Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Status
The first post of this thread is a WikiPost and can be edited by anyone with the appropiate permissions. Your edits will be public.
You really have a fetish with the first post, that is not a proper fix, a proper fix would be at VBIOS level or BIOS level, but as usual instead having a chat with REAL information we lose time pointing to stuff that I saw already.
Sorry if my questions looks like a n00b one but I really seek for dev answers...

If the dev’s had a answer it would be in post 1 already.
 
  • Like
Reactions: dfranetic
You really have a fetish with the first post, that is not a proper fix, a proper fix would be at VBIOS level or BIOS level, but as usual instead having a chat with REAL information we lose time pointing to stuff that I saw already.
Sorry if my questions looks like a n00b one but I really seek for dev answers...
Apple used the same plist approach to achieve the same thing. The problem even existed and exists on official supported systems (I assume mostly equipped with NVIDIA GOUs) - you will find a lot of hits when searching the "black screen" in the depth of the web.

Apple did not plan to have NVIDIA cards installed into these iMacs - so it is in IMHO a valid approach. Your definition of *proper* seems to be not the only existing one.

The heavy 2009 problem could be in fact cured by modding the iMac firmware with parts of the 2010 model. This would be a proper fix (at least according to your system of definitions), but because patching the firmware has not been a real success story here (search for EFI boot screen mods of AMD cards) I would not follow really this way.

Add your definition of proper in your own first post here - if you seek precise answers ask more precise questions - it is that easy!

EDIT: A lot of owners of NVIDIA cards would appreciate a vBIOS based fix, I am sure!
 
Last edited:
amdvbflash shows valid version and product name for the EEPROM contents - so I'd say that flashing has indeed been successful. It doesn't show anything for the new VBIOS you are flashing, but if I remember correctly that happens when old and new VBIOS have identical version and product name.

As the translation into English isn't very clear, I maybe didn't understand your actual problem yet. So you try to flash the VBIOS but the old VBIOS stays in place?
Maybe you're trying to flash the VBIOS in an HP machine? Many HP machines will always use and load the VBIOS from the system firmware and not from the graphics card itself - and it is VERY difficult to change that VBIOS in the system firmware.
Not in the HP model refresh VBiOS Oh, is the use of MXM PCIE development debug card adapter plug in the PC desktop refresh or CH341A different VBiOS are the same results.

In the HP notebook disassembly MXM graphics card is very inconvenient part of HP's MXM graphics card VBiOS is integrated in the motherboard BIOS is not in the graphics card EEPROM are blank this I know, such MXM PCIE development debug card is very convenient MXM-A MXM-B AMD NVIDIA MXM3.0 MXM3.1 from the oldest gt103 to The latest rtx5000 gtx3090 are supported will not be incompatible because of various VBiOS problems.

Translated with www.DeepL.com/Translator (free version)
 

Attachments

  • IMG_9649.JPG
    IMG_9649.JPG
    509.8 KB · Views: 93
  • IMG_9666.JPG
    IMG_9666.JPG
    443.7 KB · Views: 94
  • IMG_9667.JPG
    IMG_9667.JPG
    500.2 KB · Views: 89
  • Like
Reactions: Nguyen Duc Hieu
Hi everyone, I'm reporting a successful installation of a GTX 870M (N15E-GT-A2) in a 12,2 (27-inch Mid 2011).
I've followed all the steps described here, and it's running great in Big Sur, with the help of OCLP.

The last problem that I'm facing is the maximum brightness level, that's at around 70% of the original card. From what I've read in the first post and in other replies, the kext files necessary to enable full brightness should be already installed with OCLP and Big Sur.

I've tried to add (drag and drop) the AppleBacklightFixup in Kext Utility v2.6.6, but it doesn't help.
What am I missing ?

Thank you all for your guidance and hard work. It's much appreciated.
 
Hi everyone, I'm reporting a successful installation of a GTX 870M (N15E-GT-A2) in a 12,2 (27-inch Mid 2011).
I've followed all the steps described here, and it's running great in Big Sur, with the help of OCLP.

The last problem that I'm facing is the maximum brightness level, that's at around 70% of the original card. From what I've read in the first post and in other replies, the kext files necessary to enable full brightness should be already installed with OCLP and Big Sur.

I've tried to add (drag and drop) the AppleBacklightFixup in Kext Utility v2.6.6, but it doesn't help.
What am I missing ?

Thank you all for your guidance and hard work. It's much appreciated.
Yes, it is already included. It is not any longer a kernel extension, it is just a nice OC based patch. So you can do literally nothing. Installing the AppleBacklightFixup will not help in your situation.

Please read back the last two pages and find out and post your LCD panel type. If it is a non standard one we can can dive deeper.
 
Yes, it is already included. It is not any longer a kernel extension, it is just a nice OC based patch. So you can do literally nothing. Installing the AppleBacklightFixup will not help in your situation.

Please read back the last two pages and find out and post your LCD panel type. If it is a non standard one we can can dive deeper.
Indeed, when running the ioreg command, I'm getting this :

??????
?<"x"o??UL
???)P0 5

???)P0 5

And when I check the profile of the display, under "mmod" it says the model is - 0000A007
 
Indeed, when running the ioreg command, I'm getting this :

??????
?<"x"o??UL
???)P0 5

???)P0 5

And when I check the profile of the display, under "mmod" it says the model is - 0000A007
I need both Manufacturer 00000610 for me and Model 0000A007 for me in my iMac12,2.

If you have the same just re create the OCLP EFI folder with your NVIDIA card installed and replace the old version with the newly created one and reboot into this OC version and into Big Sur.

If this still does not work I have no idea.
BTW: You will not get the old 100% back!
 
I need both Manufacturer 00000610 for me and Model 0000A007 for me in my iMac12,2.

If you have the same just re create the OCLP EFI folder with your NVIDIA card installed and replace the old version with the newly created one and reboot into this OC version and into Big Sur.

If this still does not work I have no idea.
BTW: You will not get the old 100% back!
Yes, Manufacturer is also 00000610

I hope I understand you correctly :
To re create the EFI folder, I ran OCLP-TUI, then built OpenCore (option 1) and then installed OpenCore to the internal drive (option 2) and booted from the EFI partition newly created.

If so, still no luck. The level stays the same. It's definitely usable, but another 10% would have been so much better.
Thanks for your time.
 
Apple used the same plist approach to achieve the same thing. The problem even existed and exists on official supported systems (I assume mostly equipped with NVIDIA GOUs) - you will find a lot of hits when searching the "black screen" in the depth of the web.

Apple did not plan to have NVIDIA cards installed into these iMacs - so it is in IMHO a valid approach. Your definition of *proper* seems to be not the only existing one.

The heavy 2009 problem could be in fact cured by modding the iMac firmware with parts of the 2010 model. This would be a proper fix (at least according to your system of definitions), but because patching the firmware has not been a real success story here (search for EFI boot screen mods of AMD cards) I would not follow really this way.

Add your definition of proper in your own first post here - if you seek precise answers ask more precise questions - it is that easy!

EDIT: A lot of owners of NVIDIA cards would appreciate a vBIOS based fix, I am sure!
The black screen bug I am talking is about the early init of the card where there is no plist or kernel loaded, the display fire up only if alt is pressed, and if not and any osx/macos is booted it remains black, windows will fire the display proper, this only happen on k610m, k1000m, k1100m and k2100m. To be "c l e a r" this bug is on 2011 27" iMac not tested on 2009 or 2010! And not present on k3000m k3100m or k4100m that I tested! I don't see the logic on how Apple can play a role on this since we are talking on same kepler arch, IMHO is the VBIOS bug, more precise the efi part of it that was used for those cards that I mentioned earlier. I am still hopping for a good clue on what/how/where/maybe doable fix can be done.
 
Indeed, when running the ioreg command, I'm getting this :

??????
?<"x"o??UL
???)P0 5

???)P0 5

And when I check the profile of the display, under "mmod" it says the model is - 0000A007
@todor7 in 21.5" 2011 I have recently received, I also get similar output. On your system, with your card, do you have brightness slider available at all?
 
The black screen bug I am talking is about the early init of the card where there is no plist or kernel loaded, the display fire up only if alt is pressed, and if not and any osx/macos is booted it remains black, windows will fire the display proper, this only happen on k610m, k1000m, k1100m and k2100m. To be "c l e a r" this bug is on 2011 27" iMac not tested on 2009 or 2010! And not present on k3000m k3100m or k4100m that I tested! I don't see the logic on how Apple can play a role on this since we are talking on same kepler arch, IMHO is the VBIOS bug, more precise the efi part of it that was used for those cards that I mentioned earlier. I am still hopping for a good clue on what/how/where/maybe doable fix can be done.
You cannot see, it? I can!

After changing the GPU one has very often (or sometimes) to reset the PRAM to get it working. Unfortunately with this action the black screen drama often starts.

The current cure with the plist approach does not survive a PRAM reset, but just booting into a plist patched macOS version sets obviously a bit again stored *permanently* in the PRAM to have a working internal LCD on the next boot, directly after post (even without pressing alt/option on boot). Additionally it fires up the internal LCD (after the PRAM reset) and a complete boot. So we are done here, I hope - I have never checked which PRAM bit it is!

The problem of the Late 2009 is, you cannot simply press alt/option to get the internal screen fired up unlike the Mid 2010 and 2011, where you can force the screen to turn on, again. This is more severe! Therefore I currently tend to patch those system in advance or use OC with the agdpmod=vit9696 boot-arg.

I have tested it on all systems with a lot of different NVIDIA cards, not only on the Mid 2011. I agree, a vBIOS fix (which has to be applied to all published versions) would be a great achievement. Unfortunately it happens with all NVIDIA cards, trust me, all cards (the only one I have never tested myself is the 870/880) on the Late 2009 and Mid 2010 and with some on the 2011.

So I cannot agree with your analysis that it is a simple vBIOS problem of only a few cards limited to the Mid 2011.
 
@todor7 in 21.5" 2011 I have recently received, I also get similar output. On your system, with your card, do you have brightness slider available at all?
@Tom.Tom Yes, I get a slider and theoretically goes to 100%, but the actual brightness is way less than that. I was just comparing it side by side to another 2009 27" Imac, and it looks more like 50-60% of the original level.
 
Hi! I have a question: where booscreen is located?

I mean some ami bioses are contain bootswitch program that can do same as bootloader.

But is it possible to access bootscreen without any drive or ssd connected?

I mean that software should contain Unicode text with proposal to connect to WiFi. The WiFi keys are located in nvram?


@Ausdauersportler , @internetzel

?
 
....

The current cure with the plist approach does not survive a PRAM reset, but just booting into a plist patched macOS version sets obviously a bit again stored *permanently* in the PRAM to have a working internal LCD on the next boot, directly after post (even without pressing alt/option on boot). Additionally it fires up the internal LCD (after the PRAM reset) and a complete boot. So we are done here, I hope - I have never checked which PRAM bit it is!
...
Finally getting some real info from you! :)
Well this is what I want to avoid, booting into a patched OS to "repair" the problem.
If nobody know what NVRAM variable needs to be set to get it done, I will dump the NVRAM before and after to see what change where made.
Like I said, best would be to fix it in VBIOS.
My analysis is based on what cards I had in hand, still is "weird" why only on lower end cards I had this "bug".
 
Finally getting some real info from you! :)
Well this is what I want to avoid, booting into a patched OS to "repair" the problem.
If nobody know what NVRAM variable needs to be set to get it done, I will dump the NVRAM before and after to see what change where made.
Like I said, best would be to fix it in VBIOS.
My analysis is based on what cards I had in hand, still is "weird" why only on lower end cards I had this "bug".
If it's really just a variable in the NVRAM, it might be quite easy to write an EFI driver that does that same change.
That EFI driver could be loaded from a file system by OpenCore but could also be appended to the VBIOS and hence be loaded automatically. There's even another possibility to have an EFI driver loaded from a file system on bootup automatically, the "Driver####" approach which consists in adding a special NVRAM variable.
 
  • Like
Reactions: THeKiNGs
Hey. Is it possible to install an amd firepro m5100 or m6100 in a iMac 2011 21.5 inch? They are at a pretty good price currently on eBay and wanted to make sure that if I picked one up than it would be supported. Thanks in advance
 
Hi! I have a question: where booscreen is located?

I mean some ami bioses are contain bootswitch program that can do same as bootloader.

But is it possible to access bootscreen without any drive or ssd connected?

I mean that software should contain Unicode text with proposal to connect to WiFi. The WiFi keys are located in nvram?


@Ausdauersportler , @internetzel

?
There's a boot picker application which is stored in the EFI firmware. Another EFI application with "boot screen" is the target disk mode which provides access to a Mac's hard drive via FireWire or Thunderbolt. Another one not available on older models (2009 and 2010?) is internet recovery.
Those applications can be run by without any drive connected, by keeping certain keys pressed after powering on.
And yes, the WiFi keys are stored in NVRAM, at least on models with internet recovery.
 
Hi friends, I need help with OCLP
I have 765m card Installed A1312,, I did update High sierra to Catallina Using Dosdude,I have boot screen, Catallina works fine, but Not Brightness adjust, How do I install OCLP Phttps://github.com/dortania/OpenCore-Legacy-Patcher in my iInstallation Catallina? I can't understand the steps
 
If it's really just a variable in the NVRAM, it might be quite easy to write an EFI driver that does that same change.
That is my plan to make a simple driver that would check if variable exist on boot and if not set it, if there is no other possible/permanent fix.
I will also check on OC source what agdpmod=vit9696 does or even better ask vit9696 for help :)
 
Hey. Is it possible to install an amd firepro m5100 or m6100 in a iMac 2011 21.5 inch? They are at a pretty good price currently on eBay and wanted to make sure that if I picked one up than it would be supported. Thanks in advance
Check the tables on the first post for supported cards, MXM-B cards will not fit and run in a stable way in 21.5 models. The power draw is too high.
 
Last edited:
If it's really just a variable in the NVRAM, it might be quite easy to write an EFI driver that does that same change.
That EFI driver could be loaded from a file system by OpenCore but could also be appended to the VBIOS and hence be loaded automatically. There's even another possibility to have an EFI driver loaded from a file system on bootup automatically, the "Driver####" approach which consists in adding a special NVRAM variable.
We are going round in circles here? Loading something from the file system on boot is not a proper fix according to the weird definition we got and we have already such a solution called OpenCore with the agdpmod=vit9696.

Appending a fix directly to the NVIDIA vBIOS would be the only valid approach.

But:

Since all 2009/2010 systems use OC anyway to install post High Sierra macOS versions only the vBIOS fix would be a step forward. So it would be of real use only when doing the very first installation or adding a new virgin disk (which can be preconfigured with OC in advance). It is a boot strap problem, not one of daily use.

The cards @THeKiNGs mentioned (K610M, K1x00M, K2x00M) with the Mid 2011 all need OpenCore to enable brightness control. So one would add a solution not needing OpenCore just to install the very same OpenCore in the next step?

We need to have a definition of a proper analysis of a problem, too.

AMD cards do not have this problem.
 
Last edited:
@Ausdauersportler, can you confirm that to "re create the OCLP EFI folder", I need to :

Run OCLP-TUI on the machine running Big Sur itself with the NVIDIA 870M, then built OpenCore (option 1) and then install OpenCore to the internal drive (option 2) and boot from the EFI partition ?

Sorry but I just want to make sure that I'm not skipping something. Thank you.
 
Check the tables on the first post for supported cards, MXM-B cards will not fit and run in a stable way in 21.5 models. The power draw is too high.
According to https://www.techpowerup.com/gpu-specs/firepro-m5100.c2517, the FirePro m5100 is mxm-a therefore it would fit in my mac. The power withdraw is unknown though. While I was exploring the website, I found that the AMD FirePro w5170m has a base clock speed of 900MHz while the m5100 only has a 750MHz base clock speed. The AMD FirePro w5170m is also an mxm-a card this card can also fit into my mac. The card doesn't need external power for it to function, and the card is also a year newer. They go for about 80-85$ so price to performance is pretty good there.
 
There's a boot picker application which is stored in the EFI firmware. Another EFI application with "boot screen" is the target disk mode which provides access to a Mac's hard drive via FireWire or Thunderbolt. Another one not available on older models (2009 and 2010?) is internet recovery.
Those applications can be run by without any drive connected, by keeping certain keys pressed after powering on.
And yes, the WiFi keys are stored in NVRAM, at least on models with internet recovery.
After newer latest research found a lot of confirmations that Boot screen is powered by GOP driver too. So actually there are two methods: inject new GOP driver to platform bios; use that GOP driver and with tianacore build GOP OpRom for GPU (https://www.workofard.com/2020/12/aarch64-option-roms-for-amd-gpus/)
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.