Sidebar Sidebar
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

MP 1,1-5,1 Manually Configured OpenCore on the Mac Pro

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.
prefuse07 said:
Yeah that's right you set it using the DAW.

I am having some weird stuff happen though.... I can set to 96kHz with a buffer rate of 512, and with Ableton open, there is no latency at all -- clear as day.

Then as soon as I quit Ableton, the latency comes back..... I've tried playing around adding some negativity to bring the latency down to absolute 0, and even to negative values such as -45ms, but I still get issues as soon as I quit Ableton. Really odd!


Edit:

There has got to be a way to fix these latency issues.... This is NOT acceptable

In my DAW I can set the sample rate to 192kHz @ 1024 bit size and I have a 12.8ms overall latency, which is within the 13ms allowed range, but as soon as I exit the DAW, the latency issues persist.

I've tried switchin USB ports with no luck -- wonder if it's my USB cable? hmmm I didn't have any issues like this with High Sierra though -- and I think now this was one of the reasons I decided to stay on that OS for as long as I did.
Click to expand...
Let me ask a basic question here, your Xeons are Nehalem (X55xx/L55xx/E55xx) yet?

Nehalem dual processors are not supported after 10.14.3 - in reality support officially ended with 10.11.6, but the problems started with 10.14.4. This is not an OpenCore issue.

The first collateral damage of the Nehalem support being removed by Apple with Mojave and later releases are audio stuttering/cracking. This is not news and if you want to have audio working correctly after 10.14.3, you need to upgrade to Westmere Xeons.
 
  • Like
  • Love
Reactions: paalb and prefuse07
There has got to be a way
tsialex said:
Let me ask a basic question here, your Xeons are Nehalem (X55xx/L55xx/E55xx) yet?

Nehalem dual processors are not supported after 10.14.3 - in reality support officially ended with 10.11.6, but the problems started with 10.14.4. This is not an OpenCore issue.

The first collateral damage of the Nehalem support being removed by Apple with Mojave and later releases are audio stuttering/cracking. This is not news and if you want to have audio working correctly after 10.14.3, you need to upgrade to Westmere Xeons.
Click to expand...


Ah ha! So it's hardware related after all..... very well, I guess I have no choice but to do the 12 core mod! (was holding off, but now that you have enlightened me, pulling the trigger).

Thank you!!!!
 
Looking for some help ....

I have been successfully using OpenCore to boot Big Sur and to update Catalina using a NVIDIA GTX680 (flashed). As Monterey is dropping the NVIDIA Kepler drivers I picked up a used MSI RX 560 Aero 4 GB and I am trying to get it working so I can upgrade to Monterey.

So far I can boot into Mojave natively and it works fine (though I do miss the native boot picker) and is recognized as a Radeon RX 560 (0x67ff Device ID). I can also boot fine into my Legacy Windows 10 install.

I am having trouble with booting Big Sur using OpenCore.

I generally use OpenCore on a USB stick but in my testing am using a rescue CD so I can boot into the OpenCore (OpenCanopy) boot picker just holding the "c" key.

I see the boot picker fine, select Big Sur, see the Apple logo. The progress bar starts fine but when it gets near half way across small purple horizontal lines appear randomly across the top 1/4 or so of the display and after 10 seconds or so the screen goes black and I have to shut down. I never get to the desktop.

The only thing that I can see in my config.plist that might be causing it is that I have DirectGOPRendering set to TRUE.

I also am not doing anything with acceleration at this point - i just want to get it working first.

I would appreciate any pointers/advice on getting it working as it is fine when booting natively into Mojave. I have attached a copy of my config.plist that is working fine booting Big Sur with my NVIDIA GTX680

Regards,
@sfalatko
 

Attachments

  • config.plist.zip
    3.1 KB · Views: 73
sfalatko said:
I see the boot picker fine, select Big Sur, see the Apple logo. The progress bar starts fine but when it gets near half way across small purple horizontal lines appear randomly across the top 1/4 or so of the display and after 10 seconds or so the screen goes black and I have to shut down. I never get to the desktop.
Click to expand...
I also am not doing anything with acceleration at this point - i just want to get it working first.
Click to expand...
Without WEG, the purple lines are a common thing.

The only thing that I can see in my config.plist that might be causing it is that I have DirectGOPRendering set to TRUE.
Click to expand...
This should only affect the boot menu.

I have attached a copy of my config.plist that is working fine booting Big Sur with my NVIDIA GTX680
Click to expand...
Because you're spoofing the iMacPro board ID, you'll need WEG and the agdpmod=pikera device property or boot-arg to enable all video outputs. Provisionally, you can also try other outputs on your card. Another option is to use the MacPro7,1 board ID (as recommended in the latest version of the guide).
 
sfalatko said:
Looking for some help ....

I have been successfully using OpenCore to boot Big Sur and to update Catalina using a NVIDIA GTX680 (flashed). As Monterey is dropping the NVIDIA Kepler drivers I picked up a used MSI RX 560 Aero 4 GB and I am trying to get it working so I can upgrade to Monterey.

So far I can boot into Mojave natively and it works fine (though I do miss the native boot picker) and is recognized as a Radeon RX 560 (0x67ff Device ID). I can also boot fine into my Legacy Windows 10 install.

I am having trouble with booting Big Sur using OpenCore.

I generally use OpenCore on a USB stick but in my testing am using a rescue CD so I can boot into the OpenCore (OpenCanopy) boot picker just holding the "c" key.

I see the boot picker fine, select Big Sur, see the Apple logo. The progress bar starts fine but when it gets near half way across small purple horizontal lines appear randomly across the top 1/4 or so of the display and after 10 seconds or so the screen goes black and I have to shut down. I never get to the desktop.

The only thing that I can see in my config.plist that might be causing it is that I have DirectGOPRendering set to TRUE.

I also am not doing anything with acceleration at this point - i just want to get it working first.

I would appreciate any pointers/advice on getting it working as it is fine when booting natively into Mojave. I have attached a copy of my config.plist that is working fine booting Big Sur with my NVIDIA GTX680

Regards,
@sfalatko
Click to expand...

You have not completely followed the guide in #post 1 - especially from the section titled Complete your setup.

Your system has probably booted fine but do not have screen output and most likely need a specific boot argument agdpmod=pikera but that cannot work until you install Lilu.kext & Whatevergreen.kext detailed in that section. Had you previously setup Screen Sharing (in System Preferences) you can most likely login remotely from another Mac on your network.

As you do not have an AMD 5700XT or higher, you can safely set DirectGOPRendering=false.

I suggest you read the guide completely.
Alternatively, start with the pre-built package by Martin Lo here
 
cdf said:
Without WEG, the purple lines are a common thing.


This should only affect the boot menu.


Because you're spoofing the iMacPro board ID, you'll need WEG and the agdpmod=pikera device property or boot-arg to enable all video outputs. Provisionally, you can also try other outputs on your card. Another option is to use the MacPro7,1 board ID (as recommended in the latest version of the guide).
Click to expand...
@cdf - thanks.

I will search around and dig into WEG and agdpmod device property/boot-arg. Would I need that if I move to spoofing a MacPro7,1? I've started looking into graphics hardware acceleration - what's the general feeling on it if I'm not doing any heavy graphics/video work?

regards,
@sfalatko
 
MacNB2 said:
You have not completely followed the guide in #post 1 - especially from the section titled Complete your setup.

Your system has probably booted fine but do not have screen output and most likely need a specific boot argument agdpmod=pikera but that cannot work until you install Lilu.kext & Whatevergreen.kext detailed in that section. Had you previously setup Screen Sharing (in System Preferences) you can most likely login remotely from another Mac on your network.

As you do not have an AMD 5700XT or higher, you can safely set DirectGOPRendering=false.

I suggest you read the guide completely.
Alternatively, start with the pre-built package by Martin Lo here
Click to expand...
Thanks @MacNB2, I have read the guide previously but with the perspective of NVIDIA.

I'll head back to the guide and delve into the "Complete Your Setup" section.

Regards,
@sfalatko
 
sfalatko said:
I will search around and dig into WEG and agdpmod device property/boot-arg. Would I need that if I move to spoofing a MacPro7,1? I've started looking into graphics hardware acceleration - what's the general feeling on it if I'm not doing any heavy graphics/video work?
Click to expand...
I should mention that WEG stands for WhateverGreen. You won't need the device property or boot-arg if spoofing the MacPro7,1 board ID. To get the most out of your hardware, you should really enable hardware acceleration. It will clean up those purple lines and give you DRM.
 
cdf said:
I should mention that WEG stands for WhateverGreen. You won't need the device property or boot-arg if spoofing the MacPro7,1 board ID. To get the most out of your hardware, you should really enable hardware acceleration. It will clean up those purple lines and give you DRM.
Click to expand...
@cdf - are you saying that if spoofing a MacPro7,1 I wouldn't need step 6 - device property with the path to my graphics card? If that is so then the guide isn't quite clear that it isn't necessary when spoofing a MacPro7,1.
 
sfalatko said:
@cdf - are you saying that if spoofing a MacPro7,1 I wouldn't need step 6 - device property with the path to my graphics card? If that is so then the guide isn't quite clear that it isn't necessary when spoofing a MacPro7,1.
Click to expand...
No. That step (which actually no longer includes the agdpmod property because of the change to MacPro7,1 spoofing) is still needed for the other properties (rebuild-device-tree and unfairgva).
 
cdf said:
No. That step (which actually no longer includes the agdpmod property because of the change to MacPro7,1 spoofing) is still needed for the other properties (rebuild-device-tree and unfairgva).
Click to expand...
@cdf - thanks for the explanation and help.

Followed the guide and created a new RescueCD. Now booted into Big Sur via OpenCore 0.7.5 with my RX 560.

Non of the artifacts now that I have Lilu and WEG installed.

The only thing I am not sure about is Hardware Acceleration. I installed VideoProc and it is not detecting my Video Card and only showing CPU acceleration.

Is there anything more I need to do for Hardware Acceleration other than what is in post #1?

@sfalatko
 
Can anyone help me solve / fix one thing (that I believe is an OC thing)?

When I boot into my system after choosing Mac in boot picker, then navigate to Sys Prefs > Displays

It shows the following

Even if I choose "Scaled" The only available options are 1080i or 720p

Screen Shot 2021-11-08 at 6.57.33 PM.png


(sometimes it gets scaled into 1080i when I log in)


If I power off / power back on my monitor, I then see this in Sys Prefs > Displays

Screen Shot 2021-11-08 at 7.00.03 PM.png


Any idea what in my config.plist is causing this / or how to fix it?

config.plist is attached.
 

Attachments

  • config.plist.zip
    5.2 KB · Views: 72
sfalatko said:
@cdf - So how do you confirm that Hardware Acceleration is working? In Videoproc Vlogger it does not detect a GPU (says N/A).
Click to expand...

It's a "cosmetic" Videoproc bug.
It's because Videoproc does not detect the "AMD" in the model name of the GPU and assumes it's an unknown GPU but it does use GPU's HEVC & H.264 encode/decode features.

If you wish VideoProc to display the GPU name, then you can inject the model name into OpenCore's DeviceProperties settings. E.g. For my RX 580 I use "AMD Radeon RX 580" string like this:

Screenshot 2021-11-09 at 21.51.01.png

My RX580 is in Slot 1.

Videoproc now displays:

Screenshot 2021-11-09 at 21.45.37.png
You can use the same trick with other devices. E.g. I have a Sabrent Rocket NMVe blade in Slot-3 and I name it "Sabrent NVMe M.2" :

Screenshot 2021-11-09 at 21.59.51.png

So now, in System Profiler under PCI it displays:

Screenshot 2021-11-09 at 21.47.18.png

All cosmetic...nice to have.
 
Last edited:
  • Like
Reactions: splifingate, JedNZ and hwojtek
MacNB2 said:
It's a "cosmetic" Videoproc bug.
It's because Videoproc does not detect the "AMD" in the model name of the GPU and assumes it's an unknown GPU but it does use GPU's HEVC & H.264 encode/decode features.

If you wish to VideoProc to display the GPU name, then you can inject the model name into OpenCore's DeviceProperties settings. E.g. For my RX 580 I use "AMD Radeon RX 580" string like this:

View attachment 1906880

My RX580 is in Slot 1.

Videoproc now displays:

View attachment 1906908
You can use the same trick with other devices. E.g. I have a Sabrent Rocket NMVe blade in Slot-3 and I name it "Sabrent NVMe M.2" :

View attachment 1906886

So now, in System Profiler under PCI it displays:

View attachment 1906887

All cosmetic...nice to have.
Click to expand...
@MacNB2 - Thanks. Nice add.
 
I tried doing a clean install of os 12 today and I received that Unable to update the firmware error. I am using matins latest 75. Do I need additional config to do a clean install of OS12? Im looking around but I'm kinda lost now. Can I get a point in the right direction. Thank you!
 
r6novak said:
I tried doing a clean install of os 12 today and I received that Unable to update the firmware error. I am using matins latest 75. Do I need additional config to do a clean install of OS12? Im looking around but I'm kinda lost now. Can I get a point in the right direction. Thank you!
Click to expand...
 
  • Like
Reactions: r6novak
Hi guys, I have issues with sleep mode on my 5,1 in Monterey with OC 075. any suggestions? My mac won‘t sleep anymore...
 
hwojtek said:
Great tip! Can we have it in JSON please? (example PciRoot(0x0)/Pci(0x3,0x0)/Pci(0x0,0x0) entry)
Click to expand...

Thanks.
JSON ?
I think you meant raw XML plist format ?

If so, you add following key to the Deviceproperties>>Add section:

XML: 
            <key>PciRoot(0x0)/Pci(0x3,0x0)/Pci(0x0,0x0)</key>
            <dict>
                <key>model</key>
                <string>AMD Radeon RX 580</string>
            </dict>

The actual path of the device depends on which slot you install your GPU.
For Slot-1 it is PciRoot(0x0)/Pci(0x3,0x0)/Pci(0x0,0x0)

For other Slots, use gfxutil as described by @cdf in #Post 1 to find the path.

I do not edit raw XML plist files but use a plist editor (PListEdit Pro) which takes care of formatting, syntax checking & correctness so that I do not have to worry about whether I put /dict or key/ or /key, etc in the correct place or in correct oreder/format.

EDIT:
JSON is not my expertise but may be like this in Python:

JSON: 
        'DeviceProperties': {
            'Add': {
                'PciRoot(0x0)/Pci(0x3,0x0)/Pci(0x0,0x0)': {
                    'model': 'AMD Radeon RX 580'
                }
            }
        }

NOTE indentation and formatting is very important relative to other entries and the above example is just the substance and not necessarily correct format.
 
Last edited:
  • Like
Reactions: hwojtek
MacNB2 said:
Thanks.
JSON ?
I think you meant raw XML plist format ?
Click to expand...
Oh, sorry. I've spent all my day with JSON configs on a linux machine and I wrote it just automatically. Yes, I meant XML, obviously. Just the raw snippet of XML from config.plist as seen in BBEdit or any other simple text editor. Sorry for putting you through the hassle of creating a JSON interpretation of it.
The point is - I am not exactly sure of the syntax, I did try adding a <name>here goes the name</name> between <built-in>whatever</built-in> and <pci-aspm...>whatever</pci-aspm...> for my NVMe controller and I ended up with disk seen as removable, instead of seeing "here goes the name" in System Profiler so a raw snippet would help. I use BBEdit to fine-tune my configs and not PlistEdit.
 
hwojtek said:
Oh, sorry. I've spent all my day with JSON configs on a linux machine and I wrote it just automatically. Yes, I meant XML, obviously. Just the raw snippet of XML from config.plist as seen in BBEdit or any other simple text editor. Sorry for putting you through the hassle of creating a JSON interpretation of it.
Click to expand...
No problem.

hwojtek said:
The point is - I am not exactly sure of the syntax, I did try adding a <name>here goes the name</name> between <built-in>whatever</built-in> and <pci-aspm...>whatever</pci-aspm...> for my NVMe controller and I ended up with disk seen as removable, instead of seeing "here goes the name" in System Profiler so a raw snippet would help. I use BBEdit to fine-tune my configs and not PlistEdit.
Click to expand...

Because of syntax issues, I do not use BBEdit or Textedit. A plist editor focuses you on the "substance" of what you want to create rather that the syntax (which it takes care of). Yes, you still need to know "what" you wish to edit. Honestly, try a demo copy and you'll wonder why you never had it years ago.
Anyway, I guess I am going off topic and not related to Opencore.

Regarding your NVMe controller, the DeviceProperties would look something like this:
XML: 
<key>PciRoot(0x0)/Pci(0x1,0x0)/Pci(0x0,0x0)/Pci(0x2,0x0)/Pci(0x0,0x0)</key>
<dict>
    <key>built-in</key>
    <data>00</data>
    <key>name</key>
    <string>My NVMe specific name</string>
    <key>pci-aspm-default</key>
    <integer>2</integer>
</dict>

Where PciRoot(0x0)/Pci(0x1,0x0)/Pci(0x0,0x0)/Pci(0x2,0x0)/Pci(0x0,0x0) is the path for Slot-3
If your NMVe is in a different slot, then you need to find your path and change the above.
 
Last edited:
  • Like
Reactions: hwojtek
You must log in or register to reply here.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Top Bottom
Back