AMD Polaris IDs on Sierra / High Sierra

[Mork vom Ork]

just built it myself via "Build_Clover_4.5.0.sh" script here via XCode. Give it a try, it's not that hard to learn how to use.

[Max.1974]

Tks Mork!! Nice to know about it!! But too advanced for me      Enjoy it!!! 

[Picasso]

@Mork vom Ork

 

What about GPU test? Are the same results as before or better?

[omnitrio]

maybe be useful for owner's rx 560

simple cosmetic patch

 

 

			<dict>
				<key>Comment</key>
				<string>Change AMD R9 xxx to RX 560</string>
				<key>Disabled</key>
				<false/>
				<key>Find</key>
				<data>
				UjkgeHh4
				</data>
				<key>Name</key>
				<string>AMD9500Controller</string>
				<key>Replace</key>
				<data>
				UlggNTYw
				</data>
			</dict>

[dreamcat4]

Some benchmarks of RX460 2gb on HighSierra (17A291j).

 

[edited] OK so i understand there is still no resolution yet: The black screen during clover boot for rx 460. I also searched for uefi GOP driver BIOS mod (for my rx 460 card)... cannot find any.

 

And that Mork vom Ork tried 1 month ago to dig into it, but unfortunately no luck either.

[ITzTravelInTime]

so guys, what have i to do if i want to use an rx 480 in my hackintosh? i want to use it as card for open cl in rendering in final cut and other software and use my nvidia card for the display

[stehor]

read the thread then u have success

[Max.1974]

so guys, what have i to do if i want to use an rx 480 in my hackintosh? i want to use it as card for open cl in rendering in final cut and other software and use my nvidia card for the display

Hi my friend, try used this kexts on post with Kext Utility by your on risk, because the gentlemans that deleted my post have frayd to said the another great site help us hackintoshers like this site too, and both just try help everyone!!

 

So just add on your config plist this:

 

 

 

<key>KextsToPatch</key>

<array>

 

<dict>

<key>Comment</key>

<string>Change_#_of_RX4x0_CUs-©_by_okrasit_2016</string>

<key>Disabled</key>

<false/>

<key>Find</key>

<data>SLgCAAAAAQAAAEiJQ1THQ3wIAAAA</data>

<key>Name</key>

<string>AMDRadeonX4100</string>

<key>Replace</key>

<data>SLgEAAAAAQAAAEiJQ1THQ3wSAAAA</data>

</dict>

<dict>

<key>Comment</key>

<string>Remove_CU_limit_of_RX4x0-©_by_okrasit_2016</string>

<key>Disabled</key>

<false/>

<key>Find</key>

<data>D0LIiYuAAAAARIizmQAAAESIcyA=</data>

<key>Name</key>

<string>AMDRadeonX4100</string>

<key>Replace</key>

<data>kJCQiYuAAAAARIizmQAAAESIcyA=</data>

</dict>

<dict>

<key>Comment</key>

<string>Change_init_from_BAFFIN_to_ELLESMERE-©_by_Fl0r!an_2016</string>

<key>Disabled</key>

<false/>

<key>Find</key>

<data>6EmF/v++SAEAAEyJ9w==</data>

<key>Name</key>

<string>AMDRadeonX4100</string>

<key>Replace</key>

<data>6EbkAAC+SAEAAEyJ9w==</data>

</dict>

<dict>

<key>Comment</key>

<string>PP_DisablePowerContainment=1</string>

<key>Disabled</key>

<false/>

<key>InfoPlistPatch</key>

<true/>

<key>Name</key>

<string>AMD9500Controller</string>

<key>Find</key>

<data>PGtleT5QUF9EaXNhYmxlUG93ZXJDb250YWlubWVudDwva2V5PjxpbnRlZ2VyPjA8L2ludGVnZXI+</data>

<key>Replace</key>

<data>PGtleT5QUF9EaXNhYmxlUG93ZXJDb250YWlubWVudDwva2V5PjxpbnRlZ2VyPjE8L2ludGVnZXI+</data>

</dict>

<dict>

<key>Comment</key>

<string>Change "R9xxx" to "RX 480" by CONATH</string>

<key>Disabled</key>

<false/>

<key>Find</key>

<data>

OSB4eHgAQVRZLFBhcnQjAA==

</data>

<key>Name</key>

<string>AMD9500Controller</string>

<key>Replace</key>

<data>

WCA0ODAAAAAAAAAAAAAAAA==

</data>

</dict>

 

And add on Clover Configurator > Devices > ATI 0x67DF1002

RX 480 KEXTS.zip

[Gigamaxx]

so guys, what have i to do if i want to use an rx 480 in my hackintosh? i want to use it as card for open cl in rendering in final cut and other software and use my nvidia card for the display

 

Sierra 12.6 and High Sierra dont require any fixes for RX cards except for the R9 xxx name.     As long as you have a Symbios iMac 14.2-18.1 or Mac Pro 4.1 + they wget full acceleration and 36 CUs without any need to alter kexts.

 

Nvidia card in slot 0 Rx in Slot 1 for a X99 type setup usually works.  Here's my Risen 7 1700 Rx 480 using Nvidia Gt210 as IGPU or Helper card.

[telepati]

Sierra 12.6 and High Sierra dont require any fixes for RX cards except for the R9 xxx name.     As long as you have a Symbios iMac 14.2-18.1 or Mac Pro 4.1 + they wget full acceleration and 36 CUs without any need to alter kexts.

 

Nvidia card in slot 0 Rx in Slot 1 for a X99 type setup usually works.  Here's my Risen 7 1700 Rx 480 using Nvidia Gt210 as IGPU or Helper card.

 

IGPU primary and still has black screen right?

[Ur-Quan]

IGPU primary and still has black screen right?

 

Nope

[MacNB]

Nope

You mean "yes" since the iGPU is still required as primary AND you still get black screen on the AMD ports during boot. Right ?

[surfinchina]

You mean "yes" since the iGPU is still required as primary AND you still get black screen on the AMD ports during boot. Right ?

I would have thought so.

However, I found somewhere, and can't find it now, a vega hackintosh which didn't need igpu and didn't have black screen and worked oob. I'm dubious until there is more evidence though.

Who knows, if it's true they're quite likely to extend that feature through all the next gen GPUs for Mac Pro - 5xxx, R9 Fury and Vega.

[jnolla]

Sierra 12.6 and High Sierra dont require any fixes for RX cards except for the R9 xxx name.     As long as you have a Symbios iMac 14.2-18.1 or Mac Pro 4.1 + they wget full acceleration and 36 CUs without any need to alter kexts.

 

Nvidia card in slot 0 Rx in Slot 1 for a X99 type setup usually works.  Here's my Risen 7 1700 Rx 480 using Nvidia Gt210 as IGPU or Helper card.

 

I'm still struggling getting RX580 working with Vega as IGPU for an X99 platform on High Sierra. I tried injecting different FB through the clover startup menu, specifically for the RX580 without any luck. Is there anything needed on the clover config for IGPU on x99? As I'm troubleshooting, I was thinking about PCI-E assignment for MacOS.

 

On my standard setup, I have the main GPU on the 16/X8_1 slot, and the secondary (RX580) on the 16/X8_3 slot. On this configuration it boots to black screen. No changes even if apply different FB names (Guariba, Orinoco, Dayman). When I look at the IOREG output, the VEGA card is on slot BR3A@3

 

When I swap cards and install the main GPU on the 16/X8_3 slot, and the secondary (RX580) on the 16/X8_1 slot. I get reboots @ gioscreenlockstate. No changes even if apply different FB names (Guariba, Orinoco, Dayman). When I look at the IOREG output, the VEGA card is on slot BR2A@2

 

I figured that because the OS sees 16/X8_3 slot as BR2A@2, maybe the VEGA needs to be in this slot? How can I get it to stop rebooting?

 

[jnolla]

I would have thought so.

However, I found somewhere, and can't find it now, a vega hackintosh which didn't need igpu and didn't have black screen and worked oob. I'm dubious until there is more evidence though.

Who knows, if it's true they're quite likely to extend that feature through all the next gen GPUs for Mac Pro - 5xxx, R9 Fury and Vega.

 

All evidence is here. Nothing dubious about it. 

http://www.insanelymac.com/forum/topic/325120-vega-frontier-edition-on-high-sierra/page-2

[Gigamaxx]

I'm still struggling getting RX580 working with Vega as IGPU for an X99 platform on High Sierra. I tried injecting different FB through the clover startup menu, specifically for the RX580 without any luck. Is there anything needed on the clover config for IGPU on x99? As I'm troubleshooting, I was thinking about PCI-E assignment for MacOS.

 

On my standard setup, I have the main GPU on the 16/X8_1 slot, and the secondary (RX580) on the 16/X8_3 slot. On this configuration it boots to black screen. No changes even if apply different FB names (Guariba, Orinoco, Dayman). When I look at the IOREG output, the VEGA card is on slot BR3A@3

 

When I swap cards and install the main GPU on the 16/X8_3 slot, and the secondary (RX580) on the 16/X8_1 slot. I get reboots @ gioscreenlockstate. No changes even if apply different FB names (Guariba, Orinoco, Dayman). When I look at the IOREG output, the VEGA card is on slot BR2A@2

 

I figured that because the OS sees 16/X8_3 slot as BR2A@2, maybe the VEGA needs to be in this slot? How can I get it to stop rebooting?

 

Screen Shot 2017-07-24 at 8.21.30 AM.png

It is hard to get dual AMD cards to work, my Haswell pentium supports 2 RX 480s using IGPU (Framebuffer only no acceleration) the Polaris cards work well in some open CL apps, a slight reduction in open GL performance but still working.

 

However, if using Braodwell, Skylake or KabyLake systems it is very hard if not impossible to get dual AMD Polaris cards to work. There is one person I know of who has disabled most of his IGPU drivers on a Skylake and has partial dual card support but like my setup with the Haswell pentium.

 

If you really want to get RX 480 working you may need to get a cheap Gt210 or other Nvidia card as helper or simulated iGPU. $30 US for a cheap helper card, I have found the Nvidia cards work great for this purpose and usually allow full sleep wake function for the RX.

 

I think you will have better luck running dual Vegas than a combo of Polaris and Vega. It may work with your Skylake or Kaby setup if you disable the IGPU drivers.

[jnolla]

It is hard to get dual AMD cards to work, my Haswell pentium supports 2 RX 480s using IGPU (Framebuffer only no acceleration) the Polaris cards work well in some open CL apps, a slight reduction in open GL performance but still working.

 

However, if using Braodwell, Skylake or KabyLake systems it is very hard if not impossible to get dual AMD Polaris cards to work. There is one person I know of who has disabled most of his IGPU drivers on a Skylake and has partial dual card support but like my setup with the Haswell pentium.

 

If you really want to get RX 480 working you may need to get a cheap Gt210 or other Nvidia card as helper or simulated iGPU. $30 US for a cheap helper card, I have found the Nvidia cards work great for this purpose and usually allow full sleep wake function for the RX.

 

I think you will have better luck running dual Vegas than a combo of Polaris and Vega. It may work with your Skylake or Kaby setup if you disable the IGPU drivers.

 

Keeping my hopes up. I can have them boot, but no acceleration if I fake a random ID. It might be something missing in Clover.

 

 

[MacNB]

All evidence is here. Nothing dubious about it. 

http://www.insanelymac.com/forum/topic/325120-vega-frontier-edition-on-high-sierra/page-2

Not quite.

You have not shown the steps you taken to show a single Vega works without any modifications to kext, etc.

You have not shown that when booting via Clover, that you can boot the Mac OS in Verbose mode and see the output on a monitor attached to the Vega port.

You have not posted the IOREG file showing a system setup with a single Vega card while the Intel GPU (iGPU) is disabled.

You have mixed up your issues of trying get 580 working together with Vega which makes it difficult to understand what you are doing.

 

Keep up the good work.

[Brumbaer]

I can not speak for Polaris, but the Vega, does work without helper card or igpu and it displays info from the BIOS start screen into macos.

And no I will not waste time to make a video, believe it or not or get a vega card and check yourself.

 

The drivers are not very mature and are still worked on. In DP4 the LuxMark and Geekbench OpenCL benchmarks do not work anymore. In DP3 they did.

 

Under DP3 the OpenCL benchmarks were at par with those of the R9Nano. But under Windows they are about 1.5 times as high as their R9 counterparts. Which makes sense as the clock is suppose to be about 1.5 times higher.

IoReg and LuxMark claim that the card runs with 1600 MHz, which might be true or not.

The card gets extremely hot under macos, so it might be that the max. freq. of the card is set to 1600MHz, but the actual frequency never gets as high because of temperature.

 

Hopefully DP5 will address those problems.

[surfinchina]

I can not speak for Polaris, but the Vega, does work without helper card or igpu and it displays info from the BIOS start screen into macos.

And no I will not waste time to make a video, believe it or not or get a vega card and check yourself.

 

The drivers are not very mature and are still worked on. In DP4 the LuxMark and Geekbench OpenCL benchmarks do not work anymore. In DP3 they did.

 

Under DP3 the OpenCL benchmarks were at par with those of the R9Nano. But under Windows they are about 1.5 times as high as their R9 counterparts. Which makes sense as the clock is suppose to be about 1.5 times higher.

IoReg and LuxMark claim that the card runs with 1600 MHz, which might be true or not.

The card gets extremely hot under macos, so it might be that the max. freq. of the card is set to 1600MHz, but the actual frequency never gets as high because of temperature.

 

Hopefully DP5 will address those problems.

 

Thanks for the info Brumbaer

Good news re the helper card and showing the bios information!

Bad news re the speed and the heat... Hopefully they will sort it.

 

Is it using all the ports?

And, my nano won't work accelerated in HS so vega still seems the way to go, unless I stick with Sierra. Have you got your nano working in HS? 

[fantomas]

Guys, I have neither Polaris nor Vega but I'm just as excited for you about it as much as you are.    But you know that we have dedicated topics to their discussions.  

 

I know your discussions will swim through these two architectures and I understand, I really do, but please do your best to be focused on the main discussion.

 

BR

 

fantomas1

[Brumbaer]

Thanks for the info Brumbaer

Good news re the helper card and showing the bios information!

Bad news re the speed and the heat... Hopefully they will sort it.

 

Is it using all the ports?

And, my nano won't work accelerated in HS so vega still seems the way to go, unless I stick with Sierra. Have you got your nano working in HS? 

All ports on the Vega can be used. I tried each on on it's own, but only  two simultaneously. I haven't tested whether all 4 work simultaneously.

The R9Nano does not work under HS. I assume there is a driver at Apple seemingly using the same structure as the Vega Driver as hinted by some files in DP3.

But they are gone under DP4. The Vega Kexts do not hint openly that they would be Fijii compatible. Some Baffin kext might be, but I didn't find one.

With hindsight  I wouldn't have bought a Vega Frontier yet, I'd wait till it's clear wether it will be running smoothly and will be here to stay (in macos), it would be a pity to have a second dead end like the R9Nano. 

Let's stop the Vega/Nano discussion here, before we overstay our welcome 

[joalmoju]

http://www.insanelymac.com/forum/topic/321371-lilu-—-kext-and-process-patcher/page-3?hl=%2Bwhatevergreen#entry2469740

[firefly_can_fly]

Guys, i have a question - does anyone who have XFX RX 460 or XFX RX 560 managed to make it even work, on any bootloader, but without dynamically adding other bios or etc?

[Cyberdevs]

Boot to black screen is over for AMD GPUs

 

Thanks to @vit9696 and @vandroiy2012 with the new Lilu.kext v1.1.6 and WhateverGreen.kext v.1.0.0 now we can use AMD GPUs without the need of a helper card.