[Guide] Installing macOS Mojave (10.14.2) on Intel NUCi5BEH using Clover UEFI + updating to Mojave (10.14.6) on post#2

[HolyPeyote]

I spent hours past days to get 96 khz audio working on the tb3 apollo interface. It simply stopped at 48khz 64 buffer before the sound got scrambled by poor processing. My usb Apogee duet did not have this problem And could easily do 96khz 128 buffer. After indexing I noticed that the performance got a bit better but still not able to do 96khz. This i found very unusual as the performance @48khz was really really good. After I disabled speedstep in the bios I was finally able to do 96 khz at 256 buffer but after next reboot again no 96khz.

Because speedstep disabling for at least once did fix it, feeling was that the cpu simply did not run high enough to get 96khz starting. After downloading inter power gadget I saw that the cpu indeed was fluctuating (so the sppedstep bios setting was simply ignored somehow after reboot). When i got into the test tab of power gadget and run test ‘all threads’ the 4 cores ran at full speed. That fixes the problem. I can even go as low as 96khz/64 buffer now which gives me a very nice 2.58 ms overall latency without dropouts.

So that was the situation and kind of workaround fix, run the test in intel power gadget, let it run and have great performance in ableton. Now i’m looking for a more neat solution. So my question is, is it somehow possible to let clover run the cpu a full speed without having to go to power gadget and let the test run to achieve this? The bios option worked for once but i can’t get it to work with any setting i tried again.

ps. The performance for music production is very good on this i5. To give an idea : it runs 30 3 voice arturia cs-80 instances 9 bar loops @96khz 64 buffer With an overall latency of 2.58ms which is simply amazing. It does so while keeping the cpu 72c. i ll put it an akasa silent case to finally have my first high performance 96khz silent daw. Very happy with it.

It’s the sum of all parts, change one thing and the outcome could be different. The apogee drivers manages to do 96 khz easily on low running cpu while the tb3 interface did not at all.

 

[Leesureone]

I spent hours past days to get 96 khz audio working on the tb3 apollo interface. It simply stopped at 48khz 64 buffer before the sound got scrambled by poor processing. My usb Apogee duet did not have this problem And could easily do 96khz 128 buffer. After indexing I noticed that the performance got a bit better but still not able to do 96khz. This i found very unusual as the performance @48khz was really really good. After I disabled speedstep in the bios I was finally able to do 96 khz at 256 buffer but after next reboot again no 96khz.

Because speedstep disabling for at least once did fix it, feeling was that the cpu simply did not run high enough to get 96khz starting. After downloading inter power gadget I saw that the cpu indeed was fluctuating (so the sppedstep bios setting was simply ignored somehow after reboot). When i got into the test tab of power gadget and run test ‘all threads’ the 4 cores ran at full speed. That fixes the problem. I can even go as low as 96khz/64 buffer now which gives me a very nice 2.58 ms overall latency without dropouts.

So that was the situation and kind of workaround fix, run the test in intel power gadget, let it run and have great performance in ableton. Now i’m looking for a more neat solution. So my question is, is it somehow possible to let clover run the cpu a full speed without having to go to power gadget and let the test run to achieve this? The bios option worked for once but i can’t get it to work with any setting i tried again.

ps. The performance for music production is very good on this i5. To give an idea : it runs 30 3 voice arturia cs-80 instances 9 bar loops @96khz 64 buffer With an overall latency of 2.58ms which is simply amazing. It does so while keeping the cpu 72c. i ll put it an akasa silent case to finally have my first high performance 96khz silent daw. Very happy with it.

It’s the sum of all parts, change one thing and the outcome could be different. The apogee drivers manages to do 96 khz easily on low running cpu while the tb3 interface did not at all.

CPU performance is typically bios driven, an example being overclocking or undervolting. If the native bios supports those kinds of options then you can use them to tune the cpu. I could be wrong but I don’t think there is an alternative way within Clover or OSX to modify cpu performance beyond what you have already discovered in intel power gadget.

 

[HolyPeyote]

The workaround will do. Two step 5 second manual action. It is weird though that it ignores the speedstep disable while it did the first time after disabling. Seems my bios is corrupt or something. It’s a wonky bios, i also had to hookit up to another monitor one time to get disappeared bios options visible again. Toke me some time to find that out and was almost repackaging for rma.

I read some about setting max speed in clover but it seems not to let it run at max all the time but more for kind of offsetting . Thanks for the confirmation.

 

[Leesureone]

The workaround will do. Two step 5 second manual action. It is weird though that it ignores the speedstep disable while it did the first time after disabling. Seems my bios is corrupt or something. It’s a wonky bios, i also had to hookit up to another monitor one time to get disappeared bios options visible again. Toke me some time to find that out and was almost repackaging for rma.

I read some about setting max speed in clover but it seems not to let it run at max all the time but more for kind of offsetting . Thanks for the confirmation.

I was thinking about this more and I'm not sure if it will help but you can try modifying the entry in the config.plist that enables native power management. The shot below is from Xcode but you can do it in Clover Configurator as well. Change the SSDT/ Generate/ PluginType to NO

 

[HolyPeyote]

@Leesureone thanks for the suggestion, i tried it but now it fluctuates all the way up and down and does not even play audio anymore over the interface.
Went back to your original config.plist

Also i made a mistake in instance count it’s 30 3 voice cs-80 @256 buffer. At the very low 2.58 ms latency 64 buffer it’s only 7 instances. Still quite good, never been able to achieve such a low latency on my other systems.

Btw that new nuc10i7 you were mentioning seems to quite outbenches the nuc8i7. It is almost as fast as the i7 hexa 16 inch new mbp.
But I read on this forum that it might take a while before it will be supported in hackintosh.

 

[Leesureone]

@Leesureone thanks for the suggestion, i tried it but now it fluctuates all the way up and down and does not even play audio anymore over the interface.
Went back to your original config.plist

Also i made a mistake in instance count it’s 30 3 voice cs-80 @256 buffer. At the very low 2.58 ms latency 64 buffer it’s only 7 instances. Still quite good, never been able to achieve such a low latency on my other systems.

Btw that new nuc10i7 you were mentioning seems to quite outbenches the nuc8i7. It is almost as fast as the i7 hexa 16 inch new mbp.
But I read on this forum that it might take a while before it will be supported in hackintosh.

That's great you're getting that level of performance even if it takes a work around. I'm probably going to be one of the early adopters of the NUC10 model, had one ordered but intel hasn't filled orders yet. Soonest I've heard they'd be available will be Early February now.

 

[HolyPeyote]

Thanks for your help.
kext added, but the microSD drive made a very strange noise (like Floppy drives in the old days...) and the microSD card wasn't recognized.
I'll make another try when I get back home with another microSD card.

That sound is called coil whine. It is caused by variations on CPU load (the mac mini 2018 suffers from it too) Unfortunately I just updated my bios from 0074 to 0077 ( it supposed to have a power management fix) and it now suffers from coil whine quite a bit (nice power management fix ..... ) Due to the new ME version it can’t be downgraded back to 0074
The coil whine was definately not there on 0074.
The intel power gadget all thread test fixes this coil whine though and I have to use that anyway to have 96khz working on my tb3 interface. But it really sucks that a bios update induced coil whine, working on normal power management mode is impossible for me anymore as i am really allergic to this modulating high pitch noises.

So warning for those who are not bios 0076 yet (ME firmware version change) and don’t want coil whine: do not update past 0075 because you can’t go back!

 

[temuryusuf]

Brilliant guide! Everything works perfectly - audio was crackly over 3.5mm but was fixed with AppleALC.kext
I still can't get sleep working properly for the life of me though. The machine goes to sleep fine, wakes up to the password screen fine, but maybe 3 seconds after putting in the password and getting to desktop, it hangs and restarts automatically. What could possibly be the issue?

 

[Leesureone]

Brilliant guide! Everything works perfectly - audio was crackly over 3.5mm but was fixed with AppleALC.kext
I still can't get sleep working properly for the life of me though. The machine goes to sleep fine, wakes up to the password screen fine, but maybe 3 seconds after putting in the password and getting to desktop, it hangs and restarts automatically. What could possibly be the issue?

Please update your profile to show what hardware/ NUC model you are using. It makes it infinitely easier to support you by narrowing down the platform. Sleep problems are typicalyl a function of the USB port limit patch not being correctly implemented. Can you confirm you are using my EFI folder and what the contents of your EFI/ACPI/patched folder is?

Sleep issues can also be caused by devices connected to your device, external audio devices for example or KVM Switches. If that is true try unplugging eventing except your mouse and keyboard and try again.

 

[jamikus]

Needing to reboot does not sound like the screen saver is related. One thing to check is if you applied the bios settings correctly per the first post. USB devices can also cause sleep problems, your keyboard and mouse shouldn't be a problem but if you have anything else plugged in unplug it and then test what happens. Manually tell it to sleep. You can also try a different keyboard mouse set up.
The graphic issues can be a number of issues but it can be related to the cable or the monitor itself, possibly the adapter but you'd have to try a different setup to test. The USBc/ Thunderbolt jack can be used for display as well unless you had other plans for it.

Ok, I during last days I tried couple of things without success, but this is important:

2018 Mac Mini Static after sleep awakening

I'm having the static/snow screen display issue after awakening from sleep on my 2018 Mac mini. By static/snow, I mean the monitor display looks like the pattern an old TV using an antenna would display if it lost its signal. I returned my first 2018 mac mini because of this issue and the...

forums.macrumors.com

Basically it shows this problem is essential to current Apples, not only hackintoshes.
It exhibits as "snowing" on some LCDs or 20 sec delay on others, after long sleep.

Only workaround available which is only working in some cases seems to be to buy a display port connection (LCD) or to pull cables out and in each time.

Its quite shocking considering Apple is apparently not abble to address this issue for whole last year. It would be funny if hackintosh community was actually able to solve it