iM 4,1-14,4 2011 iMac Graphics Card Upgrade

[dfranetic]

@AirpodsNow and @Surf2bikes : It seems that you both experience the same (random) sleep-related kernel panic on a WX4130 card with Elpida VRAM ICs. (I guess that you both have the same type of VRAM ICs, but only the last half of the part number is visible in @AirpodsNow 's photos.) In a PM @Surf2bikes informed me that he can see the OC bootpicker with the original WX4130 VBIOS. (So it should be easy to revert to @internetzel 's VBIOS in case the original one does not fire up the internal/external display.)

Try this (if it is within your comfort level): Flash the WX4130 with the original VBIOS, see if it boots into macOS on internal/external display, and (if it does) re-enable sleep. Test if the sleep issue persists.

There's a somewhat remote chance that this is related to Elpida VRAM ICs. It is a strange coincidence that the only two users (I know of) reported this issue - and you both have those VRAM ICs.

 

[antoniomac]

Hi, I haven't been able to find a solution for over 20 days
imac 11.3 - k2100M - Catalina
I managed to install the brightness adjustment but not the graphics acceleration
Please help me, I don't know what to do anymore

 

[Surf2bikes]

@AirpodsNow and @Surf2bikes : It seems that you both experience the same (random) sleep-related kernel panic on a WX4130 card with Elpida VRAM ICs. (I guess that you both have the same type of VRAM ICs, but only the last half of the part number is visible in @AirpodsNow 's photos.) In a PM @Surf2bikes informed me that he can see the OC bootpicker with the original WX4130 VBIOS. (So it should be easy to revert to @internetzel 's VBIOS in case the original one does not fire up the internal/external display.)

Try this (if it is within your comfort level): Flash the WX4130 with the original VBIOS, see if it boots into macOS on internal/external display, and (if it does) re-enable sleep. Test if the sleep issue persists.

There's a somewhat remote chance that this is related to Elpida VRAM ICs. It is a strange coincidence that the only two users (I know of) reported this issue - and you both have those VRAM ICs.

Thanks @dfranetic . I will give it a try. Please allow me some time to do it. I will report the test results here.

 

[internetzel]

@AirpodsNow and @Surf2bikes : It seems that you both experience the same (random) sleep-related kernel panic on a WX4130 card with Elpida VRAM ICs. (I guess that you both have the same type of VRAM ICs, but only the last half of the part number is visible in @AirpodsNow 's photos.) In a PM @Surf2bikes informed me that he can see the OC bootpicker with the original WX4130 VBIOS. (So it should be easy to revert to @internetzel 's VBIOS in case the original one does not fire up the internal/external display.)

Try this (if it is within your comfort level): Flash the WX4130 with the original VBIOS, see if it boots into macOS on internal/external display, and (if it does) re-enable sleep. Test if the sleep issue persists.

There's a somewhat remote chance that this is related to Elpida VRAM ICs. It is a strange coincidence that the only two users (I know of) reported this issue - and you both have those VRAM ICs.

As far as I know, Elpida VRAM is very common on WX4130s - and the VBIOS supports both Elpida EDW4032BABG and Samsung.

 

[Ausdauersportler]

Hi, I haven't been able to find a solution for over 20 days
imac 11.3 - k2100M - Catalina
I managed to install the brightness adjustment but not the graphics acceleration
Please help me, I don't know what to do anymore

First post of this thread,

Spoiler: Installing macOS Mojave/Catalina/Big Sur/Monterey using OpenCore (unsupported)

Install Catalina using OCLP with moderate spoofing - this will solve all of your problems. Do not use dosdude1 patcher!

The docs are unchanged for at least half a year, so you cannot really missed it there.

 

[basslib]

Congratulations!
Honestly I don't think Apple cares one bit about running an operating system they don't approve of on their hardware. With Hackintoshes, you can say it's a violation of Terms of Service because it says "For use on Apple branded systems". But with updating an actual Apple computer to a newer OS, there is nothing in the Terms of Service that prohibits it. You're perfectly safe from a legal standpoint to update.

Yes, you're right, in fact I took the right opportunity!


Also because I don't work with this computer, I don't produce with it and consequently, I don't earn money with it either.

The next step? When I have time, I will bring back my first ever Apple device, an iMac G5 PowerPC, to keep it in the best possible condition.

 

[Ausdauersportler]

Yes, you're right, in fact I took the right opportunity!

View attachment 1948454
Also because I don't work with this computer, I don't produce with it and consequently, I don't earn money with it either.

The next step? When I have time, I will bring back my first ever Apple device, an iMac G5 PowerPC, to keep it in the best possible condition.

Everything is working? Especially sleep & wake, too?
(M4000 in iMac11,2)

 

[basslib]

Everything is working? Especially sleep & wake, too?
(M4000 in iMac11,2)

I'll get off work in a few hours and report back. There was an improvement in performance in general and something I never expected and was very pleased about, the yellowish display is back to white.

 

[nikey22]

Hello everyone,

I've been testing my AMD Firepro W7170M card for a while now:

It is currently in it's experimental phase, and I'm still modifying the vbios rom as we speak to improve compatibility and give it some practical functionality in our class machines. There were some initial set backs with this card. There are still some power gating issues to work out as well with hangs on reboot that I'll have to speak to Aus about. But it shows promise.

-The AMD W7170M is basically a Radeon R9 M395X, possibly a M390X in some deviceIDs.
-The R9 M395X was a 'special upgrade' that people could order for their 27" iMac 5K 2015 and so therefore I figured it would have some native support in macOS.
-Apple uses the AMDTongaGraphicsAccelerator package coupled with the AMD9000Controller.kext to initialize the card.
-I was able to get both the internal & external monitors working when I figured out their respective Display_Object_Path i2cid & senseids.
-base vram is 1250Mhz, but I noticed that all modules that come with this card are capable of much higher frequencies with different timings:

tRCDW-tRCDWA-tRCDR-tRCDRA-tRC-tCL-tRFC
1375 MHz: 12-12-19-19-58-19-123 (Samsung K4G41325FC, 4096MB, GDDR5)
1375 MHz: 15-15-20-20-55-18-89 (Elipda EDW4032BABG, 4096MB, GDDR5)
1375 MHz: 14-14-19-19-57-17-136 (Hynix H5GC4H24AJR, 4096MB, GDDR5)

-The performance is actually great!



I will publish the vbios in due time after more experimentation.

insanely great!
-Steve Jobs

 

[vinaypundith]

Yes, that error 43 has always been plaguing the EG VBIOS versions. You should search the thread for this issue - @Nick [D]vB also struggled with it - and solved it but didn't tell what exactly was the solution: Feb 23, 2020 .

I'm pretty sure that's the exact post where I downloaded my EG VBIOS from. He seems to say that he fixed the issue in legacy BIOS booting mode but it persists in EFI mode? I don't quite understand what he meant by "display configuration". What tools are y'all using to examine and modify these vBIOSes? Or do yall understand the hex code that's in them directly? I'd love to learn, and be able to help figure out stuff too.

So I tried to go ahead and install Windows in legacy BIOS mode - and voila, managed to successfully destroy both the nicely working macOS 12.2 RC install and my OpenCore EFI, and still not have a working Windows installation! I was trying to do the kind of hybrid partition table needed to install Windows in legacy BIOS/MBR mode, but apparently the Boot Camp Assistant in macOS 12.2 RC can't even proceed without a Windows USB drive, it won't take a Windows installer DVD. And my iMac can't boot USB drives in legacy BIOS mode. So I tried to boot to a Linux live boot DVD (Linux Mint 19.1 to be specific, as it was the only Linux disk that I had that seemed to work) and use it's Disks tool to do the partitioning, since I know that it will create a hybrid partition table when you try to use it to add a new partition. Instead, when I resized the APFS volume to make space for the Windows partition, it corrupted the APFS volume and somehow made the OpenCore EFI unbootable. Oh well..... :-(

 

[basslib]

Everything is working? Especially sleep & wake, too?
(M4000 in iMac11,2)

Everything works like a charm, better than with High Sierra, no glitches at shutdown, faster and more optimised than ever!

Brightness control: works better than with the original GPU, faster response times.

Sleep and wake: nothing to say, works great.

 

[internetzel]

I'm pretty sure that's the exact post where I downloaded my EG VBIOS from. He seems to say that he fixed the issue in legacy BIOS booting mode but it persists in EFI mode? I don't quite understand what he meant by "display configuration". What tools are y'all using to examine and modify these vBIOSes? Or do yall understand the hex code that's in them directly? I'd love to learn, and be able to help figure out stuff too.

So I tried to go ahead and install Windows in legacy BIOS mode - and voila, managed to successfully destroy both the nicely working macOS 12.2 RC install and my OpenCore EFI, and still not have a working Windows installation! I was trying to do the kind of hybrid partition table needed to install Windows in legacy BIOS/MBR mode, but apparently the Boot Camp Assistant in macOS 12.2 RC can't even proceed without a Windows USB drive, it won't take a Windows installer DVD. And my iMac can't boot USB drives in legacy BIOS mode. So I tried to boot to a Linux live boot DVD (Linux Mint 19.1 to be specific, as it was the only Linux disk that I had that seemed to work) and use it's Disks tool to do the partitioning, since I know that it will create a hybrid partition table when you try to use it to add a new partition. Instead, when I resized the APFS volume to make space for the Windows partition, it corrupted the APFS volume and somehow made the OpenCore EFI unbootable. Oh well..... :-(

I guess that issue might be solved by using Apple's Windows graphics drivers - but those drivers need the subsystem vendor and device IDs to match exactly.
I could modify the existing VBIOS accordingly - but at least for the Ellesmere-GPUs (WX7100, RX480) the correct subsystem IDs make the macOS drivers crash.

 

[internetzel]

Hello everyone,

I've been testing my AMD Firepro W7170M card for a while now:

View attachment 1948589

It is currently in it's experimental phase, and I'm still modifying the vbios rom as we speak to improve compatibility and give it some practical functionality in our class machines. There were some initial set backs with this card. There are still some power gating issues to work out as well with hangs on reboot that I'll have to speak to Aus about. But it shows promise.
View attachment 1948581
-The AMD W7170M is basically a Radeon R9 M395X, possibly a M390X in some deviceIDs.
-The R9 M395X was a 'special upgrade' that people could order for their 27" iMac 5K 2015 and so therefore I figured it would have some native support in macOS.
-Apple uses the AMDTongaGraphicsAccelerator package coupled with the AMD9000Controller.kext to initialize the card.
-I was able to get both the internal & external monitors working when I figured out their respective Display_Object_Path i2cid & senseids.
-base vram is 1250Mhz, but I noticed that all modules that come with this card are capable of much higher frequencies with different timings:

tRCDW-tRCDWA-tRCDR-tRCDRA-tRC-tCL-tRFC
1375 MHz: 12-12-19-19-58-19-123 (Samsung K4G41325FC, 4096MB, GDDR5)
1375 MHz: 15-15-20-20-55-18-89 (Elipda EDW4032BABG, 4096MB, GDDR5)
1375 MHz: 14-14-19-19-57-17-136 (Hynix)H5GC4H24AJR, 4096MB, GDDR5)

-The performance is actually great!

View attachment 1948587

I will publish the vbios in due time after more experimentation.

insanely great!
-Steve Jobs

That is indeed great news!
Most interesting to me would be to know whether you found the way to make the internal screen power up (including backlight) at boot without needing an external display or a hardware modification.

 

[mitesh33]

Following on from my previous post a couple of days ago.

Ignore this, I got it working, I selected AMD GPU under OCLP settings then it worked. So not sure why it did this.

Thanks in Advanced

 

[Ausdauersportler]

Following on from my previous post a couple of days ago.

I've got a 2011 iMac 27 inch 12,2. I used the green Dell WX4130 card which I previously had running on my 2009 iMac 27inc 11,1 (custom flashed Rom from here) and working flawlessly for over year or so with not a single problem what so ever.

I've put it my iMac 2011 27 and I can see the first OS user login screen (which is pretty much instant as soon as iMac is on), however the screen then goes black. Any ideas? I've done SMC/PRAM reset too.

I have read the first page but it doesn't mention any issues using the WX4130 on the iMac 2011 27inch 12,2.

I'm using OCLP v0.4.1 using default settings.

Thanks in Advanced

High Sierra?

Known issues:

1. Initial setup: Using High Sierra the GOP vBIOS disables the internal LCD on power on.
   Work around: Use an additional external display on boot.
   If you plan to use High Sierra on a regular basis you need to implement this hard wire mod.

 

[m0bil]

I guess that issue might be solved by using Apple's Windows graphics drivers - but those drivers need the subsystem vendor and device IDs to match exactly.

Would it be possible to modify Apple's Windows graphics drivers .inf to change subsystem IDs, or is something hardcoded inside the binaries ??

 

[mitesh33]

High Sierra?

Known issues:

1. Initial setup: Using High Sierra the GOP vBIOS disables the internal LCD on power on.
   Work around: Use an additional external display on boot.
   If you plan to use High Sierra on a regular basis you need to implement this hard wire mod.

I'm using OSX Monterey and it now works. I will update over the next few days how its running.
Thanks again

 

[vinaypundith]

I guess that issue might be solved by using Apple's Windows graphics drivers - but those drivers need the subsystem vendor and device IDs to match exactly.
I could modify the existing VBIOS accordingly - but at least for the Ellesmere-GPUs (WX7100, RX480) the correct subsystem IDs make the macOS drivers crash.

That sounds worth a try - to tweak the driver to use a subsystem ID that Apple's drivers will work with. the WX4130 is of a very slightly different architecture - Baffin i/o Ellesmere - so maybe that gives it a chance of working

 

[vinaypundith]

What value should I changed the sub vendor ID to? would it be the value seen in a 2017 21.5" iMac or something that has a Radeon Pro 560 GPU? if so, what's that value?

Side note - it seems that even the basic Windows inbox driver doesn't work right with the WX4130 / "RX560" with the EG2 VBIOS, going by the display backlight not coming on

 

[rcmn]

Finally it happens, I got my 2010 iMac fully upgraded and running with Catalina now, AWESOME and thanks you for the help in here

Now with
Intel core i7-870 3.93 GHz
24 GB ddr3 ram
240 gb ssd disk
nVidia Quadro k4100m
Broadcom BCM5701
Bluetooth upgrade
AWESOME, now the machine I usefull the next many years again.

 

[mitesh33]

Following on earlier post. The iMac 27inch 2011 12,2 didn't sleep/wake correctly using default OCLP settings, you need to change under Miscellaneous Settings
- Set Hibernation Workaround: True

I will now test of the next few weeks and report any problems with this setup I encounter. Hopefully it will smoothly as the same WX4130 did in my old iMac 27inch 2009 11,1.

Cheers

 

[internetzel]

What value should I changed the sub vendor ID to? would it be the value seen in a 2017 21.5" iMac or something that has a Radeon Pro 560 GPU? if so, what's that value?

Side note - it seems that even the basic Windows inbox driver doesn't work right with the WX4130 / "RX560" with the EG2 VBIOS, going by the display backlight not coming on

View attachment 1948704

The subsystem IDs in that place don't matter at all, unfortunately. It's more complicated and @nikey22 posted how and where to adjust it: #15,486 .
We'd have to find out the subsystem device and vendor ID of an Apple RX[4,5]60(X) with 2 GB of VRAM.

 

[vinaypundith]

The subsystem IDs in that place don't matter at all, unfortunately. It's more complicated and @nikey22 posted how and where to adjust it: #15,486 .
We'd have to find out the subsystem device and vendor ID of an Apple RX[4,5]60(X) with 2 GB of VRAM.

Eeeeee.... Okay. That's gonna take me a bit to figure out. I'll give it a go tonight.
I have an iMac 2017 with a Radeon Pro 555 2GB - no use, right? for finding subsystem ID

Side note, Any chance of being able to use an MXM-II GeForce 9600M GT in an iMac8,1 with the MXM-II slot? Or stock HP VBIOS not gonna be compatible?

 

[Surf2bikes]

@AirpodsNow and @Surf2bikes : It seems that you both experience the same (random) sleep-related kernel panic on a WX4130 card with Elpida VRAM ICs. (I guess that you both have the same type of VRAM ICs, but only the last half of the part number is visible in @AirpodsNow 's photos.) In a PM @Surf2bikes informed me that he can see the OC bootpicker with the original WX4130 VBIOS. (So it should be easy to revert to @internetzel 's VBIOS in case the original one does not fire up the internal/external display.)

Try this (if it is within your comfort level): Flash the WX4130 with the original VBIOS, see if it boots into macOS on internal/external display, and (if it does) re-enable sleep. Test if the sleep issue persists.

There's a somewhat remote chance that this is related to Elpida VRAM ICs. It is a strange coincidence that the only two users (I know of) reported this issue - and you both have those VRAM ICs.

Here is a little interim report of my testing yesterday night :
- I flashed the WX4130 back to the Original Vbios
- I successfully booted into Monterey with external screen (Samsung TV)
- I enabled/disabled sleep/wake feature 3-4 times
- I did run several 4K videos, Games, Valley benchmark, etc. with NO issues . However I could test only for around 2 hours
.....................................................................................................
- The I flashed back to wx4130_GOP.ROM and did leave sleep/wake activated
- the iMac worked well for 2-3 hours with 3-4 wake/sleep events and the I had 2 sudden crashes within 30 mn - no heavy GPU load, just regular office work. I have attached a crash report from console (unfortunately I cannot read it) and not sure if this is helpful.

This may be not representative as too short testing. I might extend this testing with original VBios.

EDIT: Original BIOS added

 

[internetzel]

Here is a little interim report of my testing yesterday night :
- I flashed the WX4130 back to the Original Vbios
- I successfully booted into Monterey with external screen (Samsung TV)
- I enabled/disabled sleep/wake feature 3-4 times
- I did run several 4K videos, Games, Valley benchmark, etc. with NO issues . However I could test only for around 2 hours
.....................................................................................................
- The I flashed back to wx4130_GOP.ROM and did leave sleep/wake activated
- the iMac worked well for 2-3 hours with 3-4 wake/sleep events and the I had 2 sudden crashes within 30 mn - no heavy GPU load, just regular office work. I have attached a crash report from console (unfortunately I cannot read it) and not sure if this is helpful.

This may be not representative as too short testing. I might extend this testing with original VBios.

Please upload the original VBIOS - or post a link to the post where you had uploaded it. I'll check for differences.