MP 1,1-5,1 MacPro5,1: BootROM thread | 144.0.0.0.0

[what2be]

How do you reset the EFI? I have a highpoint 7110 card setup in raid 0 that was bootable and working fine in Catalina with a RX5700 XT that is flashed. The problem started when I upgraded to Big Sur.

What I did: I put a old SATA drive in my 5,1 cMP and made a USB installer and everything went fine, Big Sur booted up but I couldn't get Big Sur to recognize my 7110 card so I downloaded the EFI package from highpoint for Mac but for the life of me it would not boot up into the EFI shell. All it would do is make my screen go blank as soon as I picked it from the startup manager screen. So then I tried using clover with their shell, and then I tried rEFInd boot manager and inserted a shell.efi into it so I could boot into the shell , none of those things worked. I need to boot into the Mac EFI shell so I can run the .nsh commands to initialize the card and delete/create the array. But now I have another problem, now when I leave both the RX5700 and the 7110 card in and try to boot the machine it just either stays on a black screen , or if I hold the option key down it will stay stuck on a white screen. Ive reset the pram, nvram, etc a hundred times, the only way to get it to boot is to remove the 7710 card from the machine and then it will boot up just fine. The only thing I can think of is the EFI table is corrupted because I've spent the last 2 days on this and tried so many different things I'm now out of ideas.

Any suggestions?

 

[tsialex]

How do you reset the EFI? I have a highpoint 7110 card setup in raid 0 that was bootable and working fine in Catalina with a RX5700 XT that is flashed. The problem started when I upgraded to Big Sur.

What I did: I put a old SATA drive in my 5,1 cMP and made a USB installer and everything went fine, Big Sur booted up but I couldn't get Big Sur to recognize my 7110 card so I downloaded the EFI package from highpoint for Mac but for the life of me it would not boot up into the EFI shell. All it would do is make my screen go blank as soon as I picked it from the startup manager screen. So then I tried using clover with their shell, and then I tried rEFInd boot manager and inserted a shell.efi into it so I could boot into the shell , none of those things worked. I need to boot into the Mac EFI shell so I can run the .nsh commands to initialize the card and delete/create the array. But now I have another problem, now when I leave both the RX5700 and the 7110 card in and try to boot the machine it just either stays on a black screen , or if I hold the option key down it will stay stuck on a white screen. Ive reset the pram, nvram, etc a hundred times, the only way to get it to boot is to remove the 7710 card from the machine and then it will boot up just fine. The only thing I can think of is the EFI table is corrupted because I've spent the last 2 days on this and tried so many different things I'm now out of ideas.

Any suggestions?


View attachment 1660364

MP5,1 EFI part of the BootROM firmware is not resettable or even modifiable in any way that you are thinking, for all effects is read-only. You probably are mixing the EFI partition where you can install and configure things, and resides on the disk, with the EFI part of the BootROM firmware that is stored on a SPI flash memory on the Mac Pro backplane.

Your problem for sure is elsewhere. Since you used Clover, the most probable is that using it you completely trashed the NVRAM volume, one boot is enough for damaging the NVRAM. If you previously saved a dump of from your Mac Pro BootROM, try flashing it back with ROMTool. If not, you will need a firmware reconstruction.

The UEFI firmware update for SSD7110 is to be installed on the card SPI flash memory, not on your Mac Pro BootROM. Btw, I'm not sure of the SSD7110, but you don't need to use any of the High Point configuration EFI tools with a Mac Pro that have SSD7101A-1/7102/7103/7104. The Mac Pro firmware itself initialises, and boot, from the card alone even without any drivers, then you use the HighPoint web configurator (HPTWEBGUI) to do anything configuration wise - after the HPT drivers are installed. Use HighSierra or Mojave with a supported GPU to do all the configuration, RX 5700 only works with 10.15.1+.

Don't use a bootable RAID array after High Sierra, you are asking for trouble. Use one NVMe drive/blade for booting and then you can create an array for scratch/fast storage/etc.

 

[what2be]

MP5,1 EFI part of the BootROM firmware is not resettable or even modifiable in any way that you are thinking, for all effects is read-only. You probably are mixing the EFI partition where you can install and configure things, and resides on the disk, with the EFI part of the BootROM firmware that is stored on a SPI flash memory on the Mac Pro backplane.

Your problem for sure is elsewhere. Since you used Clover, the most probable is that using it you completely trashed the NVRAM volume, one boot is enough for damaging the NVRAM. If you previously saved a dump of from your Mac Pro BootROM, try flashing it back with ROMTool. If not, you will need a firmware reconstruction.

The UEFI firmware update for SSD7110 is to be installed on the card SPI flash memory, not on your Mac Pro BootROM. Btw, I'm not sure of the SSD7110, but you don't need to use any of the High Point configuration EFI tools with a Mac Pro that have SSD7101A-1/7102/7103/7104. The Mac Pro firmware itself initialises, and boot, from the card alone even without any drivers, then you use the HighPoint web configurator (HPTWEBGUI) to do anything configuration wise - after the HPT drivers are installed. Use HighSierra or Mojave with a supported GPU to do all the configuration, RX 5700 only works with 10.15.1+.

Don't use a bootable RAID array after High Sierra, you are asking for trouble. Use one NVMe drive/blade for booting and then you can create an array for scratch/fast storage/etc.

Well, that's a lot to take in! Thank you for such a detailed reply. Ill try to be concise and address what my main concern was (and still is) below.

1) Re: clover, No backup of Mac Pro BootROM, but its booting right as long as I leave the SSD7110 out of the machine. Does that mean the bootrom is ok?

2) Re: UEFI firmware, Yes, the UEFI firmware update for the SSD7710 is installed onto the cards flash, and it must be run from a fat32 formatted USB key, which I made correctly but when I hold down the ALT key and start boot manager, I pick, EFI Boot and it just goes to a white screen and that's it. If I reset the pram and don't use the ALT key it simply boots into a grey screen and does nothing. I believe its because the boot.efi they supply is for a EFI that's not compatible with the one apple uses in the 5,1 Mac Pro. I can take the card and put it into a windows machine, create a different USB key with the windows UEFI files for the SSD7110 and it works perfect. But in the Mac, no way. And since I can't configure the array (or see it) I can't copy anything to it.

3) And that is my issue, ever since I put the card in a window machine, ran the flash for the card and then tried to put it back into my Mac I can no longer boot from anything on my Mac. Most of the time it just makes the chime noise really fast over and over again but the chime is a 1/4 of the length it usually is, it just goes, boing, boing, boing. Instead of BOOOOOOOIIINNNGG like it normally does. I need to figure out how to get the dame USB key to work with the Mac flash files on the USB key.

 

[tsialex]

Well, that's a lot to take in! Thank you for such a detailed reply. Ill try to be concise and address what my main concern was (and still is) below.

1) Re: clover, No backup of Mac Pro BootROM, but its booting right as long as I leave the SSD7110 out of the machine. Does that mean the bootrom is ok?

2) Re: UEFI firmware, Yes, the UEFI firmware update for the SSD7710 is installed onto the cards flash, and it must be run from a fat32 formatted USB key, which I made correctly but when I hold down the ALT key and start boot manager, I pick, EFI Boot and it just goes to a white screen and that's it. If I reset the pram and don't use the ALT key it simply boots into a grey screen and does nothing. I believe its because the boot.efi they supply is for a EFI that's not compatible with the one apple uses in the 5,1 Mac Pro. I can take the card and put it into a windows machine, create a different USB key with the windows UEFI files for the SSD7110 and it works perfect. But in the Mac, no way. And since I can't configure the array (or see it) I can't copy anything to it.

3) And that is my issue, ever since I put the card in a window machine, ran the flash for the card and then tried to put it back into my Mac I can no longer boot from anything on my Mac. Most of the time it just makes the chime noise really fast over and over again but the chime is a 1/4 of the length it usually is, it just goes, boing, boing, boing. Instead of BOOOOOOOIIINNNGG like it normally does. I need to figure out how to get the dame USB key to work with the Mac flash files on the USB key.

You can eliminate the Mac Pro BootROM from your problems easily. Dump your BootROM with ROMTool, save it securely (you will need to flash it back later) then flash the generic firmware upgrade image, MP51.fd from 10.14.6 (144.0.0.0.0). This will get you a Mac without serial number or any hardwareIDs, but a fresh NVRAM and will be useful for diagnosing.

Btw, like I written before use an AppleOEM GPU and 10.13.6/10.14.6. Eliminate your RX 5700 and unsupported macOS from the equation.

 

[Dewdman42]

@tsialex, I have question for you. I read all your posts with interest because you seem to know what you're doing.

I am contemplating whether to go Open Core on my MP5,1. Its not clear to me exactly where OpenCore gets installed. Does some part of it get installed into the logic board, or is all of it installed on my boot disk? In other words, theoretically, if I try out OpenCore and then want to pull out my boot disk and put another disk without Opencore into that same bay, will things be back to normal? Or will OpenCore have modified other flash memory somewhere?

I have seen you mention various possibilities for bricking a MP5,1 and its not entirely clear to me what the risky things are to do, and why it happens. I THINK I am understanding that the NVRAM could get corrupted to a degree that it won't even be possible to clear it again without special tools. Does part of Open Core get installed into NVRAM or is it simply that bad configurations of OpenCore could end up overwriting the NVRAM and bricking the mac?

Is there any other way to backup the NVRAM and get it back other then these other tools? I am not afraid of using some kind of restoration tool to restore the NVRAM? The possibility of bricking my mac has my a little scared to try out OpenCore. I don't want to do any crazy spoofing or anything, I just want to follow the default instructions on that other thread for a vanilla open core Catalina, Big Sure and probably Windows10 eventually, but I read some posts from you that Windows setups have sometimes corrupted the NVRAM, which makes me afraid again..

 

[tsialex]

@tsialex, I have question for you. I read all your posts with interest because you seem to know what you're doing.

I am contemplating whether to go Open Core on my MP5,1. Its not clear to me exactly where OpenCore gets installed. Does some part of it get installed into the logic board, or is all of it installed on my boot disk? In other words, theoretically, if I try out OpenCore and then want to pull out my boot disk and put another disk without Opencore into that same bay, will things be back to normal? Or will OpenCore have modified other flash memory somewhere?

I have seen you mention various possibilities for bricking a MP5,1 and its not entirely clear to me what the risky things are to do, and why it happens. I THINK I am understanding that the NVRAM could get corrupted to a degree that it won't even be possible to clear it again without special tools. Does part of Open Core get installed into NVRAM or is it simply that bad configurations of OpenCore could end up overwriting the NVRAM and bricking the mac?

Is there any other way to backup the NVRAM and get it back other then these other tools? I am not afraid of using some kind of restoration tool to restore the NVRAM? The possibility of bricking my mac has my a little scared to try out OpenCore. I don't want to do any crazy spoofing or anything, I just want to follow the default instructions on that other thread for a vanilla open core Catalina, Big Sure and probably Windows10 eventually, but I read some posts from you that Windows setups have sometimes corrupted the NVRAM, which makes me afraid again..

OpenCore is stored only on the EFI partition of your disk, nothing in the BootROM/logicboard/whatever.

Clover is the boot loader that can't be used and damage the Mac Pro NVRAM. OpenCore was designed to be used on Macs from the ground up and don't cause adverse effects.

NVRAM is part of the BootROM, you can only backup it dumping the full firmware image from the SPI flash memory, using flashrom/ROMTools.

 

[Dewdman42]

( moved to Open Core thread)

 

[tsialex]

Thank you for clarifying, sorry for simple questions...



Right so I interpret that to mean that Clover can and will write into the NVRAM, which is the BootROM (which is not on any disk), and that is how NVRAM can get corrupted that people have described on this forum thread. But you believe OpenCore should not make those kinds of mistakes in corrupting the NVRAM, so should be ok to try it out without too much concern about corrupting the NVRAM.

Does that include when I boot to MSWIndows? I read some post on one of these threads that there was some situations with MS Windows where it might have overwritten past some buffer limits in NVRAM or something and possibly corrupted it, even using OpenCore EFI windows install?



Thank you.

So a few more questions to help me figure out the best workflow to attempt this stuff... I need to understand exactly what the proper procedure would be to rollback out of OpenCore if I decide its not working for me.

So I think you're saying that if NVRAM gets corrupted because of a stupid boot loader (Clover), then might have to get some tricky software or maybe even hardware-level tools to restore the NVRAM back to a state where the machine can even be booted at all...ie...bricked. I'm not prepared to deal with that, and I definitely won't use Clover, so hopefully NVRAM wouldn't get corrupted any other way. If that is unlikely scenario with OpenCore, then I won't worry about it right now.

However, one other question.. Let's say I want to try OpenCore... Assume my system has 3 or 4 SSD's installed, any of which could be used for booting if I choose. Ok, so I have normal non-OpenCore Mojave installed on my primary drive. let's say I setup a completely separate drive with the OpenCore boot loader and install Catalina. Fine so far. And in those steps the NVRAM is changed (I think) to indicate that the mac should boot with OpenCore Bootloader which is on only one of the drives in the computer.

Alright, so I do that, I get it all working and play with it a while, but then for whatever the reason I want to pull it out and go back to my old non-OpenCore Mojave boot. I presume in that case I could not just pull out the OpenCore drive and boot up the machine, I would need to also configure the NVRAM to now look back to the original drive with non-OpenCore Mojave on it...then I'd be good to go back and could remove the OpenCore drive entirely and boot up with original Mojave, right?

what would happen if I pulled out the open core drive without reconfiguring NVRAM (if I have that right) to look back for the original Mojave non-OpenCore bootable volume? That's not really completely a brick because I could put the OpenCore drive back in, fix it and move on, but I just want to understand the ramifications of rollback before I try anything.

I think ultimately I will want the drive in drive bay #1 to have the open core boot loader and Mojave, and then I will have various other drives or partitions with Catalina, BigSur and Windows10, maybe even linux later. Sounds like rolling it back would simply be remove the bootloader and reconfigure the NVRAM and reboot, then it would be back to normal MacOS approach. yea?

And also, is there any reason to believe that my Mojave boot volume would act funny while booting under OpenCore...would it require hardware acceleration hacks and what not like is required for Catalina and Big Sur?

Please ask your OpenCore related questions on the OpenCore thread, here is not the place.

 

[Dewdman42]

sorry will remove the post and repost it there. I just wanted to hear from you specifically and here you were.. ;-) regards.

 

[tsialex]

sorry will remove the post and repost it there. I just wanted to hear from you specifically and here you were.. ;-) regards.

Please don't ask questions directly to me on the OpenCore thread, there lot's of people there that can help you too.

I have limited time and I usually only answer questions related to firmware/low level development/hardware, not questions about OC installs/configuration.

 

[Dewdman42]

I respect that. I've already posted it though. hehe. sorry. what I have observed on that other thread is too many cooks in the kitchen, I saw your repeated advice to a void spoofing the MP5,1 as an iMac, yet still that is what people there are saying to do... Your words have gained my respect much more, but i understand if you don't have time to respond.

 

[tsialex]

I respect that. I've already posted it though. hehe. sorry. what I have observed on that other thread is too many cooks in the kitchen, I saw your repeated advice to a void spoofing the MP5,1 as an iMac, yet still that is what people there are saying to do... Your words have gained my respect much more, but i understand if you don't have time to respond.

Spoofing is a mean to an end, you shouldn't use it if you don't needed it since there are drawbacks and initially had adverse effects with certain variables inside the NVRAM stores, but this specific issue has been solved.

In my personal opinion, the only reasonable use of spoofing, before BigSur release, was to have VideoToolBox acceleration. Lot's of people are now using spoofing to have hardware assisted video decompression/compression/DRM support with Safari, but it's a very cumbersome/blunt solution that changes lot's of things and a more elegant/precise one needs to be developed.

OC works fine, if you have the hardware configuration required and a clean BootROM (save a BootROM dump before anything), you shouldn't have problems.

Btw, if you have an early-2009 Mac Pro cross-flashed to MP5,1 firmware via Netkas forum/MacEFIROM tools, I strongly recommend that you don't use OpenCore before a full BootROM reconstruction.

 

[Dewdman42]

its a mid 2010, with updated airport extreme. 12 core at 3.46 cpu. I think it should everything as updated as a 5,1 can be..

 

[what2be]

You can eliminate the Mac Pro BootROM from your problems easily. Dump your BootROM with ROMTool, save it securely (you will need to flash it back later) then flash the generic firmware upgrade image, MP51.fd from 10.14.6 (144.0.0.0.0). This will get you a Mac without serial number or any hardwareIDs, but a fresh NVRAM and will be useful for diagnosing.

Btw, like I written before use an AppleOEM GPU and 10.13.6/10.14.6. Eliminate your RX 5700 and unsupported macOS from the equation.

OK, I should have kept it simple from the start, just as you suggested.
1) Installed ROMTool, dumped my bootrom.
2) Removed 5700XT and put in Radeon HD 5870, Installed Catalina on a old SATA drive In my Mac Pro.
3) Booted a usb key with the Highpoint bios and it successfully flashed the 7710 SPI.
4) Created a USB Catalina install and booted from it and installed catalina onto the SATA drive.It installed successfully.
5) Installed my Highpoint 7110 drivers, rebooted and then verified both loaded via system report, software, extensions.
6) Ran Disk Utility, Formatted the raid as OSX Journaled. Installed Carbon Copy Cloner and cloned the SATA drive to the Highpoint 7110 partition. Removed the SATA drive, booted from the 7710 and everything is great. Works perfect.

I then put the SATA drive back into the Mac Pro, formatted it, installed Big Sur and it booted fine but after installing the 7710 drivers it only showed 1 of the 2 drivers loaded when viewed in system report. When I ran the Highpoint web gui instead of showing adapter 1 (NVME) and 2 (SAS) it only showed adapter 2 (SAS). so.....

I ran Disk Utility and it showed the 3 blades installed on my NVME card so I created a RAID set via Disk Utility and it worked fine (although I couldnt boot off it and benchmarking it with AJA dropped read/write speeds (see attached) Since the whole point of buying that card was to use the NVME portion of it to boot off of and the SAS portion of it to replace the SATA drives I previously had setup as a raid set its really not much use for me to keep in this configuration.

While researching I found a post from someone that has a very similar issue and the person who replied sounds like they are correct. If so, then I assume the only way to resolve my issue is to hope Highpoint comes out with a new driver that can boot their cards from Big Sur.

Is there any way to verify if mine is corrupted or not? I never got clover to boot so perhaps it didnt damage my NVRAM?
Can you provide a link to rebuild my NVRAM aka, firmware reconstruction as you had previously mentioned if need be?

Thanks!

 

[tsialex]

OK, I should have kept it simple from the start, just as you suggested.
1) Installed ROMTool, dumped my bootrom.
2) Removed 5700XT and put in Radeon HD 5870, Installed Catalina on a old SATA drive In my Mac Pro.
3) Booted a usb key with the Highpoint bios and it successfully flashed the 7710 SPI.
4) Created a USB Catalina install and booted from it and installed catalina onto the SATA drive.It installed successfully.
5) Installed my Highpoint 7110 drivers, rebooted and then verified both loaded via system report, software, extensions.
6) Ran Disk Utility, Formatted the raid as OSX Journaled. Installed Carbon Copy Cloner and cloned the SATA drive to the Highpoint 7110 partition. Removed the SATA drive, booted from the 7710 and everything is great. Works perfect.

I then put the SATA drive back into the Mac Pro, formatted it, installed Big Sur and it booted fine but after installing the 7710 drivers it only showed 1 of the 2 drivers loaded when viewed in system report. When I ran the Highpoint web gui instead of showing adapter 1 (NVME) and 2 (SAS) it only showed adapter 2 (SAS). so.....

I ran Disk Utility and it showed the 3 blades installed on my NVME card so I created a RAID set via Disk Utility and it worked fine (although I couldnt boot off it and benchmarking it with AJA dropped read/write speeds (see attached) Since the whole point of buying that card was to use the NVME portion of it to boot off of and the SAS portion of it to replace the SATA drives I previously had setup as a raid set its really not much use for me to keep in this configuration.

While researching I found a post from someone that has a very similar issue and the person who replied sounds like they are correct. If so, then I assume the only way to resolve my issue is to hope Highpoint comes out with a new driver that can boot their cards from Big Sur.

Is there any way to verify if mine is corrupted or not? I never got clover to boot so perhaps it didnt damage my NVRAM?
Can you provide a link to rebuild my NVRAM aka, firmware reconstruction as you had previously mentioned if need be?

Thanks!

You are still holding it wrong, never diagnose Mac Pro hardware with spoofed or hacked installs, the results are not meaningful. Use completely supported/vanilla macOS installs - with an AppleOEM GPU like HD 5870, you need to test from High Sierra.

The way Apple implemented Big Sur kext support makes extremely difficult, if not impossible unless Apple starts to include HighPoint drivers with Big Sur install, to boot from an array - you can't load the HighPoint kext from the sealed snapshot.

 

[Steve Expat]

@tsialex, I know you're getting a lot of questions. I hope you don't mind one more. I have a 2009 cMP that has been upgraded steadily since about 2016 (Amfeltech Squid carrier card, m.2 PCIe boot drives, graphics cards, CPUs, etc..). I'm currently running 10.14.6 on an NVME m.2 in PCIe slot 3. I have a 12-core (dual x5690s) and a flashed MSI Armor RX580X.

The problem started when I bought and attached an LG 32UL500-W 4k display. The LG is connected via the DisplayPort (and set at 3840 x 2160 (2160p 4K UHD - Ultra High Definition) but the UI Looks like: 1920 x 1080 @ 60 Hz.

Connected to the DVI port is a HP 27xi which has a native resolution of 1920 x 1080 (1080p FHD - Full High Definition), however, I set this monitor so that the UI Looks like: HiDPI 960 x 540 @ 60 Hz.

When I boot the MP, I have to unplug the LG display (and I unplug almost every USB cable that could draw power), before the computer will just show a grey screen (the Apple logo usually doesn't appear on the first attempt). I usually have to press the power button to shut down and reboot once or twice more to get the Apple logo to appear on the HP display, then I can connect the LG display and everything works fine.

I considered that I may be drawing too much power. Maybe I am. But I'm not convinced. Have you seen this problem, and do you know of a fix? This was not happening until I got the LG 4k display. It got worse last week when I replaced the flashed Gigabyte Windforce AMD R9 280x 3GB card with the flashed MSI Armor Radeon RX 580 8 GB. I've considered looking into Open Core, which I'll do later this week. I want to figure out if I'm drawing too much power. I am powering quite a few "low power draw" USB devices (3 dongles and an audio interface), and have other USB devices connected that are powered.

My specs are:

Boot ROM Version: 144.0.0.0.0
Mac Pro (Early 2009)
2 x 3.46 GHz 6-Core Intel Xeon
96 GB 1333 MHz DDR3
Macintosh 1TB NVME
Radeon RX 580 8 GB

Thank you kindly for any advice you can offer.

Regards,
Steve

 

[tsialex]

@tsialex, I know you're getting a lot of questions. I hope you don't mind one more. I have a 2009 cMP that has been upgraded steadily since about 2016 (Amfeltech Squid carrier card, m.2 PCIe boot drives, graphics cards, CPUs, etc..). I'm currently running 10.14.6 on an NVME m.2 in PCIe slot 3. I have a 12-core (dual x5690s) and a flashed MSI Armor RX580X.

The problem started when I bought and attached an LG 32UL500-W 4k display. The LG is connected via the DisplayPort (and set at 3840 x 2160 (2160p 4K UHD - Ultra High Definition) but the UI Looks like: 1920 x 1080 @ 60 Hz.

Connected to the DVI port is a HP 27xi which has a native resolution of 1920 x 1080 (1080p FHD - Full High Definition), however, I set this monitor so that the UI Looks like: HiDPI 960 x 540 @ 60 Hz.

When I boot the MP, I have to unplug the LG display (and I unplug almost every USB cable that could draw power), before the computer will just show a grey screen (the Apple logo usually doesn't appear on the first attempt). I usually have to press the power button to shut down and reboot once or twice more to get the Apple logo to appear on the HP display, then I can connect the LG display and everything works fine.

I considered that I may be drawing too much power. Maybe I am. But I'm not convinced. Have you seen this problem, and do you know of a fix? This was not happening until I got the LG 4k display. It got worse last week when I replaced the flashed Gigabyte Windforce AMD R9 280x 3GB card with the flashed MSI Armor Radeon RX 580 8 GB. I've considered looking into Open Core, which I'll do later this week. I want to figure out if I'm drawing too much power. I am powering quite a few "low power draw" USB devices (3 dongles and an audio interface), and have other USB devices connected that are powered.

My specs are:

Boot ROM Version: 144.0.0.0.0
Mac Pro (Early 2009)
2 x 3.46 GHz 6-Core Intel Xeon
96 GB 1333 MHz DDR3
Macintosh 1TB NVME
Radeon RX 580 8 GB

Thank you kindly for any advice you can offer.

Regards,
Steve

Please don't hijack threads, this has noting to do with the topic of the thread.

Anyway, your problem for sure is not power draw related, unless you are powering the GPU with the wrong cables. At least, it's not the type of power draw that will cause any sweat for a perfectly working Mac Pro.

Something is going on awry with your hardware, maybe your PSU started to die around the time you installed your 4k display, maybe your the power plane of your backplane is defective, maybe a cable is damaged/shorted - lot's of possibilities here.

I'd start diagnosing the Mac Pro running AHT and ASD, if the tools show you an error, you will have a clue on what it's needed to investigate.

 

[Steve Expat]

Please don't hijack threads, this has noting to do with the topic of the thread. Anyway, your problem for sure is not power draw related, at least it's not the type of power draw that will cause any sweat for a perfectly working Mac Pro.

Something is going on awry with your hardware, maybe your PSU started to die around the time you installed your 4k display, maybe your the power plane of your backplane is defective, maybe a cable is shorted - lot's of possibilities here.

I'd start diagnosing it with AHT/ASD, if the tools show an error, you will have a clue on what you need to investigate.

My apologies. I didn't realize I hijacked this thread. My first thought (other than PSU), was that my issue may have something to do with the BootROM version in this 4,1/5,1 which is why I left it here. Thanks for the tips.

 

[tsialex]

My apologies. I didn't realize I hijacked this thread. My first thought (other than PSU), was that my issue may have something to do with the BootROM version in this 4,1/5,1 which is why I left it here. Thanks for the tips.

BootROM problems are rarely intermittent, this don't seem related to any BootROM problems that I know of.

The only GPUs that are directly affected by NVRAM problems are NVIDIA TITAN Black and TITAN X - caused by a very weird interaction between MP4,1>5,1 NVRAM and some settings stored on the NVRAM by old NVIDIA web drivers.

While your Mac Pro can have BootROM problems, I'm almost certain that it has some sort of problems since it's a 11 year old Mac and worse, it's an hybrid MP4,1>5,1 (MP5,1 EFI, MP4,1 NVRAM/BootBlock/hardwareIDs), what you described is essentially something power related.

You can dump your BootROM and flash the 144.0.0.0.0 MP51.fd from 10.14.6 to eliminate any chance that your Mac Pro BootROM has any culprit on this. The generic upgrade image don't have hardwareIDs, so, you can't login to iCloud/Messages/Facetime while flashed with MP51.fd/without the hardwareIDs, but it will re-initialize the NVRAM from scratch.

Btw, from experience, it's never just one thing that cause weird problems.

 

[macguru8]

Hello tsialex, I was trying to PM you but I went to the wrong cite

 

[paulcons]

@tsialex You won't remember, but a long ways back you gave me some advice and answered some PIA question about taking my cMP up to the 144.0.0 boot rom. Followed them to a T and everything worked great (almost, there are some issues with code signing with the last HS security update that have killed 3 apps for me, none are critical though). So THANKS!

 

[xch]

Hello, I have a 4.1>5.1 Mac Pro. Can someone tell me how can I know if my BootROM or NVRAM is corrupted or damaged, before it gets bricked? Thanks.

 

[tsialex]

Hello, I have a 4.1>5.1 Mac Pro. Can someone tell me how can I know if my BootROM or NVRAM is corrupted or damaged, before it gets bricked? Thanks.

This is a too complex topic, you really need to know what you are looking for, but if you know your way around an hex editor:

1. you can see the empty space between the first NVRAM volume VSS store and the second one, if you have variables up to the beginning of the second VSS store, your NVRAM is full,
2. count the MemoryConfig variables, if you have more MemoryConfigs than the quantity of DIMMs installed on your CPU tray inside the first VSS store, your Mac Pro NVRAM is suffering from variable duplication/multiplication,
3. count the bluetoothActiveControllerInfo variables per VSS store, if you have more than one per VSS store, your Mac Pro NVRAM is suffering from variable duplication/multiplication.

There are several other things that show NVRAM fragmentation/corruption, but are too complex to explain and outside the scope of a rapid post.

 

[xch]

This is a too complex topic, you really need to know what you are looking for, but if you know your way around an hex editor:

1. you can see the empty space between the first NVRAM volume VSS store and the second one, if you have variables up to the beginning of the second VSS store, your NVRAM is full,
2. count the MemoryConfig variables, if you have more MemoryConfigs than the quantity of DIMMs installed on your CPU tray inside the first VSS store, your Mac Pro NVRAM is suffering from variable duplication/multiplication,
3. count the bluetoothActiveControllerInfo variables per VSS store, if you have more than one per VSS store, your Mac Pro NVRAM is suffering from variable duplication/multiplication.

There are several other things that show NVRAM fragmentation/corruption, but are too complex to explain and outside the scope of a rapid post.

Thank you very much @tsialex. I'm very new to this things and honestly I don't understand much of what you're saying. Just start reading this threads and got a bit worried. I tried OpenCore for professional reasons and I'm afraid that could be a problem for my 4.1-5.1

 

[tsialex]

Thank you very much @tsialex. I'm very new to this things and honestly I don't understand much of what you're saying. Just start reading this threads and got a bit worried. I tried OpenCore for professional reasons and I'm afraid that could be a problem for my 4.1-5.1

I'll send you a PM with instructions, get everything done and I'll diagnose your dump against my signatures/tools.