Z490 & Z590 - Will Z590 ever have macOS Support ?

[CODYQX4]

Try enabling wake from Thunderbolt devices and native os security, and disable Thunderbolt USB support. According to Gigabyte 'Thunderbolt USB support' is buggy and is for the preboot environment.

Gave that a try, no change. Hoping putting a GPU in will change things. Not having a GPU makes a lot of things buggy. GT 710 and a USB 3 Type A to Type C adapter should arrive tomorrow for more testing.

 

[c-o-pr]

Followup on my previous headache re a Asus BIOS upgrade and Sabrent Rocket 4 NVMe death at same time.

After working through all my config and arranging another backup I updated to latest Asus Z590 Hero BIOS, 0707, running a EVO 970 Plus with no problemo.

The Sabrent death was just a coincidence, infant mortality. Engaged Sabrent customer support directly they do not seem enthusiastic. But we plodded through their RMA process, and they inspected the drive are sending a replacement and covering shipping.

Re comments on overlocking this Asus, I've never done this before, but I give this board high marks for both making it easy to get a stable OC within Intel's limits, and to also expose every possible detail of timing and power config to if you want to stress it to the max and eke out the last 3% you can do it.

Lesson learned, i9-10900 OC and binning is so well understood that a stable 4.9 OC is a sure thing, and the "AI" system for OC leaves just a smidge on the table to ensure that voltage/power are within healthy limits for long term use. Trying to push a bit more is very easy, and exposes an obvious power pitfall. These Intel parts come properly maxed out of box.

The G-Skil Ripjaws is worked flawless at its promised 3600 timings under XMP II and lets me push freq a tad more to 3800 with no complaint.

OTOH, pushing CPU too hard by disabling Intel's limits is quite practical and the board auto-mode will let e take it right to edge where power consumption soars and it eventually conks.

My feeling is that everything about this is so well baked it's boring, unless like me you have no experience so it's sort of neat to see for myself.

End result is a proper combo of thermals and power yields a reliable 5GHz multi-core at 1.35–1.4V, low-80s and runs a non-AVX full load all day with no glitches.

GB5 of 1330 / 11175
Cb.R23 of 15500

The BIOS is set to endlessly retrain OC so it seeks and finds a great combo of limits while ambient temp can go wherever. I'm impressed.

And compared to old Mac Pro 3,1 this thing is lightening.

When new Sabrent comes in I'll cross my fingers and try again, then move my world over from old Mac Pro and look back fondly at a very reliable Apple unit that's not only served me for 13 years, taken upgrades with ease, and with just one exception of Parallels VM, is still a useful machine today.

One last thought: As I was working on OC, I wanted to completely start over on OC / cooler training to I reset the CMOS and when it came up my EVO was not found. I was "oh crapola!" It took me the longest 10 mins I've ever spend grinding through 1000s OS BIOS options looking for the one that sets PCIe bifurcation for the M_2 NVMe slot when using last-gen CPU, and, when I found it Phew! — It had been reset under assumption of use of primary NVMe slot (which is PCIe 4 only so this 10900K doesn't support). As Orson Welles once said,"Too Much Johnson."

 

[xtreem.p]

@dehjomz, do you still have the TB3-hotplug.aml for the Z590 Vision D by any chance?

I'm giving this some extra time to convince me to keep it

 

[dehjomz]

@dehjomz, do you still have the TB3-hotplug.aml for the Z590 Vision D by any chance?

I'm giving this some extra time to convince me to keep it

View attachment 518504

Seems that you have Thunderbolt working.

Here you go my friend. Nothing special, I just used the Asus Z590 Hero 13 TBT-hotplug as a guide, and modified it slightly. The Maple Rridge NHI is on Root Port 5 on both the Hero and the Vision D. Hopefully, it's helpful to you.

Only thing I will note is that hot-plugging in macOS doesn't appear to be working at the moment. It is working in Windows and Linux, however. Perhaps the SSDT needs more modifications?

When I hot-plug a Thunderbolt device in macOS, I've noticed that, even though it is not recognized, with the SSDT applied, when I suspend into s3 sleep and when the system resumes from sleep, the formerly unrecognized Thunderbolt device is then recognized.

Seems that Gigabyte has some ACPI code that is executed during resume from S3 sleep that probes the bus, and then notifies that OS kernel that a Thunderbolt device has attached. If only we could find out what that code is, and somehow run it when Thunderbolt devices are hot-plugged... or maybe macOS needs an updated Thunderbolt driver for Maple Ridge? I don't know. All I know is, it's weird that an unrecognized device is then recognized after entering into and resuming from s3 sleep.

As quirky as the board is, I'm keeping it. Other than the Thunderbolt hot-plug thing in macOS, it works stable for me. Overclocks Rocket Lake, 33 degrees at idle, not much to complain about. For stability, I've enabled a per-core overclock. And I disabled TVB frequency clipping to ensure the board holds the 8-core turbo ratio under sustained load, instead of clocking down to 4.6 GHz. Gigabyte is working behind the scenes to rectify the various issues, there is an updated Maple Ridge NVM, as well as unreleased BIOSs that seem to improve hot-plugging support in Windows/Linux. Hopefully, they push BIOS updates soon.

There still is an issue of hot-plugging not working for certain JHL6240 devices, but all the others (JHL6340,JHL6540, JHL7540,JHL8440) seem to hot-plug fine in Windows/Linux.

 

[xtreem.p]

@dehjomz Hey! Thanks buddy!

This is weird since I don't seem to have RP05 here

Any words of wisdom?

 

[CODYQX4]

@dehjomz Hey thanks buddy!

This is weird since im dont have RP05 here

Any word of wisdom?

View attachment 518519

I'm going to try the SSDT he just posted to see if anything changes. I also haven't been able to get RP05 to show.

I only have a Thunderbolt monitor, and probably can't drive it since I had to put a GT 710 in and disable the iGPU entirely, and that GPU doesn't have a DisplayPort to use.

One thing I noticed, is despite the SSDT I am using not working (HackinDROM Asus ROG Z590 Maximus XIII Hero except I changed GC-Titan Ridge to GC-Maple Ridge as the device name), is that I can plug/hotplug USB-C 3.0 devices in.

I wouldn't have expected this to work. As @CaseySJ found on Z390 DESIGNARE, the later BIOS that adds Resizable BAR breaks SSP1/SSP2, which are USB 3 devices connected to the thunderbolt port. Despite the fact that I don't have SSP1/2 on Z590 Vision D, plugging a USB 3 drive in using an Anker USB A to USB C adapter lets me hotplug the device in either Thunderbolt port. I can see them connect in IORegistry and the flash drive mounts in Finder as expected.

This is not Thunderbolt hotplug, but the USB capability works when I wouldn't expect it to in this case.

EDIT: Gave it a try, no luck. Seems to be minor cosmetic changes over the previous one.

I'm able to at least map the USB 3 USB-C ports with USBMap. I called them SST1 and SST2 to identify them as the thunderbolt ports, but distinguish them from the SSDT calling them SSP1 and SSP2.

 

[dehjomz]

Are you all also loading SSDT-DTPG?

 

[xtreem.p]

Yessir, so far no go

 

[CODYQX4]

Are you all also loading SSDT-DTPG?

Yes I'm loading that SSDT and earlier than TB3 SSDT if that matters like it does kexts.

 

[xtreem.p]

I tested more with the UAD stuff, and I get often communication UAD error -38. Plugins get disabled. I tested same setup with my Mac Mini M1, and there are no issues. I guess Gigabyte still need to tweak the TB firmware.

Still no hot-plug capabilities. I have to cold boot for the devices to be recognize. Warm boot doesnt work with TB devices. @dehjomz, this is the same experience you had with this mobo right?