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.
cdf said:
Let's address this one first.

Unfortunately, you seem to be dealing with one of the exceptional cases where your GPU lacks the necessary acpi-path property for adding other properties. The quickest solution for now is just to use boot-args. In your config, remove

XML: 
<key>PciRoot(0x0)/Pci(0x3,0x0)/Pci(0x0,0x0)/Pci(0x0,0x0)/Pci(0x0,0x0)</key>
<dict>
<key>rebuild-device-tree</key>
<integer>0</integer>
<key>unfairgva</key>
<integer>1</integer>
<key>shikigva</key>
<integer>80</integer>
</dict>

and replace

XML: 
<string>-no_compat_check</string>

by

XML: 
<string>-no_compat_check wegtree=1 unfairgva=1 shikigva=80</string>

Now with these changes, check that DRM works.
Click to expand...

Yup, confirming -- moving it all back into boot-args seems to do the trick.

So a note to anyone reading this:

Sapphire Pulse RX-5700XT [11293-01-20G] -- lacks the appropriate acpi-path, so if you want DRM, you need to add the above as boot-args instead of within DeviceProperties.

Thanks again @cdf for all your help!
 
Hello, Just noticed restarting the mac with 0.7.6 installed does not reboot, load screen hangs around one third bar.
Have to shut down and start up. Has anyone else noticed this?
MacPro 5,1 OS Monterey.
 
prefuse07 said:
Sapphire Pulse RX-5700XT [11293-01-20G] -- lacks the appropriate acpi-path
Click to expand...
What is your path?

To further configure WhateverGreen for hardware acceleration, you'll need a special utility:

Utility ⠀ ⠀ ⠀ ⠀ ⠀ ⠀ ⠀ ⠀ ⠀⠀​
Description ⠀ ⠀ ⠀ ⠀ ⠀ ⠀ ⠀ ⠀ ⠀ ⠀​
What to do ⠀ ⠀ ⠀ ⠀ ⠀ ⠀ ⠀ ⠀ ⠀ ⠀ ⠀ ⠀ ⠀ ⠀ ⠀ ⠀ ⠀ ⠀ ⠀ ⠀ ⠀ ⠀​
gfxutilFor getting device pathsDownload the latest RELEASE version

➌ Copy the executable "gfxutil" to your home folder.
➍ In Terminal, enter the following command:

xattr -c gfxutil

➎ Get the path of your graphics card by entering the following command:

./gfxutil -f display | sed 's|.*= ||'
 
cdf said:
You have a space before the *. Remove it to get the properly formatted path, which seems to be
PciRoot(0x0)/Pci(0x3,0x0)/Pci(0x0,0x0)/Pci(0x0,0x0)/Pci(0x0,0x0)
Click to expand...
prefuse07 said:
./gfxutil -f display | sed 's|. *= ||' returns the following path:
Code: 
% /PCI0@0/IOU0@3/PXS1@0/pci-bridge@0/display@PciRoot(0x0)/Pci(0x3,0x0)/Pci(0x0,0x0)/Pci(0x0,0x0)/Pci(0x0,0x0)
Therefore, the address: Pci(0x3,0x0)/Pci(0x0,0x0), is the path of my GPU, right? (PCIe slot 1)
Click to expand...
This will not work unless you create an SSDT with the correct path. The pci-bridge@0 has to be eliminated by reattaching the GPU directly in the SSDT.
 
I had Catalina running via OpenCore on my Mac Pro 5,1 / Radeon RX 580 (thanks to this forum) and decided to upgrade to Monterey. Open Core Legacy Patcher (0.4.1) made the transition seamless - wow! Before the upgrade, I used Geekbench 5 to measure CPU and GPGPU performance, just out of curiosity. The CPU results weren't changed much (though they did drop a few %). However, the "Compute Benchmark" results dropped significantly. I re-ran the OCLP post-install, and got the same results.

Catalina Geekbench 5 Compute Benchmark:
- OpenCL: 43418

Monterey (OCLP 0.4.1) Compute Benchmark:
- OpenCL: 25636
- Metal: 26230

This is not affecting my installation or day to day enjoyment of the system at all - GPGPU is not part of my daily diet. I thought some folks would be curious about this result, and someone might have an idea of what's going on.
 
Last edited:
  • Like
Reactions: JeDiGM
N4LYD said:
Catalina Geekbench 5 Compute Benchmark:
- OpenCL: 43418

Monterey (OCLP 0.4.1) Compute Benchmark:
- OpenCL: 25636
- Metal: 26230
Click to expand...
Benchmark again, put the computer to sleep, wait 60 seconds, wake and benchmark again. This used to be a remedy sometime ago, I believe this is no longer the case but perhaps?
BTW my RX570 8GB just clocked 38700 OpenCL, 43927 Metal (down from 41719/46880 on 11.6.1)
 
Last edited:
hwojtek said:
Benchmark again, put the computer to sleep, wait 60 seconds, wake and benchmark again. This used to be a remedy sometime ago, I believe this is no longer the case but perhaps?
BTW my RX570 8GB just clocked 38700 OpenCL, 43927 Metal (down from 41719/46880 on 11.6.1)
Click to expand...
I read about the "sleep/retry" thing, and thought I'd done it, but your "wait 60 seconds" suggestion was new, and sure enough, after > 60 seconds of sleep, waking the machine up, the Geekbench Compute numbers dramatically improved:

OpenCL: 45252
Metal: 48651

Then after 20 minutes of uptime, re-ran, and saw:
Metal: 27382

Finally, did a sleep, and watched the power on my USB devices get cut after about 20 seconds, so rebooted then, and sure enough, the score returned to 46874.

Checking the post you mentioned, this issue seems to have either gone away for most folks after Big Sur, or they aren't checking it. The system runs great otherwise, so not sure if others are having it happen but not noticing.

[edit: the slower performance seems to happen at around 5 minutes of uptime. Very odd.]
 
Last edited:
  • Like
Reactions: JeDiGM and hwojtek
NC12 said:
@cdf I noticed in the OpenCore guide there is an option to enable the boot chime on hackintoshes. Could this same guide be used to change the boot sound to the modern boot chime on the mac pro?
Click to expand...
While you can't really change the original boot chime, you could always enable an additional (modern) boot chime that would play when OpenCore loads. This could be useful for those Mac Pros that lost their boot chime with MP51.0087.B00. See posts #9,995 and #10,006 for hints on enabling audio preboot support.
 
  • Like
Reactions: NC12
N4LYD said:
or they aren't checking it.
Click to expand...
A quick way to check this is to use the benchmark feature in AirOut 2:

forums.macrumors.com

AirOut: A solution to the racing fan bug

Introduction The racing fan bug is a well known issue that affects certain cMP systems with upgraded video cards, particularly systems with SMC version 1.39f5. On these systems the PCI, PS, and booster fans spin abnormally fast at idle, causing unnecessary noise. Although the issue appears to...
forums.macrumors.com forums.macrumors.com
 
cdf said:
While you can't really change the original boot chime, you could always enable an additional (modern) boot chime that would play when OpenCore loads. This could be useful for those Mac Pros that lost their boot chime with MP51.0087.B00. See posts #9,995 and #10,006 for hints on enabling audio preboot support.
Click to expand...
Oh so you could disable the original boot sound via terminal and then have the new boot sound play when opencore loads?
 
NC12 said:
Oh so you could disable the original boot sound via terminal and then have the new boot sound play when opencore loads?
Click to expand...
I don't believe this has been tested, but in principle, the OC boot chime would also be disabled, because it also respects the SystemAudioVolume NVRAM variable...
 
cdf said:
I don't believe this has been tested, but in principle, the OC boot chime would also be disabled, because it also respects the SystemAudioVolume NVRAM variable...
Click to expand...
No. Or not quite! ;) If the in-OS sound is muted then yes (as with the real Mac boot chime) the OC boot chime will be muted. However, there is an NVRAM variable StartupMute which mutes the Mac boot chime on real Macs - or at least it does on later firmware.... OC Audio/PlayChime can be configured to be always on (Enabled), always off (Disabled), or respect to StartupMute (Auto). So the trick for this to work is to find out if the MacPro firmware respects StartupMute. If it does, you are good.

The variable is a single byte value in the Apple config NVRAM section 7C436110-AB2A-4BBB-A880-FE41995C9F82 and the value I have seen recommended to set is the ASCII code for '1' (i.e. 0x31, dec. 49), which definitely works (apparently any non-zero byte should also work). So if I've done my sums right for the NVRAM settings, it should be:

Code: 
    <key>NVRAM</key>
    <dict>
        ...
        <key>Add</key>
        <dict>
            ...
            <key>7C436110-AB2A-4BBB-A880-FE41995C9F82</key>
            <dict>
                ...
                <key>StartupMute</key>
                <data>MQ==</data>
                ...
            </dict>
        </dict>
        ...
    </dict>
 
  • Like
Reactions: Macschrauber and cdf
Bmju said:
So the trick for this to work is to find out if the MacPro firmware respects StartupMute.
Click to expand...
In Monterey, this variable can also be easily set in System Preferences > Sound > Sound Effects by checking or unchecking "Play sound on startup". Unfortunately, the only way I've ever been able to disable the chime is by muting the system sound...
 
  • Like
Reactions: Bmju
cdf said:
In Monterey, this variable can also be easily set in System Preferences > Sound > Sound Effects by checking or unchecking "Play sound on startup". Unfortunately, the only way I've ever been able to disable the chime is by muting the system sound...
Click to expand...
The variable gets set, but has no effect? That's too bad. PlayChime would need a fourth setting, then. Forced or something. But it's a bit of a niche market - especially as the OC chime comes quite a lot later in boot; well after the mac itself will have responded to some possible hotkeys, though before OC responds to any - so OC's chime is a less useful indicator on a mac than a hack.
 
hwojtek said:
Memory leak in AMD kext?
Click to expand...
It's an "all-or-nothing" behavior, not a slow degradation, so I don't think it's a driver memory leak. Plus my system isn't showing an unbounded growth of pinned memory - even after multiple days of uptime.
 
Bmju said:
The variable gets set, but has no effect? That's too bad. PlayChime would need a fourth setting, then. Forced or something. But it's a bit of a niche market - especially as the OC chime comes quite a lot later in boot; well after the mac itself will have responded to some possible hotkeys, though before OC responds to any - so OC's chime is a less useful indicator on a mac than a hack.
Click to expand...
Actually @cdf, I caught up with what you were saying already: if there's no known way to mute the native chime except by muting the in-OS sound, then it pretty much doesn't work. I guess you could have a shutdown script to mute the sound, and an OpenCore NVRAM variable to unmute it, or something, but it's getting pretty baroque for a feature of questionable value anyway.
 
  • Like
Reactions: cdf
cdf said:
This guide explains how to use the excellent OpenCore boot loader to install, run, and update the latest version of macOS on the MacPro5,1, resulting in a clean, unaltered operating system just like on a supported Mac.

View attachment 1899553
Click to expand...
Internal PCIe drives

I would like to ask you to give commands to determine the path of pcie cards with nvme ssd gfxutil, as they were made for old versions of opencore (for example 0.6.9). Since the terminal command in the current version of the opencore installation instructions does not give me all the paths of my nvme ssd. In the old version, I entered the psi address from my SystemInfo into terminal command line. I have record in my note with my all pci paths, but now I want to move nvme ssd to other slots, and gfxutil with current command line will not show me the path to one of my NVME SSD. Thanks in advance. MacPro 4.1/5.1, dual 5680, 96Gb samsung RAM, 1Tb nvme 970EvoPlus, 2Tb crutial P2, 1Tb Crutial NAND SATA M2, USB3 pcie card, OC 0.7.7, MacOS Monterey (Mojave, BigSur)
 
Last edited:
PeterHolbrook said:
I've just updated my cMP to Monterey 12.2 (OpenCore 0.7.7). No issues. SurPlus wasn't necessary.
Click to expand...
I also updated Monterey this morning, but I have a config with sur plus, is it really no longer needed? Or is the community still accumulating experience and statistics?
 
Gustav Holdoff said:
I would like to ask you to give commands to determine the path of pcie cards with nvme ssd gfxutil, as they were made for old versions of opencore (for example 0.6.9). Since the terminal command in the current version of the opencore installation instructions does not give me all the paths of my nvme ssd. In the old version, I entered the psi address from my SystemInfo into terminal command line. I have record in my note with my all pci paths, but now I want to move nvme ssd to other slots, and gfxutil with current command line will not show me the path to one of my NVME SSD.
Click to expand...
Sure. First carry out steps 1 and 2 of Internal PCIe drives. Then
  • Open System Information and find the name of your disk in the PCI section
    Look for "AHCI Controller" or "NVM Express Controller." You should see something like pci0000,0000.
  • In terminal, enter ./gfxutil -f pci0000,0000. Make sure to replace pci0000,0000 with the actual name of your disk.
This should return the desired path.

I'd like to understand why the command isn't working for you, though. Please enter system_profiler SPPCIDataType in Terminal, and post the output. Is your disk listed there? If so, which one is it? Thanks!
 
  • Like
Reactions: Gustav Holdoff
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