MP 1,1-5,1 Activate AMD hardware acceleration

[cmphal]

I don’t think there’s a consensus. My current hypothesis is that this is due to a) having concurrent APFS volumes or b) FeatureUnlock.

@cdf, as data for your hypothesis I only have a single APFS volume on the SSD Volume in question.

Will probably try a clean install of 12.6.3, which is my second question, 12.6.1 is when I joined this thread, I hadn't realized that I had missed the 12.6.2 update. Nor do I see a 12.6.3 update. Should I automatically be seeing those updates?

Currently using MartinLo's 8.6.0, firmware 9144.0.8.6.0 and I don't seem to be receiving macOS updates. Here's what Software Update is reporting.

 

[antonlinus]

what could be the issue when my bootpicker is static? like i see it but i don't see the selected drive changing on screen (althou it is changing it). I was working before and just suddently stopped...

thanks for any advice

Solved this by doing a NVRAM reset in case anyone else struggles with this…

 

[antonlinus]

@cdf, as data for your hypothesis I only have a single APFS volume on the SSD Volume in question.
View attachment 2147251
Will probably try a clean install of 12.6.3, which is my second question, 12.6.1 is when I joined this thread, I hadn't realized that I had missed the 12.6.2 update. Nor do I see a 12.6.3 update. Should I automatically be seeing those updates?

Currently using MartinLo's 8.6.0, firmware 9144.0.8.6.0 and I don't seem to be receiving macOS updates. Here's what Software Update is reporting.
View attachment 2147252View attachment 2147253

You neet to enable the VMM flag and check the UpdateSMBIOS settings in your OC config.plist to update your OS safely. See this video (turn on cc)

Make sure to reverse those edits after updating since they are not meant to be permanent and can affect power management

 

[cdf]

@cdf, as data for your hypothesis I only have a single APFS volume on the SSD Volume in question.
View attachment 2147251

Thanks for the info. Do you have any other disks that use APFS?

Will probably try a clean install of 12.6.3, which is my second question, 12.6.1 is when I joined this thread, I hadn't realized that I had missed the 12.6.2 update. Nor do I see a 12.6.3 update. Should I automatically be seeing those updates?

Yes, but this also depends on your OC configuration (maybe not using SecureBootModel or missing firmware features).

You neet to enable the VMM flag and check the UpdateSMBIOS settings in your OC config.plist to update your OS safely.

Again, this depends on your configuration (maybe by design to accommodate certain setups and older versions of macOS). Updates without the VMM flag have been possible since Big Sur.

 

[cmphal]

Do you have any other disks that use APFS?

Yes, a TimeMachine volume, and a User Volume which are each on separate HDD

 

[captainhulk]

legacy or updated card?

Legacy

 

[cmphal]

Why is this so fragile? Am I in danger of bricking my MacPro?

Okay, I mounted my EFI and copied config.plist to my desktop. Then edited the copy to change <key>Cpuid1Mask</key> from all "A" to <data>AAAAAAAAAAAAAACAAAAAAA==</data>. I also changed <key>UpdateSMBIOS</key> to <false/>. I copied back to the EFI Partition and I can't reboot. I copied the file using finder and now thinking I should have done it with cp in terminal.

Nothing. Tried Option-Command-P-R for three chimes but still no boot screen. If machine left alone it would repeat reboot chime and long wait with nothing on the screen.

I tried putting an old boot drive in and it still wouldn't boot. Though that was resolved by pulling the Sonnet Card with Boot SSD. I believe the issue there is that the OpenCore on the PCIe mounted SSD is always grabbing things first so no way to get control other than pulling the card. I then was able to boot using an old install of macOS 11.6.2 under MartinLo's 7.3.1.

Did I misinterpret the edits I was supposed to make to config.plist? Any thoughts as to what I did wrong? Or suggestions as to how to get back to macOS 12.6.1 or update to 12.6.3. Sort of feeling that I should have left things alone. But also worried about the fragility. I had a lot of problems when I first upgraded from Big Sur to Monterey but chalked that up to the PCIe mounted SSD.

BTW, my opening question should be interpreted as a lament and not a complaint. I continue to be amazed at the longevity of the cMP product line.

 

[h9826790]

Why is this so fragile? Am I in danger of bricking my MacPro?

Okay, I mounted my EFI and copied config.plist to my desktop. Then edited the copy to change <key>Cpuid1Mask</key> from all "A" to <data>AAAAAAAAAAAAAACAAAAAAA==</data>. I also changed <key>UpdateSMBIOS</key> to <false/>. I copied back to the EFI Partition and I can't reboot. I copied the file using finder and now thinking I should have done it with cp in terminal.

Nothing. Tried Option-Command-P-R for three chimes but still no boot screen. If machine left alone it would repeat reboot chime and long wait with nothing on the screen.

I tried putting an old boot drive in and it still wouldn't boot. Though that was resolved by pulling the Sonnet Card with Boot SSD. I believe the issue there is that the OpenCore on the PCIe mounted SSD is always grabbing things first so no way to get control other than pulling the card. I then was able to boot using an old install of macOS 11.6.2 under MartinLo's 7.3.1.

Did I misinterpret the edits I was supposed to make to config.plist? Any thoughts as to what I did wrong? Or suggestions as to how to get back to macOS 12.6.1 or update to 12.6.3. Sort of feeling that I should have left things alone. But also worried about the fragility. I had a lot of problems when I first upgraded from Big Sur to Monterey but chalked that up to the PCIe mounted SSD.

BTW, my opening question should be interpreted as a lament and not a complaint. I continue to be amazed at the longevity of the cMP product line.

Using Finder is fine. The EFI partition is just a very basic FAT32 partition. No permission issue.

Did you check if there is any typo?

Sadly, OpenCore has very little to no resist to typo. And it is that fragile.

To overcome that, the OpenCore team actually provided an ocvalidate tool to check if there is any error in the config plist. That can pick up most syntax error. However, since only minior change required if using my pre-configured OC package. And very very few people fall into this kind of no boot issue after following that video to mod the config. Therefore, I haven't put that tool inside the package.

If you has a 11.6.2 and OC 0.7.3 as backup. You can ...

1) plug in a USB flash drive into the USB 2.0 front port of the cMP

2) Format the USB drive to GUID

3) Mount the EFI partition of the USB drive

4) Install OC 0.8.8 onto the USB drive

5) Run the bless tool (assume ONLY the USB's EFI partition is mounted)

6) Reboot

This should allow you to boot to OC 0.8.8. If this works, then you can re-install your 12.6.1 back in.

Once you can boot to 12.6.1, you can fix the OpenCore installed onto that drive. Then mod the config plist on the USB flash drive's EFI directly (no need to copy that to desktop). Double check no typo, then reboot.

If anything goes wrong again, just pull the USB drive, then your cMP should boot to other known good OC automatically.

Another way to fix the OC on the 12.6.1 drive is to connect that externally (e.g. via USB to SATA / PCIe cable). Then mount the EFI drive and fix the config plist.

 

[antonlinus]

Why is this so fragile? Am I in danger of bricking my MacPro?

Okay, I mounted my EFI and copied config.plist to my desktop. Then edited the copy to change <key>Cpuid1Mask</key> from all "A" to <data>AAAAAAAAAAAAAACAAAAAAA==</data>. I also changed <key>UpdateSMBIOS</key> to <false/>. I copied back to the EFI Partition and I can't reboot. I copied the file using finder and now thinking I should have done it with cp in terminal.

Nothing. Tried Option-Command-P-R for three chimes but still no boot screen. If machine left alone it would repeat reboot chime and long wait with nothing on the screen.

I tried putting an old boot drive in and it still wouldn't boot. Though that was resolved by pulling the Sonnet Card with Boot SSD. I believe the issue there is that the OpenCore on the PCIe mounted SSD is always grabbing things first so no way to get control other than pulling the card. I then was able to boot using an old install of macOS 11.6.2 under MartinLo's 7.3.1.

Did I misinterpret the edits I was supposed to make to config.plist? Any thoughts as to what I did wrong? Or suggestions as to how to get back to macOS 12.6.1 or update to 12.6.3. Sort of feeling that I should have left things alone. But also worried about the fragility. I had a lot of problems when I first upgraded from Big Sur to Monterey but chalked that up to the PCIe mounted SSD.

BTW, my opening question should be interpreted as a lament and not a complaint. I continue to be amazed at the longevity of the cMP product line.

Do you have another option to access the OC EFI partition, like throu target mode or plugging the drive into a working machine?
My guess would be you mistyped something and therefore corrupted the Config.plist.
Like for example writing </false> instead of <false/>

 

[antonlinus]

Stupid question: Should i get a bootpicker when holding down alt booting native mojave with a Radeon VII and no OC?
Like for example to select booting off a USB

 

[tsialex]

Stupid question: Should i get a bootpicker when holding down alt booting native mojave with a Radeon VII and no OC?
Like for example to select booting off a USB

Nope, only flashed cards with Mac EFI can provide the BootPicker without OpenCore help.

 

[antonlinus]

Nope, only flashed cards with Mac EFI can provide the BootPicker without OpenCore help.

is it possible / helpfull to flash a Radeon VII with a Mac EFI?
If have been struggling with a fully clean from scratch build of my cMP for days now and its super frustrating to guess boot volumes blindly all the time when OC ist not active...

 

[tsialex]

is it possible / helpfull to flash a Radeon VII with a Mac EFI?

I wouldn't do it now, there are Mac Pro BootROM modifications/improvements on this topic in the near future, with free and commercial solutions being developed and tested. Since it's not public yet, I can't give you more info, but works 100% with RX 580/VEGA 56/RX 6600 and most recent AMD GPUs - I don't own a VII and don't know if works or not, VII cards are somewhat problematic.

If it's critical for you to it right now, it's possible to flash the VII, but it's not a flashing yourself procedure. You can ask MacVidCards Europe if they can provide flashing service for the VII.

 

[antonlinus]

I wouldn't do it now, there are Mac Pro BootROM modifications/improvements on this topic in the near future, with free and commercial solutions being developed and tested. Since it's not public yet, I can't give you more info, but works 100% with RX 580/VEGA 56/RX 6600 and most recent AMD GPUs - I don't own a VII and don't know if works or not, VII cards are somewhat problematic.

If it's critical for you to it right now, it's possible to flash the VII, but it's not a flashing yourself procedure. You can ask MacVidCards Europe if they can provide flashing service for the VII.

i guess its not that critical no. i can still use the terminal to select the bootvolume prior to rebooting but i will be looking forward to what you guys have cooking then

thanks a lot!

 

[Scubaman]

Anybody got Ventura working via OpenCore yet? I'm more than happy with Monterey on my Mac Pro 5,1 but Ventura beckons!!

 

[14dcutaneo]

Anybody got Ventura working via OpenCore yet? I'm more than happy with Monterey on my Mac Pro 5,1 but Ventura beckons!!

If you want Ventura support, you'll have to use OpenCore Legacy Patcher which uses root patches to gain support. Check to see if your hardware configuration has full support, as Navi GPU's aren't supported yet. Vanilla OpenCore like this will not work on Ventura.

 

[14dcutaneo]

This may be a long shot but I'm curious as to if anyone else is experiencing this. For the last two MacOS versions (12.6.2 and 12.6.3) and using the 0.8.7 and 0.8.8 package, I've noticed that around every two or so hours something within macOS just decides to basically read my NVME SSD to death. The system gets slowed down to a crawl. The reads in/sec go up to about 3-4000, and the writes go all the way down to 0 leaving the system to be unusable unless I restart or wait about 30 minutes. The amount of Data read/sec is about 50MB and is pretty much static the whole time. I've checked so many processes, and none of them have either excessive Bytes Read or high CPU utilization. Yea, I've done a malware scan and nobody is remotely connecting to my computer. For the life of me I cannot figure out what is causing macOS to read my SSD to the point of a crawl. But what bothers me most is that the processes within Activity monitor dont show any one process thats showing a high amount of "Bytes Read". The total amount of "Data Read" by activity monitor can show 500GB, meanwhile the process with the highest amount of "bytes read" only has 20GB, and totaling up the rest of them lead nowhere to 500GB.

Anyone ever experience anything like this before?

 

[Macschrauber]

is it possible / helpfull to flash a Radeon VII with a Mac EFI?
If have been struggling with a fully clean from scratch build of my cMP for days now and its super frustrating to guess boot volumes blindly all the time when OC ist not active...

You can make an OpenCore CD described by CDF in the OpenCore thread and boot it with the c-key like in the old days OS installers to select and bless systems.

OpenCore on the Mac Pro

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.

forums.macrumors.com

Go to complete your setup and scroll down to create a rescue CD.

 

[h9826790]

This may be a long shot but I'm curious as to if anyone else is experiencing this. For the last two MacOS versions (12.6.2 and 12.6.3) and using the 0.8.7 and 0.8.8 package, I've noticed that around every two or so hours something within macOS just decides to basically read my NVME SSD to death. The system gets slowed down to a crawl. The reads in/sec go up to about 3-4000, and the writes go all the way down to 0 leaving the system to be unusable unless I restart or wait about 30 minutes. The amount of Data read/sec is about 50MB and is pretty much static the whole time. I've checked so many processes, and none of them have either excessive Bytes Read or high CPU utilization. Yea, I've done a malware scan and nobody is remotely connecting to my computer. For the life of me I cannot figure out what is causing macOS to read my SSD to the point of a crawl. But what bothers me most is that the processes within Activity monitor dont show any one process thats showing a high amount of "Bytes Read". The total amount of "Data Read" by activity monitor can show 500GB, meanwhile the process with the highest amount of "bytes read" only has 20GB, and totaling up the rest of them lead nowhere to 500GB.

Anyone ever experience anything like this before?

Have you select "All Processes"?

 

[14dcutaneo]

Have you select "All Processes"?
View attachment 2148264

Yup, all processes are selected. I've even tried to search Console for any sort of errors, but I don't exactly have any hints as to what to be searching for with something as strange as this. The only thing I'm really seeing in the console repeatedly is AMPDeviceDiscoveryAgent[593]: Entered:_AMMuxedDeviceDisconnected, mux-device:521, which im pretty sure is unrelated as it corresponds with iOS devices trying to wirelessly sync. One time the computer actually logged itself out automatically and had me re-log in to re-launch everything. Never seen anything like it.

 

[h9826790]

Yup, all processes are selected. I've even tried to search Console for any sort of errors, but I don't exactly have any hints as to what to be searching for with something as strange as this. The only thing I'm really seeing in the console repeatedly is AMPDeviceDiscoveryAgent[593]: Entered:_AMMuxedDeviceDisconnected, mux-device:521, which im pretty sure is unrelated as it corresponds with iOS devices trying to wirelessly sync. One time the computer actually logged itself out automatically and had me re-log in to re-launch everything. Never seen anything like it.

Do you have any backup boot drive?

If no, you better prepare it now, just in case.

If yes, and if you have time, please try to boot from that drive, and see if the same problem exist.

Both 12.6.2 and 12.6.3 has no such issue on my cMP.

 

[antonlinus]

You can make an OpenCore CD described by CDF in the OpenCore thread and boot it with the c-key like in the old days OS installers to select and bless systems.

OpenCore on the Mac Pro

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.

forums.macrumors.com

Go to complete your setup and scroll down to create a rescue CD.

Yes i did that now thanks for the tip!

 

[cmphal]

0.8.8 macOS 12.6.3 working - with some restart/shutdown oddities

Did you check if there is any typo?

Yes, you called it correctly, a missing '<' due to my fingers expecting karabiner-elements remappings which are in my user account but not the admin account I use for macOS installs/updates.

Also, thanks for pointing out ocvalidate. I will try to use should I run into issues in the future. I must admit that I've only given a cursory viewing of the Dortania documentation and the OpenCore on MacPro thread.

I now have 0.8.8 with macOS 12.6.3 up and running. I had to edit
/System/Volumes/Preboot/uuid/System/Library/CoreServices/.disk_label.contentDetails to get the Boot Picker's drive name to match the boot volumes name. Is that the recommended way to deal with boot volume name changes or is there some other mechanism?

I do have a couple of issues with shutdown and reboot from the Apple menubar. This behaves slightly differently for my different accounts.

In my Admin account which is the cleanest, of I select restart, I see my desktop contents and menubar disappearing, then a momentary flash of a black screen, then the desktop background image with nothing else on it, at which point it waits for two minutes, then goes black and the boot picker. If I select shutdown, similar behavior with displaying the desktop background for two minutes, and then it will shutdown. Note, when I restart after the boot picker I do get the macOS flash screen stating "Your computer restarted because of a problem". This is for both a restart or shutdown.

In my main user account, there is slightly different behavior. Restart is the same as the admin account. However, for shutdown, it never shuts down. Rather the desktop only image is displayed for five minutes after which it displays the opencore bootpicker. I can shutdown by selecting the power button in the opencore bootpicker.

In all cases, macOS thinks ""Your computer restarted because of a problem" and after login I am presented with the dialogs for automatically restarting applications. I am puzzled by that because it starts up some applications even though I had quit them prior to selecting shutdown. It doesn't start everything. Just a couple of processes.

Also, shutdown -h now, in terminal has similar behaviour and doesn't really shutdown the computer but it does get me to the boot picker screen without the five minute wait.

I'm suspecting these restart and shutdown issues are not really due to opencore. Any thoughts on how to investigate?

 

[14dcutaneo]

Do you have any backup boot drive?

If no, you better prepare it now, just in case.

If yes, and if you have time, please try to boot from that drive, and see if the same problem exist.

Both 12.6.2 and 12.6.3 has no such issue on my cMP.

Yeah, I have a vanilla Mojave boot drive for emergencies. Hopefully its just some rogue application thats not reporting itself properly. I've got plenty of time machine backups too. Thanks for the reply!

 

[14dcutaneo]

This may be a long shot but I'm curious as to if anyone else is experiencing this. For the last two MacOS versions (12.6.2 and 12.6.3) and using the 0.8.7 and 0.8.8 package, I've noticed that around every two or so hours something within macOS just decides to basically read my NVME SSD to death. The system gets slowed down to a crawl. The reads in/sec go up to about 3-4000, and the writes go all the way down to 0 leaving the system to be unusable unless I restart or wait about 30 minutes. The amount of Data read/sec is about 50MB and is pretty much static the whole time. I've checked so many processes, and none of them have either excessive Bytes Read or high CPU utilization. Yea, I've done a malware scan and nobody is remotely connecting to my computer. For the life of me I cannot figure out what is causing macOS to read my SSD to the point of a crawl. But what bothers me most is that the processes within Activity monitor dont show any one process thats showing a high amount of "Bytes Read". The total amount of "Data Read" by activity monitor can show 500GB, meanwhile the process with the highest amount of "bytes read" only has 20GB, and totaling up the rest of them lead nowhere to 500GB.

Anyone ever experience anything like this before?

I don't want to get this thread too off topic with this, but I at least want to report that I miraculously found the process causing the trouble just in case anyone experiences the same thing in their OpenCore setup.

Basically, the process UVFSService was trying to index my Windows partition. It wasn't taking up much CPU and wasn't showing anything in the "disk" tab under "bytes read" so it was hard to spot. No idea why this is the case.

Solution: Exclude the partition or disks from spotlight search by going to System Preferences > Spotlight > Privacy, and add the disk(s) to the list. Immediately UVFSService went back down to 0.0% CPU and everything went back to normal