOT work nearly "out-of-the-box" ! <--- it is really far away from working like this.

 

But here are my results:

 

• iGPU set to primary, RX480 in slot#1 as secondary, CSM enabled = verbose boot but than black screen on both monitors (each card has its own monitor, connected via Displayport)

• iGPU, CSM enabled and RX480 set to FakeID 67ef1002 via CLOVER (as for successfull boot, when RX card is the only one, but w/o accelleration) = black screens

• iGPU, CSM enabled, RX460 in Akitio Node Encloser connected via Gigabyte Alpine Ridge Thunderbolt 3 card = Login Screen comes up but: after a login immediatly Logout and back to Login Screen

• iGPU, CSM enabled, RX480 in Akitio Node Encloser connected via Gigabyte Alpine Ridge Thunderbolt 3 card = Login Screen comes up but: after a login immediatly Logout and back to Login Screen

 

so i was thinking: what else to try to make it work? Than i remember, that Apple's eGPU Developer Kit is a RX580 and Sonnet eGPU encloser. And i remember, that it is possible to flash Sapphire RX480 Nitro with VBIOS from Sapphire RX580 Nitro. Botted into WIN10, get VBIOS from TechPowerUp website, flashed onto the RX480, put that newly flashed card back into the Akitio Node and rebooted into HighSierra:

 

SystemProfiler.jpg

As you can see, full METAL support, but system remains only 0MB VRAM ?!? But it work and i have also testen HEAVEN Benchmark, which reports good framerates (havent made a full benchmark test with it so far).

My personal conclusion: WHAT A DISASTER TO MAKE THIS ONE WORK! It may be my personal hardware, which have made this report and experience so horrible. As i told earlier: N E V E R had such problems within macOS 10.12.

 

Regards...

 

EDIT1: now i know, why it is showing 0MB VRAM: it is CLOVER related, cause i don't set any value for VRAM in config.plist - exactly i have completely no entry to set this value in my config.plist and which already never was a problem so far, cause CLOVER detected VRAMsize automatically. Booted with "inject ATI = false", i will get RadeonFramebuffer, but correct VRAM size shown in system profiler. Not a problem for me and maybe it is related cause it is somewaht in the night and i am realy tired from this horrible experience.

So do you thing if will work with boot screen wen we flash gpu with bios from Sapphire RX480 Nitro? (whiteout IGPU or helper card) 

Link to comment
Share on other sites

This is the "36 CU limit patch" for High Sierra:

<dict>
    <key>Comment</key>
    <string>36_CUs_Patch</string>
    <key>Disabled</key>
    <false/>
    <key>Find</key>
    <data>SLgEAAAAAQAAAEiJQ1THQ3wJAAAA</data>
    <key>Name</key>
    <string>AMDRadeonX4250</string>
    <key>Replace</key>
    <data>SLgEAAAAAQAAAEiJQ1THQ3wSAAAA</data>
</dict>

take care that we now have to patch "AMDRadeonX4250" instead of "AMDRadeonX4100"!

The "Remove CU Limit" patch is no longer needed - or not found within "AMDRadeonX4250.kext".

 

luxmark.jpg

  • Like 1
Link to comment
Share on other sites

Rom from my RX 480 https://www.dropbox.com/s/sxoa0g0yn98ofo7/RX480Mac.ROM?dl=0

 

New Sierra loading standard framebuffer

 attachicon.gifZrzut ekranu 2017-06-07 o 12.28.12.png

Is it possible to run this on MSI Gaming X RX580?

Or is there a way to mod my bios? I searched and can't find anything about modding the bios. I see many things about RX480 - RX580 but not to get it work in macOS

Link to comment
Share on other sites

There seems to be a new checkroutine within AMD9500Controller: it checks for the Part# of the installed AMD card.

If it finds this value for Part#: 113-4E3531U or 113-4E353BU , than it sets "Radeon RX 580" as modelname and chooses "Orinoco" (which will match perfectly for RX480!) instead of "AMD,RadeonFramebuffer" as its related framebuffer.

 

Thats because Apples eGPU kit contains Sapphire AMD RX 580 PULSE as their bundled cards. You can check this with this patch within CLOVER:

<dict>
	<key>Comment</key>
	<string>Part#_Patch</string>
	<key>Disabled</key>
	<false/>
	<key>Find</key>
	<data>MTEzLTRFMzUzQlU=</data>
	<key>Name</key>
	<string>AMD9500Controller</string>
	<key>Replace</key>
	<data>MTEzLVNFMzY2QVU=</data>
</dict>

! You will have to replace the "Replace" value with Part# of your card. The one in the example above is my Part# !

To find out your part#, boot CLOVER with "Inject ATI = FALSE" and note your Partnumber.

 

This will give you:

ScreenShot2017-06-11at21.42.47.png

ScreenShot2017-06-11at22.14.30.png

  • Like 3
Link to comment
Share on other sites

There seems to be a new checkroutine within AMD9500Controller: it checks for the Part# of the installed AMD card.

If it finds this value for Part#: 113-4E3531U or 113-4E353EU , than it sets "Radeon RX 580" as modelname and chooses "Orinoco" (which will match perfectly for RX480!) instead of "AMD,RadeonFramebuffer" as its related framebuffer.

 

Amazing finding, applied this patch on my rx480, tho i still get black/green/other screen color+system freeze when using card in intense graphics ...

Link to comment
Share on other sites

There seems to be a new checkroutine within AMD9500Controller: it checks for the Part# of the installed AMD card.

If it finds this value for Part#: 113-4E3531U or 113-4E353EU , than it sets "Radeon RX 580" as modelname and chooses "Orinoco" (which will match perfectly for RX480!) instead of "AMD,RadeonFramebuffer" as its related framebuffer.

 

Thats because Apples eGPU kit contains Sapphire AMD RX 580 PULSE as their bundled cards. You can check this with this patch within CLOVER:

Yes good find!

How would that work with the nano / fury?

I'd need to find the part# .  edit: found the part#

The 9300 controller kext is loading fine, but the X4500HW services kext isn't.

 

post-1564334-0-01777400-1497220474_thumb.png

post-1564334-0-99226400-1497220489_thumb.png

post-1564334-0-90178900-1497220501_thumb.png

Link to comment
Share on other sites

There seems to be a new checkroutine within AMD9500Controller: it checks for the Part# of the installed AMD card.

If it finds this value for Part#: 113-4E3531U or 113-4E353EU , than it sets "Radeon RX 580" as modelname and chooses "Orinoco" (which will match perfectly for RX480!) instead of "AMD,RadeonFramebuffer" as its related framebuffer.

 

Thats because Apples eGPU kit contains Sapphire AMD RX 580 PULSE as their bundled cards. You can check this with this patch within CLOVER:

<dict>
	<key>Comment</key>
	<string>Part#_Patch</string>
	<key>Disabled</key>
	<false/>
	<key>Find</key>
	<data>MTEzLTRFMzUzQlU=</data>
	<key>Name</key>
	<string>AMD9500Controller</string>
	<key>Replace</key>
	<data>MTEzLVNFMzY2QVU=</data>
</dict>

! You will have to replace the "Replace" value with Part# of your card. The one in the example above is my Part# !

To find out your part#, boot CLOVER with "Inject ATI = FALSE" and note your Partnumber.

 

...

 

 
Hello Mork,
 

 

I am curious about something, you wrote that the check routine is looking for 113-4E3531U or 113-4E353EU, but in your patch you use <data>MTEzLTRFMzUzQlU=</data> for the find parameter which translates to 113-4E353BU. Was that a typo or is there another reason for using that?
Link to comment
Share on other sites

 

 
Hello Mork,
 

 

I am curious about something, you wrote that the check routine is looking for 113-4E3531U or 113-4E353EU, but in your patch you use <data>MTEzLTRFMzUzQlU=</data> for the find parameter which translates to 113-4E353BU. Was that a typo or is there another reason for using that?

 

 

is base64_encode/decode

Link to comment
Share on other sites

I am curious about something, you wrote that the check routine is looking for 113-4E3531U or 113-4E353EU, but in your patch you use <data>MTEzLTRFMzUzQlU=</data> for the find parameter which translates to 113-4E353BU. Was that a typo or is there another reason for using that?

 

Yes, was a typo, allready corrected. Thanx.

Link to comment
Share on other sites

 Share

×
×
  • Create New...