Contribute
Register

macOS Sierra PB: Need testers for new AMD Radeon drivers!

Status
Not open for further replies.
This seems to be happening a lot in various places. PikeRAlpha has a post up on his site telling how something similar happened to him.

I am using the public beta, if you go back a couple of posts, I am seeing the same 9500 / 4200 combination when using the generic approach and getting the 9515 / 4200 combination when using a named frame buffer. I have not checked very far into the 9200 controller yet.

Have you had a chance to test your RX 560 card with a newer cpu to see if it gets 10 bit graphics?

No, not yet but I'm sure it will, I'll slap my 4690k in and give some results tonight. I've been doing some video capture with my sons ElGato on the pentium and it wants 4 cores, but it works and I can hear the card spinning during the rendering and edits.
 
This seems to be happening a lot in various places. PikeRAlpha has a post up on his site telling how something similar happened to him.

I am using the public beta, if you go back a couple of posts, I am seeing the same 9500 / 4200 combination when using the generic approach and getting the 9515 / 4200 combination when using a named frame buffer. I have not checked very far into the 9200 controller yet.

Have you had a chance to test your RX 560 card with a newer cpu to see if it gets 10 bit graphics?


Something else just happened recently as well the 2 factor authentication for me is now mandatory. It started about a week ago and as often as I change and reload disks I get a lot of phone texts to add code. This is a real pain I prefer the simple single password verification.
 
As I understand it, when you disable the IGPU in your motherboard's BIOS settings, the motherboard then turns around and initializes all of the big pci slots so that any slot with a graphics card will give you some kind of graphics support. This seems to cause a problem for newer AMD cards because they do not like being re-initialized while running. There are various proposals out there as to how to fix the issue, such as re-write motherboard bioses -- a daunting task that is sure to be met with open hostility by motherboard manufactureres, an idea to issue a pci reset command during the boot process which would cause the monitor to black out for a time and then come back on, but could be very tricky to pull off as a general solution. Or root through code from other os builders to see how they pull it off, but that has issues with licensing.

Hmmm. Thanks for the explanation. Another member here ( @zakklol ) mentioned the re-initialisation issue as being the problem with AMD cards. I hadn't realised the cause.

So despite the new AMD GPU chips themselves being similar to the latest genuine Apple-used items, the video-bioses are what are causing the conflicts (?) Seems it's going to be a tough time for hackintosh GPUs over the next few months if Nvidia are no longer in genuine Macs and AMD have bios problems...
 
Something else just happened recently as well the 2 factor authentication for me is now mandatory. It started about a week ago and as often as I change and reload disks I get a lot of phone texts to add code. This is a real pain I prefer the simple single password verification.

I too prefer password verification.
 
Hmmm. Thanks for the explanation. Another member here ( @zakklol ) mentioned the re-initialisation issue as being the problem with AMD cards. I hadn't realised the cause.

So despite the new AMD GPU chips themselves being similar to the latest genuine Apple-used items, the video-bioses are what are causing the conflicts (?) Seems it's going to be a tough time for hackintosh GPUs over the next few months if Nvidia are no longer in genuine Macs and AMD have bios problems...

Nahhh, :) Nvidia will keep on releasing web-drivers and AMD users will keep using helper cards. It has been that way for some time now. But it is more of a problem for folks that use X99 types of motherboards than it is for folks that use IGPU type motherboards, so I hope folks keep this in mind as they are digging through the various code bases to solve other problems. There are other video card related issues that some consider more important at the moment than this one, so we will have to see what turns up. All part to the fun of this hobby.
 
...

My RX 560 is running on 9500controller and 4200kext. 9500 makes a comeback and the 4200kext is brand new.

I was surprised to see that the 9500 (0x67E01002 0x67EF1002 0x67FF1002 0x67C01002 0x67DF1002)
was being used as the default controller, I mostly expected to see the 9520 being used. I am not holding out any hope that the 9520 controller will gives us audio but I think Apple is giving a a chance to get the graphics part worked out before the public betas start being released later this month.
 
They did yank the 9500 before, Apple giveth, Apple Takith away.

I like the CO coming back in there HBM, HBM2? No time for now to check. Running benches on 560 in 12.6.3 public beta. I'm liking this so far it's holding its own even with a helper card.

RX 480 fired right up, and 36 CUs, this is gonna put a few threads out of business.

12.6.3 Luxmark 36 cu.png

12.6.3 BETA GEEKBENCH RX 480.png
 

Attachments

  • Screen Shot 2017-06-14 at 7.42.26 PM.png
    Screen Shot 2017-06-14 at 7.42.26 PM.png
    1.7 MB · Views: 129
  • Screen Shot 2017-06-14 at 7.48.41 PM.png
    Screen Shot 2017-06-14 at 7.48.41 PM.png
    2.6 MB · Views: 133
Last edited:
first id like to say sorry for my bad and limited ability of expression with english since its not my first language
well,as we know sierra and high sierra cannot booted without igpu, id like to share some of my "observation" about this problem. in fact i want to write this a while ago, but im busying with other things so i quickly forgot about that
correct me if am wrong,heres is the known ways to boot sierras without igpu
1.flash the card with legacy bios ONLY
if we stripped the UEFI from vbios, polaris would boot fine without igpu or helper card, thats the first hint for me which problem lying in the difference of initialize the EFI between real mac and hackintosh
2.change the device id in UEFI Vbios
if we change the Vbios device id from 67DF to anything not used by any card(like 67DD), the card would boot fine even under UEFI Vbios, i tried to using clover's load vbios function to load the very same bios, i got black screen just as theres no igpu, while flashing it to card and booting with clover results in success booting. thats the second hint for me, solution must lying in the way clover deal with uefi vbios. As we can see, load vbios failed to show same result as flashing the vbios directly into card, thats imply clover would load basic info(like device id) from vbios then passing these info to the system regardless what we changed in vbios file.
Im not intended to nail the exact cause of these problem, but i come up a workaround, if clover could let us pass any info we want to the system(like custom device id), or making a DSDT patch which fake the device id for polaris card, would it solve this problem?
 
Last edited:
I wonder if there would be a way to use EDID to send fake ID 0x67DD1002, and load Vbios at the same time? Asking clover to do 2 things simultaneously.

Have you tried using older version of clover for load Vbios, I know the clover team is busy adding many things to get Sierra to recognize the Polaris cards. Maybe a clover that doesn't try so hard to inject Polaris would let the Vbios load.


This is very interesting and your English is fine, thanks for the observations.

No sav, to habla espanol? In pregunta, tu cheque una version Clover viejo para load Vbios? Primero versionnes no hay heriamentes para injection Polaris? De versionnes nuevas tiene muchos cosas para Sierra conose Polaris.
 
I wonder if there would be a way to use EDID to send fake ID 0x67DD1002, and load Vbios at the same time? Asking clover to do 2 things simultaneously.

Have you tried using older version of clover for load Vbios, I know the clover team is busy adding many things to get Sierra to recognize the Polaris cards. Maybe a clover that doesn't try so hard to inject Polaris would let the Vbios load.


This is very interesting and your English is fine, thanks for the observations.

No sav, to habla espanol? In pregunta, tu cheque una version Clover viejo para load Vbios? Primero versionnes no hay heriamentes para injection Polaris? De versionnes nuevas tiene muchos cosas para Sierra conose Polaris.
i havent tried using older version of clover , but definitely would to, weekend maybe
ah im not speaking spanish :p
 
Status
Not open for further replies.
Back
Top