Tried that first and it's not executable independently. Again, would recommend most people just use the Apple installer with an EFI card.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
MP5,1: New firmware on 10.13.4 MAS full installer: MP51.0085.B00 **OBSOLETE**
- Thread starter tsialex
- Start date
- Sort by reaction score
- Status
Tried that first and it's not executable independently. Again, would recommend most people just use the Apple installer with an EFI card.
I used the full 10.13.4 install package from the App Store to get the firmware upgrade process going. I was expecting a smooth ride (I'd had no problems with the original update to 51.084 for HS), but it was a bit more fraught this time.
I use a Sapphire RX 560 as my main GPU, but I have a GT120 on a shelf for doing upgrades such as this. I pulled the 560 and swapped it directly with the 120 after the first shutdown. When I powered up using the long press on the switch, I got the beep as normal but nothing on the screen (even though I had my monitor connected to the DP port of the GT120). I left it about ten minutes before powering down and trying a normal boot -- still nothing on the screen and the creeping threads of panic started to wrap themselves around me.
I then pulled the GT120 and put it in slot 3 instead, with the RX 560 back in its usual home of slot 1. When I booted this time everything was fine, and System Information reported that I'd been upgraded to 51.085. I then completed the 10.13.4 upgrade using the App Store upgrade process, which ran smoothly.
As others have said, I'm not sure what this firmware upgrade does, but 10.13.4 as fixed sleep for me (it was fine under 10.12.6 but wouldn't wake up under 10.13.3) and my RX 560 is now correctly reported in About this Mac and System Profiler.
In future I won't disturb the RX 560 and will just put the GT 120 in along side it in slot 3 for doing these upgrades. For some reason it didn't like being the only card in the system in a slot previously occupied with the 560.
/csd
I use a Sapphire RX 560 as my main GPU, but I have a GT120 on a shelf for doing upgrades such as this. I pulled the 560 and swapped it directly with the 120 after the first shutdown. When I powered up using the long press on the switch, I got the beep as normal but nothing on the screen (even though I had my monitor connected to the DP port of the GT120). I left it about ten minutes before powering down and trying a normal boot -- still nothing on the screen and the creeping threads of panic started to wrap themselves around me.
I then pulled the GT120 and put it in slot 3 instead, with the RX 560 back in its usual home of slot 1. When I booted this time everything was fine, and System Information reported that I'd been upgraded to 51.085. I then completed the 10.13.4 upgrade using the App Store upgrade process, which ran smoothly.
As others have said, I'm not sure what this firmware upgrade does, but 10.13.4 as fixed sleep for me (it was fine under 10.12.6 but wouldn't wake up under 10.13.3) and my RX 560 is now correctly reported in About this Mac and System Profiler.
In future I won't disturb the RX 560 and will just put the GT 120 in along side it in slot 3 for doing these upgrades. For some reason it didn't like being the only card in the system in a slot previously occupied with the 560.
/csd
This is what I did originally -- just the GT120 was installed. This worked insofar as the firmware got updated, but there was nothing displayed on the screen, which made me nervous!
In future, I will modify my procedure to only have the GT120 installed, but install it in a slot different to the one the RX 560 was in.
/csd
Today Intel published the updated Microcode Update Guidance. The production versions of the microcode for the Mac Pro 4,1/5,1 Xeons are 0x1C/0x1E (exactly like previous Dell's servers updates).
0x1C for Nehalem EP/WS Xeons
0x1E for Westmere EP/WS Xeons
Microcode on the MP51.0085.B00 still on the unmitigated 0xF version.
Note: Xeon W3690 will not be getting the update since Intel weirdly stopped Gulftown i7/Xeons microcode update.
0x1C for Nehalem EP/WS Xeons
0x1E for Westmere EP/WS Xeons
Microcode on the MP51.0085.B00 still on the unmitigated 0xF version.
Note: Xeon W3690 will not be getting the update since Intel weirdly stopped Gulftown i7/Xeons microcode update.
Last edited:
I was seriously thinking on upgrading my single processor to W3690, already had it on my AliExpress cart. With today news, I have to forget W3690 and downgrade to W3680 or spend a lot more on a X5690.
Last edited:
According to NVIDIA, the macOS security updates in 10.13.2 or 10.13.3 patched several of the variants related to the GPU if they were used in conjunction with NVIDIA web drivers. If the MP51.0085.B00 firmware did not address the remaining patch, should we expect to see another firmware update?
Any idea on what the MP51.0085.B00 firmware did address if it wasn't this?
FYI, I'm seeing pretty comparable prices for X5690 and W3690 on quick searches on AliExpress. Looks to be less than $30 more.
Any idea on what the MP51.0085.B00 firmware did address if it wasn't this?
FYI, I'm seeing pretty comparable prices for X5690 and W3690 on quick searches on AliExpress. Looks to be less than $30 more.
I've read that the APFS module was the reason for the updated firmware - almost all of the High Sierra supported Macs got it.
I'm seeing a little bit more, W3680 for $73 and X5690 for $106.72. Thinking on getting the X5690, maybe later on the road I'll get another and install both on my dual-processor.
And yet it was only accessible with a full download. Either the firmware update was necessary or it wasn't. No idea what the thinking behind this was.
im curious why intel is segregating the microcode update to this extent...
the W3680 and W3690 are practically identical apart from clock speeds.
Does anyone here have a System that got a BIOS update with the new microcodes? (or a system that you have patched the microcodes into the BIOS yourself LGA771 on LGA 775 style)
if you do, it would be very interesting see what microcode gets loaded for a W3680 vs a W3690, as at least back in the LGA775 LGA771 days microcode was only differentiated on platform (socket type) and architecture (it was not CPUID levels of specific)
It’s CPUID based, if you look on the table, Gulftown have a different one.
Gulftown = 206C2
Westmere = 206F2
206F2 is for Westmere-EX...
206C2 is the CPUID for Gulftown Westmere-EP and Westmere-WS AFAIK
to further this here is an excerpt from the Apple XNU source code to do with CPUIDs (J137 is the code name for the iMac Pro in case anyone is wondering, also fun fact if you check out the 10.4.11 intel, version of this source code file you can see Apple where "supporting" a Lot more CPUs and that they had added Penryn support despite no Mac shipping with Penryn at the time
)(Also my bad, by CPUID levels of specific I meant the past microcodes never cared what the CPUID name string was ie they never cared about this: "Intel(R) Xeon(R) CPU X5677 @ 3.47GHz" all they cared about was the General CPUID and the Platform ID )
Last edited:
Seems Intel did a petty move - if ever Apple update microcode they will have to add checks to differentiate Gulftown/Bloomfield/Clarksfield206F2 is for Westmere-EX...
206C2 is the CPUID for Gulftown Westmere-EP and Westmere-WS AFAIK
to further this here is an excerpt from the Apple XNU source code to do with CPUIDs (J137 is the code name for the iMac Pro in case anyone is wondering, also fun fact if you check out the 10.4.11 intel, version of this source code file you can see Apple where "supporting" a Lot more CPUs and that they had added Penryn support despite no Mac shipping with Penryn at the time
View attachment 757098
CPUs…
which makes me wonder is the Microcode actually checking for the Brand/Model string or is intel Just saying the W3690 is not supported even tho it does get its microcode updated?
only way to tell is test the CPUs out in a patched system and see what Microcode version is reported exactly, Once intel puts out the microcode that the Linux guys use people should be able to use that to patch PC X58 BIOSs with the new microcode even if the system vendor does not put out an update. (in the same way people add LGA 771 Microcode to LGA 775 boards, or how I added Wolfdale/Penryn LGA775 microcode to couple LGA 775 boards I have that do/did not support Penryn and lacked microcode for them so until I patched in microcode I had no SSE4.1 or VT-x heres a couple CPU Z validations of the systems I added penryn support too
http://valid.x86.fr/b2y13x and http://valid.x86.fr/56lzvi )
Has it anything to do with verbrose text during startup with APFS and flashed Nvidia cards? Does it fixes it?
I do not know what needs to be 'fixed', but the verbose text remains unchanged with the latest firmware/OS update.
Regards, splifingate
You are right about the firmware installation issues. I made a dedicated thread about it:
https://forums.macrumors.com/threads/warning-do-not-install-high-sierra-from-usb-1st-time.2116175/
In short: Do NOT install from USB the 1st time you install High Sierra. It will not work.
https://forums.macrumors.com/threads/warning-do-not-install-high-sierra-from-usb-1st-time.2116175/
In short: Do NOT install from USB the 1st time you install High Sierra. It will not work.
Downloaded the 10.13.4 installer and upgraded from MP51.0084.B00 to the 0085.B00 version. It took a few efforts:
1. You can install without running the rest of the installer, if you are already running a current or beta version of macOS and do not want to install over that.
2. Firmware update did not complete from a boot drive connected via a third party card, it needed to be on the native controller.
3. I have a Sierra installation in one of my standard locations. The first time I waited for the light to flash plus the sound. This did not take. On the second attempt I stopped pressing the power button with the flashing light. It beeped, reboot chimed, opened the DVD drive doors, trundled, and rebooted. The EFI display driver does not work with my Dell 38" screen so I had to fly blind. But all ended well. After reboot I cancelled out of the install.
I intend to check out the speed of the APFS boot into High Sierra on a third party controller and SSD.
My God, those old spinning drives are slow especially when not having run for a while and they are updating caches.
Later: no change in boot times, but apparent improvement in actually going to sleep. I had to force it from the menu most of the time.
1. You can install without running the rest of the installer, if you are already running a current or beta version of macOS and do not want to install over that.
2. Firmware update did not complete from a boot drive connected via a third party card, it needed to be on the native controller.
3. I have a Sierra installation in one of my standard locations. The first time I waited for the light to flash plus the sound. This did not take. On the second attempt I stopped pressing the power button with the flashing light. It beeped, reboot chimed, opened the DVD drive doors, trundled, and rebooted. The EFI display driver does not work with my Dell 38" screen so I had to fly blind. But all ended well. After reboot I cancelled out of the install.
I intend to check out the speed of the APFS boot into High Sierra on a third party controller and SSD.
My God, those old spinning drives are slow especially when not having run for a while and they are updating caches.
Later: no change in boot times, but apparent improvement in actually going to sleep. I had to force it from the menu most of the time.
Last edited:
For info, my cMP can be upgrade via a SSD that installed onto the Sonnet Tempo SSD card. So, point 2 is not entirely correctly. However, I do agree that the storage and its connection method can cause firmware upgrade failure.
For point 3, if there is any display port setting in your Dell monitor, set it to DP 1.1, then the boot screen should work. And it's the very first post I've seen to confirm you don't actually need to display the firmware upgrade process, but only need a Mac EFI graphic card installed. Thanks for this critical info.
I have several versions of Hi Sierra on various partitions on a single spinner HDD but I only wanted the firmware upgrade for my 4,1-5,1. as I intend to wait until the last High Sierra upgrade is released .. . and then go with a pure, clean HS installation.
===================================
Today, for those who are not sure how to get the firmware upgrade.
My method
I did a clean install of Mavericks 10.9.5 on a spare HDD.
Logged into the App Store
Searched for "High Sierra.
Selected install.
Then the "firmware update required" message popped up.
Followed the instruction to SHUT DOWN the cMP.
Restarted holding down the power button 'til the light started blinking.
It started up in Mavericks 10.9.5 with MP51.0085.B00 showing in System Info
Done.
Now I won't be touching High Sierra 'til the Golden Master is released.
===================================
Today, for those who are not sure how to get the firmware upgrade.
My method
I did a clean install of Mavericks 10.9.5 on a spare HDD.
Logged into the App Store
Searched for "High Sierra.
Selected install.
Then the "firmware update required" message popped up.
Followed the instruction to SHUT DOWN the cMP.
Restarted holding down the power button 'til the light started blinking.
It started up in Mavericks 10.9.5 with MP51.0085.B00 showing in System Info
Done.
Now I won't be touching High Sierra 'til the Golden Master is released.
I can confirm that the firmware upgrade is no problem with drives installed onto Caldigit Fasta-6gu3 Pro.
I checked with great hopes, but alas this option does not exist in the U3818DW option screens that I could see. I had the boot screen work via HDMI, but when I tried it this time it did not. I have a regular Cinema on my other system in another part of the house, but I was too lazy to schlepp that and just prayed.
[doublepost=1524694516][/doublepost]
Okay, then I presume it was user error. Next time (if there is one,) I will only wait for the light to flicker!
- Status