Contribute
Register

[Success] ASRock Z390 Phantom Gaming-ITX + TB3 + iGPU + Mojave + SFF Build

See Post #2.


Thanks. As I'm using a supported nVidia, it should be fine, although the USB might be impacted. I'll hold off for a bit, as to have time to fix the issues if they arise.
 
Please take a video or take all screenshots from the RAM settings section. Apparently, my problem is that my DDR4 memory is too fast for a Mac, so the disks are removed incorrectly after getting out of sleep.
 
Please take a video or take all screenshots from the RAM settings section. Apparently, my problem is that my DDR4 memory is too fast for a Mac, so the disks are removed incorrectly after getting out of sleep.

So the disk sleep disconnecting issue will be solved if the XMP is turn off?

In that case, I heard from other threads that, you can try to turn on the XMP by clover instead of on BIOS.
 
So the disk sleep disconnecting issue will be solved if the XMP is turn off?

In that case, I heard from other threads that, you can try to turn on the XMP by clover instead of on BIOS.

I read a lot of information about the problem (Disk not ejected properly) on the Internet and spent a lot of time on it. Almost everyone managed to solve the problem by disabling XMP (Extreme Memory Profile) in Bios. So far, I haven’t managed to solve my problem, because I don’t know how to disable this technology in the BIOS of the ASRock Z390 Phantom Gaming-ITX motherboard. In the settings there are only items "auto", "profile 1" and "profile 2".

I know that my problem is definitely that my HyperX RAM has a frequency of 3466 MHz, although the maximum that the iMac 19.1 supports is 2666 MHz. All my automatic settings for the motherboard are more than 2800 MHz. I do not exclude that I will have to change my RAM to some other one.

ASRock Z390.jpg

ASRock Z390 2.jpg

ASRock Z390 1.jpg

ASRock Z390 2-1.jpg

ASRock Z390 3.jpg

ASRock Z390 4.jpg

ASRock Z390 5.jpg

ASRock Z390 6.jpg

ASRock Z390 7.jpg
 
Last edited:
Thank you so much @rj510 !! your guide is really well done, I have a very similar hardware and even though I never built an hackintosh I managed to make everything work.

Just two "small" questions, if you can help me.
1 the boot is quite slow (about 1'20") is it normal?
2 I have this TB3 Adapter and everything works (Ethernet,USB3 HUB,SD Card Slots),unless the HDMI Output. You think it is possible to make it work in some way?

thank youuu!! :clap:
 
Last edited:
There are many issues here. The slots and other descriptions (Clover/Arbitrary) I uploaded were for the iGPU. Once you add a graphics card, the later entries get bumped about, and this is what has led to the mis-labelled slots and cosmetic inaccuracies. I did not load correct values for a graphics card. I will have to create 2 sets of config.plists at a later date.

(Keep in mind that these are only descriptions and are cosmetic in nature, rarely affecting function.)

Until I can get new sets created, I would suggest that you disable the TB descriptions, allowing only the TB SSDT to be used for setting descriptions. I've attached here a v5 plist that removes, or disables, those entries above 1:00:0. The entries below 1:00:0 are not affected by adding a graphics card. The Broadcom card seems to get moved from 3c:00:0 (iGPU) to 3e:00:0 (with a graphics card).

The attached image showing the System Report is with an optimized Clover/Arbitrary for only iGPU (I'll supply another image optimized for Vega 56, which could be edited for any other graphics card).

If you then follow what I wrote about in the "Spoiler: Look: No TB SSDT" section, at the bottom, you can derive your own slot values. (Specifically using LSPCI and Hackintools.) This will then lead to correct entries in the "About this Mac" window. Below, I show images of my "About this Mac" window when running iGPU alone, or the iGPU with a Vega 56 (the latter does not use any special Clover/Arbitrary entries but it naturally read by Mojave 14.5 and 14.6).

I ran the Videoproc and it works just fine (see attached), showing the iGPU being activated along with the Vega 56. (I set in BIOS the PCI slot for the Vega 56 and kept the iGPU also active). And this was before I worked on correcting the slot descriptions in the Clover/Arbitrary section. I do not use NoVPAJpeg and have no idea if this is playing a role; I would suggest that you temporarily removing it to see it's influence on Videoproc.

On USB matters, attached is what is activated by my USBPorts kext as shown by Hackintools. (Unfortunately, I messed up and cut off the bottom which showed the presence of the SSPx ports, along with USR1.)

To follow up on the iGPU discussion:

I take your point @rj510 that measured performance is what matters—however I'm unable to actually measure that without removing the Vega (and would rather not because I haven't applied the framebuffer patch and hope to avoid it) and believe the evidence I do have suggests the iGPU isn't properly configured.
  1. In IOReg, all the IntelAccelerator entries appear as disconnected/red (see attached). This is with your Arbitrary iGPU device configuration enabled (and no other one in the Properties tab). [EDIT: this may just be because it's in headless mode?]
  2. UHD Graphics appears in iStat (but not Intel Power Gadget), but never has memory or utilization values other than 0%, despite rendering Youtube videos in Safari and/or hitting "refresh" in videoproc. I suspect that VideoProc is simply reporting that it can use the Vega to accelerate transcoding tasks. [EDIT: I now believe acceleration actually works, but iStat can't see the activity]
  3. Attempting to plug in a second monitor to the DisplayPort port on the motherboard (with the first remaining in the Vega) results in a total system lockup - unrecoverable. (So complete that unplugging a powered USB device (mouse with LED) and plugging it back it does not power it up again.) [EDIT: this is not ideal, but not unexpected (?) for headless mode—i.e. headless mode => the motherboard HDMI and DP ports aren't usable]
  4. LuxMark and GeekBench both do not allow the iGPU to be selected as a graphics accelerator to benchmark (see screenshots).
  5. EDIT: As I wrote in an earlier post, deleting the Device configuration in the Arbitrary section (1) doesn't affect availability of acceleration (videoproc is green) and (2) makes iGPU power and util viewable in Intel Power Gadget.
    1. In summary, I think it's best to delete the iGPU device config, since functionality/performance remains unchanged, and at least one program can actually monitor its use.​
    2. Note: in none of my tests was I able to get DRM video playback working in Safari (i.e. Netflix and some Youtube stuff). Still trying to get this working.​
    3. Note: this post in the Designare thread had a screenshot of UHD 630 activity recorded in iStat! So it must be possible! Any ideas on how?​
Another way to look at it is through benchmarks—I get 140183 OpenCL GeekBench compute score and 25813 on Luxmark/luxball. These seem... good, but not close to Windows (expected), and perhaps not what they should be if the iGPU were hypothetically contributing cycles? If you still have your Vega attached to the phantom gaming itx mobo, want to check what you get? (Or anyone else who wishes to try.)

Also, Sleep:

Have you observed Sleep and/or Power Nap/Darkwake to work correctly? I've had intermittant results. Sometimes it seems to work. Sometimes it sleeps at a higher power setting, so wakeup is faster (perhaps only the display is sleeping?), and sometimes it sleeps so... deeply that it cannot wake up, and requires a hard reboot to return to the desktop environment. So far I've been unable to isolate what combination of settings actually affects this, however.

Many thanks,
-jsil
 

Attachments

  • Screen Shot 2019-07-24 at 8.48.16 PM.png
    Screen Shot 2019-07-24 at 8.48.16 PM.png
    99.5 KB · Views: 69
  • Screen Shot 2019-07-24 at 8.49.24 PM.png
    Screen Shot 2019-07-24 at 8.49.24 PM.png
    43.8 KB · Views: 61
  • Screen Shot 2019-07-24 at 8.50.47 PM.png
    Screen Shot 2019-07-24 at 8.50.47 PM.png
    17.9 KB · Views: 63
  • Screen Shot 2019-07-24 at 8.55.27 PM.png
    Screen Shot 2019-07-24 at 8.55.27 PM.png
    49.8 KB · Views: 55
  • Screen Shot 2019-07-24 at 9.39.04 PM.png
    Screen Shot 2019-07-24 at 9.39.04 PM.png
    70.3 KB · Views: 59
Last edited:
To solve “disk not ejected properly” I tried to do literally everything:
– I reflashed the BIOS on 1.2, 1.3, 1.4 and 1.5
– Reinstalled the operating system
– Switched the frequency of RAM to 2666 MHz and 2133 MHz
– Removed Wi-Fi module
– Disconnected all keyboards, mice and other devices.
– I connected a hard disk using SATA
– Disconnected drive m.2 (Samsung 950 EVO Plus)
– Launched a cloned MacBook operating system
– I tried to use special scripts and programs.
– Turned off the water cooling
– Disconnected power buttons (and others) from the case.

Unfortunately, none of this helped. Most likely, the problem is in my RAM HyperX Fury 3466 MHz (HX434C19FBK2/32).

ASRock Z390 1.jpg

Apparently you will have to change it to Corsair Vengeance LPX DDR4 2666 MHz (CMK32GX4M2A2666C16)

Maybe someone knows that you can try again?
 
To solve “disk not ejected properly” I tried to do literally everything:
– I reflashed the BIOS on 1.2, 1.3, 1.4 and 1.5
– Reinstalled the operating system
– Switched the frequency of RAM to 2666 MHz and 2133 MHz
– Removed Wi-Fi module
– Disconnected all keyboards, mice and other devices.
– I connected a hard disk using SATA
– Disconnected drive m.2 (Samsung 950 EVO Plus)
– Launched a cloned MacBook operating system
– I tried to use special scripts and programs.
– Turned off the water cooling
– Disconnected power buttons (and others) from the case.

Unfortunately, none of this helped. Most likely, the problem is in my RAM HyperX Fury 3466 MHz (HX434C19FBK2/32).

View attachment 418493

Apparently you will have to change it to Corsair Vengeance LPX DDR4 2666 MHz (CMK32GX4M2A2666C16)

Maybe someone knows that you can try again?

Hence, this issue is RAM specific?
 
To solve “disk not ejected properly” I tried to do literally everything:
– I reflashed the BIOS on 1.2, 1.3, 1.4 and 1.5
– Reinstalled the operating system
– Switched the frequency of RAM to 2666 MHz and 2133 MHz
– Removed Wi-Fi module
– Disconnected all keyboards, mice and other devices.
– I connected a hard disk using SATA
– Disconnected drive m.2 (Samsung 950 EVO Plus)
– Launched a cloned MacBook operating system
– I tried to use special scripts and programs.
– Turned off the water cooling
– Disconnected power buttons (and others) from the case.

Unfortunately, none of this helped. Most likely, the problem is in my RAM HyperX Fury 3466 MHz (HX434C19FBK2/32).

View attachment 418493

Apparently you will have to change it to Corsair Vengeance LPX DDR4 2666 MHz (CMK32GX4M2A2666C16)

Maybe someone knows that you can try again?

Just a random thought: have you tried toggling some of the CPU sleep states in the BIOS? Which of C7 and C10 power states do you have enabled? It may be that macOS is sending the board/cpu into a power state you haven't enabled, or if you disable a deeper power saving state it could force it to stay at a higher power, and hence not try to power off the disks/RAM/whatever is the root cause of your issue. (I also don't know if disabling C7 might make things worse!)
 
Just a random thought: have you tried toggling some of the CPU sleep states in the BIOS? Which of C7 and C10 power states do you have enabled? It may be that macOS is sending the board/cpu into a power state you haven't enabled, or if you disable a deeper power saving state it could force it to stay at a higher power, and hence not try to power off the disks/RAM/whatever is the root cause of your issue. (I also don't know if disabling C7 might make things worse!)
May I know if you are also suffering the same sleep disconnecting issue in acrock 390 itx?

If not, can you share the details of your RAM? I may want to purchase the same one to avoid it. Thanks.
 
Back
Top