Z490 & Z590 - Will Z590 ever have macOS Support ?

[vipermachine]

The Z590 Vision G and Z590i from Gigabyte are very similar, so ochang's EFI is a good starting point. The USB map is the first thing to customize. It can be done by following this guide:

The New Beginner's Guide to USB Port Configuration

Important Update: June 2023 From Big Sur 11.3 on the built-in OpenCore XhciPortLimit quirk no-longer worked as described. This became a known "bug", though in reality it was Apple changing their internal code, making a simple "peek and poke" to keep it operating, a difficult task. However, with...

www.tonymacx86.com

Once that's out of the way we can tackle any subsequent post-install issue. For example, if the Ethernet port is not working or HDMI audio is not working, etc. those can be addressed one by one later on.

At this time it's probably best to stay with the original Thunderbolt firmware. Hot plug should work with most devices once the Maple Ridge Thunderbolt SSDT is installed.

I'm still willing to experiment with activating Thunderbolt Bus via SSDT-TbtOnPCH, but this can wait until the system is fully configured.

Hi @CaseySJ


I tried @ori69 EFI and had some problems with the USB and my TB equipment didnt show up.
Weird since ori has it working on his system.

I managed to make my own USB Patch,and now thunderbolt shows as expected.(No Hotplug yet,I tried the Mapple Ridge patch from you)
Here some files i extracted IOReg/Macaisl and Hackintool PCI
I also attached my EFI. just to check if al is good. and proceed from here.

I use a RX 580,i didnt do anything for igpu patching.

Thanks.

 

[dehjomz]

I am posting my EFI directory.
USB mapping done.
Audio USB works.
View attachment 513553
Thunderbolt now works every time, but is not detected in the profiler.

View attachment 513554


It disappears after sleep and it is necessary to shut down the computer (do not restart). There is no hot-plug.
I stick my connection to IO.
Sleep don't work.

It appears that Gigabyte's current Thunderbolt 4 implementation on their early Z590 BIOS may be broken as it pertains to Thunderbolt hot plug. I don't know if it's the BIOS or if it's the controller firmware that's the problem.

Outside of Windows 10, once a Thunderbolt device is disconnected from the Z590 Vision D, it will never reconnect until the system is cycled off and on. This is what is happening in Linux/macOS on the Z590 Vision D, but it doesn't happen on the MSI Z590 Ace, and, apparently according to this thread, the Asus Hero XIII. The only issue I have with the Asus Hero is that it doesn't appear to have a DisplayPort input, unlike the Vision D and the Ace.

Somehow on the Gigabyte Z590 Vision D, Thunderbolt hot plug does work in Windows, but not in the sense that we have come to understand. On the Z590 Vision D, when Windows goes to sleep and awakens, Thunderbolt devices lose power for a few milliseconds, and then Windows reengages the bus, and the driver hot plugs the device(s). This isn't a problem if your Thunderbolt device is a network card .or audio, but if your device is media, it will disconnect, which can be catastrophic.

So for example, if you boot windows on the Z590 Vision D from a USB device plugged into a Thunderbolt dock for example, upon sleep, you cannot successfully reawaken because for some unknown reason, the media disconnects before Windows has a chance to reengage the disk. The system will hang and then reboot. This doesn't happen on the MSI Z590 Ace... when it goes down to sleep, the system will continue to power Thunderbolt storage device(s), and resumes from sleep just fine. This also doesn't happen on the Z490 Vision D either.

However, this problem with hot plug on the Z590 Vision doesn't happen to USB devices. If you plug the same Windows disk that I described above into an available USB port on the Z590 Vision D (instead of a Thunderbolt port), Windows can go to sleep and reawaken just fine...the media does not disconnect. So something weird is going on with Thunderbolt hot plug on the Gigabyte Z590 Vision series. Non-functioning hot plug kinda negates the whole point of Thunderbolt.

 

[ori69]

It appears that Gigabyte's current Thunderbolt 4 implementation on their early Z590 BIOS may be broken as it pertains to Thunderbolt hot plug. I don't know if it's the BIOS or if it's the controller firmware that's the problem.

Outside of Windows 10, once a Thunderbolt device is disconnected from the Z590 Vision D, it will never reconnect until the system is cycled off and on. This is what is happening in Linux/macOS on the Z590 Vision D, but it doesn't happen on the MSI Z590 Ace, and, apparently according to this thread, the Asus Hero XIII. The only issue I have with the Asus Hero is that it doesn't appear to have a DisplayPort input, unlike the Vision D and the Ace.

Somehow on the Gigabyte Z590 Vision D, Thunderbolt hot plug does work in Windows, but not in the sense that we have come to understand. On the Z590 Vision D, when Windows goes to sleep and awakens, Thunderbolt devices lose power for a few milliseconds, and then Windows reengages the bus, and the driver hot plugs the device(s). This isn't a problem if your Thunderbolt device is a network card .or audio, but if your device is media, it will disconnect, which can be catastrophic.

So for example, if you boot windows on the Z590 Vision D from a USB device plugged into a Thunderbolt dock for example, upon sleep, you cannot successfully reawaken because for some unknown reason, the media disconnects before Windows has a chance to reengage the disk. The system will hang and then reboot. This doesn't happen on the MSI Z590 Ace... when it goes down to sleep, the system will continue to power Thunderbolt storage device(s), and resumes from sleep just fine. This also doesn't happen on the Z490 Vision D either.

However, this problem with hot plug on the Z590 Vision doesn't happen to USB devices. If you plug the same Windows disk that I described above into an available USB port on the Z590 Vision D (instead of a Thunderbolt port), Windows can go to sleep and reawaken just fine...the media does not disconnect. So something weird is going on with Thunderbolt hot plug on the Gigabyte Z590 Vision series. Non-functioning hot plug kinda negates the whole point of Thunderbolt.

The MSI MEG Z590I UNIFY - Mini-ITX is cool. Have you tried to put macOS on MSI 590 motherboards?

 

[dehjomz]

The MSI MEG Z590I UNIFY - Mini-ITX is cool. Have you tried to put macOS on MSI 590 motherboards?

I tried the Zz590 ACE, it was cool. macOS worked just fine. So did Linux.

Unlike the Vision D, MSI apparently hasn’t connected the iGPU to the Thunderbolt ports. So, if you want to use video via Thunderbolt, you have to connect a GPU and route DisplayPort from your card into the DisplayPort in. Weird.

It was very stable, no problems at all to use this board as a daily driver. But, it consumed much more power during heavy AVX workloads like Cinebench. +50 watts under Cinebench at 5.1 GHz. How is the same CPU consuming 50 more watts in the MSI board to get the same score? I lowered the voltages as far as they could stably go and set the load line calibration to the minimum value needed for stable performance at 5.1 GHz all core. If some YouTuber was reviewing Comet Lake in this board, they’d swear it was a power hungry beast, consuming over 200 watts at 5.1 GHz in Cinebench. But, if they put the same CPU into the Z490 Vvision D, it would consume literally 50 less watts for the same stable overclock, while apparently using a higher voltage. It’s crazy. But even with the 50 additional watts of power consumption, it didn’t overheat at all which was impressive.

Also, MSI’s UEFI click BIOS 5 isn’t as nice looking as the Gigabyte Vision, and the early release BIOS GUI was laggy. The early BIOS was kinda quirky, too. Sometimes it would reset itself, and I’d lose my XMP and overclock settings. To me, the only good thing going for it was the 4 M.2 slots and two extra USB2 ports as compared to the Z590 Vision (which has no external USB 2 ports anymore, just 2 internal ones).

But I had a very hard time installing MSI Mystic Light (RGB), the installer downloaded all kinds of bloatware. I literally had to install an old version first then upgrade to the new version for Mystic Light to appear. No documentation to help me.

Overall, I just couldn’t justify the cost, I don’t know what I gained, so I packed it up and shipped it out.

 

[ori69]

Hi @CaseySJ


I tried @ori69 EFI and had some problems with the USB and my TB equipment didnt show up.
Weird since ori has it working on his system.

I managed to make my own USB Patch,and now thunderbolt shows as expected.(No Hotplug yet,I tried the Mapple Ridge patch from you)
Here some files i extracted IOReg/Macaisl and Hackintool PCI
I also attached my EFI. just to check if al is good. and proceed from here.

I use a RX 580,i didnt do anything for igpu patching.

Thanks.

For me, your EFI is not working. Kernel panic crashes.
Once I was able to enter, the USB did not work.
In my config.plist XhciPortLimit has to be changed to YES, because without it, USB goes crazy.

What does the Thunderbolt system report look like on your computer?

 

[CaseySJ]

I tried the Zz590 ACE, it was cool. macOS worked just fine. So did Linux.

Unlike the Vision D, MSI apparently hasn’t connected the iGPU to the Thunderbolt ports. So, if you want to use video via Thunderbolt, you have to connect a GPU and route DisplayPort from your card into the DisplayPort in. Weird.

It was very stable, no problems at all to use this board as a daily driver. But, it consumed much more power during heavy AVX workloads like Cinebench. +50 watts under Cinebench at 5.1 GHz. How is the same CPU consuming 50 more watts in the MSI board to get the same score? I lowered the voltages as far as they could stably go and set the load line calibration to the minimum value needed for stable performance at 5.1 GHz all core. If some YouTuber was reviewing Comet Lake in this board, they’d swear it was a power hungry beast, consuming over 200 watts at 5.1 GHz in Cinebench. But, if they put the same CPU into the Z490 Vvision D, it would consume literally 50 less watts for the same stable overclock, while apparently using a higher voltage. It’s crazy. But even with the 50 additional watts of power consumption, it didn’t overheat at all which was impressive.

Also, MSI’s UEFI click BIOS 5 isn’t as nice looking as the Gigabyte Vision, and the early release BIOS GUI was laggy. The early BIOS was kinda quirky, too. Sometimes it would reset itself, and I’d lose my XMP and overclock settings. To me, the only good thing going for it was the 4 M.2 slots and two extra USB2 ports as compared to the Z590 Vision (which has no external USB 2 ports anymore, just 2 internal ones).

But I had a very hard time installing MSI Mystic Light (RGB), the installer downloaded all kinds of bloatware. I literally had to install an old version first then upgrade to the new version for Mystic Light to appear. No documentation to help me.

Overall, I just couldn’t justify the cost, I don’t know what I gained, so I packed it up and shipped it out.

It seemed the MSI Z590 Meg Ace was an ideal Z590 platform for content creators, but with power consumption being so high, I'm disappointed. But maybe all is not lost. Intel's Rocket Lake 11700 and 11900 (not the K variant) are both 65W parts so maybe power consumption will be acceptable?

 

[dehjomz]

It seemed the MSI Z590 Meg Ace was an ideal Z590 platform for content creators, but with power consumption being so high, I'm disappointed. But maybe all is not lost. Intel's Rocket Lake 11700 and 11900 (not the K variant) are both 65W parts so maybe power consumption will be acceptable?

Hi Casey,

Let me caution that I didn’t test power consumption at “stock” so it could very well be that out of the box, the power consumption figures aren’t too bad, I was testing overclocked and under load, curious to see if the CPU would overheat. I was just curious, and not expecting to see a +50W figure. Even with the extra power consumption, the CPU ran slightly cooler than it did in the Z490 Vision D. But that may be because I used more thermally conductive thermal paste in the MSI than I did with the Vision D. With a future BIOS upgrade, maybe they’ll get the power consumption in check.

The board isn’t a bad board at all, very solidly built. But is it worth $500, given that it likely won’t be able to run Alder Lake? Also, creators that use other boards (with less NMVe drives) can just buy a PCIe-to-NMVe card if they need more NMVe storage, and don’t want to sacrifice SATA ports or lose PCIe lanes in their bottom PCH PCIe port.

Also, “65W” on the non-K Intel CPUs should, in my opinion, basically just be labeled as “overly optimistic” marketing at this point because power consumption will increase under load, depending on how the motherboard implements turbo duration limits. Some motherboards don’t respect turbo duration limits at all, and essentially run in turbo mode all the time. Once the CPU is in “turbo” mode above its base clock, it will consume much more power than the 65W. And if the motherboard removes turbo duration limits altogether, power consumption will go thru the roof under load.

Anandtech has written about it here: https://www.anandtech.com/show/1634...e-i710700k-review-is-65w-comet-lake-an-option

 

[joevt]

-It seems Linux sees it as USB, check screenshot.

Did you install linux kernel 5.11 or later?
https://linuxhint.com/update_ubuntu_kernel_20_04/

Maple Ridge support is added here:
https://github.com/torvalds/linux/commit/db0746e3399ee87ee5f957880811da16faa89fb8
https://github.com/torvalds/linux/commit/5a8e3229ac27956bdcc25b2709e5d196d109a27a
It says it was added to 5.11-rc1 so 5.11 is the minimum released version.

The commit note says it's a USB4 controller with a firmware connection manager like previous Intel Thunderbolt controllers.

 

[CaseySJ]

Hi Casey,

Let me caution that I didn’t test power consumption at “stock” so it could very well be that out of the box, the power consumption figures aren’t too bad, I was testing overclocked and under load, curious to see if the CPU would overheat. I was just curious, and not expecting to see a +50W figure. Even with the extra power consumption, the CPU ran slightly cooler than it did in the Z490 Vision D. But that may be because I used more thermally conductive thermal paste in the MSI than I did with the Vision D. With a future BIOS upgrade, maybe they’ll get the power consumption in check.

The board isn’t a bad board at all, very solidly built. But is it worth $500, given that it likely won’t be able to run Alder Lake? Also, creators that use other boards (with less NMVe drives) can just buy a PCIe-to-NMVe card if they need more NMVe storage, and don’t want to sacrifice SATA ports or lose PCIe lanes in their bottom PCH PCIe port.

Also, “65W” on the non-K Intel CPUs should, in my opinion, basically just be labeled as “overly optimistic” marketing at this point because power consumption will increase under load, depending on how the motherboard implements turbo duration limits. Some motherboards don’t respect turbo duration limits at all, and essentially run in turbo mode all the time. Once the CPU is in “turbo” mode above its base clock, it will consume much more power than the 65W. And if the motherboard removes turbo duration limits altogether, power consumption will go thru the roof under load.

Anandtech has written about it here: https://www.anandtech.com/show/1634...e-i710700k-review-is-65w-comet-lake-an-option

Quite right. From the AnandTech article:

The TDP will be the maximum power under a sustained workload for which the base frequency is the minimum frequency guarantee.

	i7-11700	i7-11700K
Base Frequency	2.5 GHz	3.6 GHz
TDP (Thermal Design Point)	65W	125W

So TDP is a misleading comparison because base clocks are not the same. Clock-normalized power consumption may well be the same.

 

[dehjomz]

Quite right. From the AnandTech article:

	i7-11700	i7-11700K
Base Frequency	2.5 GHz	3.6 GHz
TDP (Thermal Design Point)	65W	125W

So TDP is a misleading comparison because base clocks are not the same. Clock-normalized power consumption may well be the same.

Exactly.

I never knew that at all until I read the article. AMD on the other hand has a power consumption close to what they advertise.