[SUCCESS] Gigabyte Designare Z390 (Thunderbolt 3) + i7-9700K + AMD RX 580

[djx8605]

Thanks for the tip on NVRAM! I did find an entry to Windows boot manager and removed it.

That seems to have removed my issue with just the F12 Boot Menu. Choosing F12 and proceeding to the Mac drive results in a successful boot now.

So this has really narrowed things down for my board. Now it is clear, I cannot have my Windows drive connected to the motherboard. I converted my Windows drive to UEFI from within Windows so that Clover could boot it. Now that the motherboard sees it as a UEFI option, that seems to be what is altering things and giving me the allocations error. I tested this with AptioMemoryFix and OsxAptioFix3Drv. While the latter still allows me to boot with iGPU enabled, it throws the allocation error when my Windows drive is connected. The. interesting thing is before I converted it to UEFI last night, I could leave it attached, as long as I didn't select it in any way.

This has really helped eliminate some things. I'm still looking for that perfect scenario of seamless dual boot.

Any ideas?

OK! We are getting somewhere

I wonder if Windows adding boot entries is doing something... Can you try this:
Boot macOS, or some other OS except Windows (Linux, WinPE etc). You will then mount the windows EFI partition, then rename the file in /EFI/Microsoft/Boot/bootmgfw.efi to /EFI/Microsoft/Boot/bootmgfw-orig.efi. Clover will scan and find the file, and if you have made a custom entry for it, just alter the name to read like the renamed file.
You can also rename bootx64.efi in /EFI/Boot, this may or may not help. The actual bootloader for Windows is the bootmgfw.efi file, it is responsible for loading and executing the BCD store before handing control over to WinLoad.efi.

Years ago I tried looking for a way to just load Winload.efi directly, but it seems that bootmgfw.efi sets up certain variables and such, so it couldn't be done
I really hate Windows persistently adding itself to NVRAM's boot option, it's like it wants to be the only bootloader or the number 1... It's as if Microsoft isn't aware that there are other bootloaders that can chain to their's and boot Windows as well lol...

 

[CaseySJ]

Thanks Casey,

I tried again but had the same issue as before. The screen would not get past a certain line of code after a short while.

However, I Googled around and found someone who was experiencing a similar issue with a slightly different Mobo. Another person provided help by sending their EFI folder to see if it would work for them. I downloaded the attached EFI folder and copied to my EFI Partition on my boot drive. Not sure why or how, but it solved my issue of not being able unlock my integrated graphics and only having only 7MB of VRAM. After rebooting, I now have fully accelerated 1536 MB of VRAM and weird choppy glitches are gone.

Here's the post where I found the file https://www.tonymacx86.com/threads/gigabyte-z390-ud-only-7mb-uhd630-mojave-10-14-2.276854/

I am attaching the EFI folder that seemed to work for me.

Wondering if you can take a look to see what the difference is.

Thanks again for all your help and creating this killer guide.

Cheers!

I see two possible explanations:

• This EFI sets device-id to 0x3E9B. But we should not set device-id at all if running Mojave 10.14.5 or later.
• This EFI sets Inject Kexts to Yes, which means it is injecting the kexts from CLOVER/kexts/Other.
  • Do you have Lilu and WhateverGreen installed in your /Library/Extensions folder (and if so, do you remember running Kext Utility to rebuild the kernel cache)?

 

[CaseySJ]

Re:10.15.1
If you update, be sure to update WEG first!
If you don’t...you’ll be sorry.

*** WARNING: Before Updating to Catalina 10.15.1 ***​

As @rondark pointed out, it is in fact necessary to update WhateverGreen to version 1.3.4 or newer before installing Catalina 10.15.1. This means:

• Downloading WhateverGreen 1.3.4 from the acidanthera GitHub site.
• Copy the kext to:
  • CLOVER/kexts/Other
  • /Library/Extensions and then run Kext Utility to rebuild the kernel cache
  • Also disable or remove any AGDP patch from Kernel and Kext Patches.

Failure to perform these steps will result in an empty screen just before the login window would have appeared.

---

 

[heythisisdave]

Hey @CaseySJ - quick questions:

1. Any ideas on what may have caused the borked VirtualSMC install (based on those instructions)? I'm running Clover v 4920 (the one that came in the original MultiBeast) - do you think that may have something to do with it? Do you recommend upgrading to the new versions of Clover? ( I remember I had a nightmare of a time with 5xxx versions of Clover and getting the right memory fix).

2. What do you think about the F8 bios - is that recommended or still in research?

3. What are your thoughts on the 5700XT's running in Catalina? Are those drivers portable to Mojave or is that just a pipedream. I'm considering upgrading to Catalina just to run the PowerColor Red Devil but not sure how stable those drivers are in either OSX or Windows right now.

 

[hawkshawwd]

Hello,

Thanks for this thorough build. I just set up mine but can't seem to get Thunderbolt 3 to work. I'm trying to use UAD Apollo x8. I has worked with a previous hackintosh with Thunderbolt 2. Could someone help me troubleshoot this?


System: Gigabyte z370 Designare, Intel i9-9900k, HyperX Predator Black 64BG 3200MHz DDR4, Sapphire Radeon RX580 8GB, Silicon Power 2TB NVMe M.2 (system drive), macOS 10.14.6 Mojave.

 

[hussar]

Hey All,

My question concerns the CalDigit drive utility:

I have 2 CalGigit T4s running via TB on my system.

Once I install the CalDigit drive utility my drives work correctly, however I cannot see the drive status in the app.

On a normal Mac the square shows green for normal and red when there's an issue-
not on my Hackintosh...
And when I click on "Device List" there's nothing.

I'm wondering if anyone else has the same issue? and if they have some sort of workaround?

 

[CODYQX4]

Updated to 10.15.1 successfully.

Then updated to F8 BIOS and lost the ability to boot with AptioMemoryFix+slide=0.

I looked at my memmap and saw the number was reduced.

For whatever reason, clearing CMOS and setting the BIOS up exactly the way I had it after the flash (which resets all the settings in the first place) resolved it.

This does seem to hint at the theory there is something in CMOS that permanently effects the memmap.

[SUCCESS] Gigabyte Designare Z390 (Thunderbolt 3) + i7-9700K + AMD RX 580

Hello all, Just wanted to touch base and mention I have XFX 5700 XT working in 15.1 beta 3. I was able to do some exports in premiere pro that worked great. I've attached Geekbench 5 compute scores as well. Very promising for future build for the community! Really looking forward to a public...

www.tonymacx86.com

My experience with memmap (iGPU off in all cases).
1. Flash to F8 BIOS+set desired BIOS settings: 160D6 pages. Couldn't boot with AptioMemoryFix+slide=0.
2. Reset CMOS: 24573 pages. I don't think I rebooted or shutdown here and can't remember if I had set the BIOS settings yet.
3. Reboot after I reset CMOS (the first time, not a second reset): 1E2EE pages.
4. Reboot and then patch my MSR because I forgot to do so: 19269 pages. This is slightly lower but close to where I was on F7. On F7 it was 19580 pages.

The values in #4 seem to be persistent after awhile, as sometimes as in 2-3 I get a different value but it doesn't persist. On F7 if I enabled iGPU and lowered DVMT preallocated to 32MB I got to 16580 pages. I was able to boot like that briefly but after a few restarts lost the ability to do so. I have only ever shorted the CMOS on F8 in all my testing. With the default 64MB DVMT I get 14580 pages and have never successfully booted like that.

I don't see DVMT settings anywhere in F8 so I can't replicate that test. I would like to test and see if clearing the CMOS consistently gives a high enough memmap to boot with iGPU enabled for the first few reboots before the memmap settles.

Things I have found to have no effect on my memmap so far (I have not cleared the CMOS before and after but across multiple reboots, shutdowns, and unplugs I end up stabilizing at the same values).
1. CSM Support Enabled/Disabled (I have only very briefly tested this and have almost always had it disabled and I don't see why enabling it would give me more memory)
2. VT-d Enabled/Disabled
3. Onboard Wi-Fi Enabled/Disabled
4. Audio Controller Enabled/Disabled
5. PCH Onboard LAN Enabled/Disabled (Annoyingly we can only disable one of two NICs in BIOS)
6. Installing an Aquantia 10G Gaming NIC (this didn't lower my memmap at all on several checks on F7)

Things that definitely effect my memmap (and do so immediately).
1. iGPU Disabled/Enabled
2. DVMT Preallocated (setting seems to be removed in F8)

Again, I have only ever cleared CMOS once (yesterday) and I don't know if clearing and then setting specific settings in a specific way can lead to different values persistently (my experience clearing CMOS lead to higher than normal available pages for a check or two and then settling), but on F7 I got consistent values changing the things that I confirmed have an effect and then changing them back.

What I would like to know is if the difference between 16580 pages on F7 and 19269 pages on F8 is due to the BIOS change itself, or the CMOS reset I did on F8 making something stick. I only had my 10GB NIC for a week on F7 but had it from the beginning on F8. I also had to clear CMOS upon flashing F8 because doing fast flash (which isn't supposed to persist anything) left me with a lower than normal memmap.

Another change I made last night (nothing related to memmap) was to go back to iMac19,1 SMBIOS and add the shiki values in, which allow previews to work without iGPU enabled. I don't know if there are any negative effects of that, if Sidecar works like that (I don't have an iPad), or if that breaks DRM.

 

[heythisisdave]

Updated to 10.15.1 successfully.

Then updated to F8 BIOS and lost the ability to boot with AptioMemoryFix+slide=0.

I looked at my memmap and saw the number was reduced.

For whatever reason, clearing CMOS and setting the BIOS up exactly the way I had it after the flash (which resets all the settings in the first place) resolved it.

This does seem to hint at the theory there is something in CMOS that permanently effects the memmap.

[SUCCESS] Gigabyte Designare Z390 (Thunderbolt 3) + i7-9700K + AMD RX 580

Hello all, Just wanted to touch base and mention I have XFX 5700 XT working in 15.1 beta 3. I was able to do some exports in premiere pro that worked great. I've attached Geekbench 5 compute scores as well. Very promising for future build for the community! Really looking forward to a public...

www.tonymacx86.com

My experience with memmap (iGPU off in all cases).
1. Flash to F8 BIOS+set desired BIOS settings: 160D6 pages. Couldn't boot with AptioMemoryFix+slide=0.
2. Reset CMOS: 24573 pages. I don't think I rebooted or shutdown here and can't remember if I had set the BIOS settings yet.
3. Reboot after I reset CMOS (the first time, not a second reset): 1E2EE pages.
4. Reboot and then patch my MSR because I forgot to do so: 19269 pages. This is slightly lower but close to where I was on F7. On F7 it was 19580 pages.

The values in #4 seem to be persistent after awhile, as sometimes as in 2-3 I get a different value but it doesn't persist. On F7 if I enabled iGPU and lowered DVMT preallocated to 32MB I got to 16580 pages. I was able to boot like that briefly but after a few restarts lost the ability to do so. I have only ever shorted the CMOS on F8 in all my testing. With the default 64MB DVMT I get 14580 pages and have never successfully booted like that.

I don't see DVMT settings anywhere in F8 so I can't replicate that test. I would like to test and see if clearing the CMOS consistently gives a high enough memmap to boot with iGPU enabled for the first few reboots before the memmap settles.

Things I have found to have no effect on my memmap so far (I have not cleared the CMOS before and after but across multiple reboots, shutdowns, and unplugs I end up stabilizing at the same values).
1. CSM Support Enabled/Disabled (I have only very briefly tested this and have almost always had it disabled and I don't see why enabling it would give me more memory)
2. VT-d Enabled/Disabled
3. Onboard Wi-Fi Enabled/Disabled
4. Audio Controller Enabled/Disabled
5. PCH Onboard LAN Enabled/Disabled (Annoyingly we can only disable one of two NICs in BIOS)
6. Installing an Aquantia 10G Gaming NIC (this didn't lower my memmap at all on several checks on F7)

Things that definitely effect my memmap (and do so immediately).
1. iGPU Disabled/Enabled
2. DVMT Preallocated (setting seems to be removed in F8)

Again, I have only ever cleared CMOS once (yesterday) and I don't know if clearing and then setting specific settings in a specific way can lead to different values persistently (my experience clearing CMOS lead to higher than normal available pages for a check or two and then settling), but on F7 I got consistent values changing the things that I confirmed have an effect and then changing them back.

What I would like to know is if the difference between 16580 pages on F7 and 19269 pages on F8 is due to the BIOS change itself, or the CMOS reset I did on F8 making something stick. I only had my 10GB NIC for a week on F7 but had it from the beginning on F8. I also had to clear CMOS upon flashing F8 because doing fast flash (which isn't supposed to persist anything) left me with a lower than normal memmap.

Another change I made last night (nothing related to memmap) was to go back to iMac19,1 SMBIOS and add the shiki values in, which allow previews to work without iGPU enabled. I don't know if there are any negative effects of that, if Sidecar works like that (I don't have an iPad), or if that breaks DRM.

Hey @CODYQX4 - this is fascinating. How are you viewing the BIOS memory? Are you dumping it in OSX or viewing it somehow through the BIOS tools?

 

[CODYQX4]

Hey @CODYQX4 - this is fascinating. How are you viewing the BIOS memory? Are you dumping it in OSX or viewing it somehow through the BIOS tools?

I just run memmap -b from the UEFI Shell which is part of Clover. This dumps the relevant information.

For me, I can only get slide=0 working. From the memmap -b dump there are columns and the first one is Type. We want one that is "Available". I use the second Available row (third actual row because second row is Type=Reserved), which the start column should be 100000 (which for slide=0 is what we use). When I list my available pages, I'm referring to the # Pages column in this row.

You can boot this board with AptioMemoryFix+slide=0 if you have enough pages. I don't know where the line is drawn to where you can and can't boot, but having the iGPU on takes enough pages so as not to boot. When I was able to boot with iGPU enabled with reduced DVMT memory, I had 16580 pages. I imagine it gets really iffy in that range, as I could boot a few times that night and then couldn't at all the next day. At 19269-19580 pages available I have never encountered the runtime error with AptioMemoryFix+slide=0, but I can't get that many pages with iGPU on, which is why everyone using AptioMemoryFix+slide=0 on a Z390 Designare has turned their iGPU off.

Theoretically, if there was some other reliable, consistent way to free up some more pages that wouldn't cause a different set of issues or tradeoffs, we could boot with iGPU enabled.

 

[totototo]

Updated to 10.15.1 successfully.

Then updated to F8 BIOS and lost the ability to boot with AptioMemoryFix+slide=0.

I looked at my memmap and saw the number was reduced.

For whatever reason, clearing CMOS and setting the BIOS up exactly the way I had it after the flash (which resets all the settings in the first place) resolved it.

This does seem to hint at the theory there is something in CMOS that permanently effects the memmap.

[SUCCESS] Gigabyte Designare Z390 (Thunderbolt 3) + i7-9700K + AMD RX 580

Hello all, Just wanted to touch base and mention I have XFX 5700 XT working in 15.1 beta 3. I was able to do some exports in premiere pro that worked great. I've attached Geekbench 5 compute scores as well. Very promising for future build for the community! Really looking forward to a public...

www.tonymacx86.com

My experience with memmap (iGPU off in all cases).
1. Flash to F8 BIOS+set desired BIOS settings: 160D6 pages. Couldn't boot with AptioMemoryFix+slide=0.
2. Reset CMOS: 24573 pages. I don't think I rebooted or shutdown here and can't remember if I had set the BIOS settings yet.
3. Reboot after I reset CMOS (the first time, not a second reset): 1E2EE pages.
4. Reboot and then patch my MSR because I forgot to do so: 19269 pages. This is slightly lower but close to where I was on F7. On F7 it was 19580 pages.

The values in #4 seem to be persistent after awhile, as sometimes as in 2-3 I get a different value but it doesn't persist. On F7 if I enabled iGPU and lowered DVMT preallocated to 32MB I got to 16580 pages. I was able to boot like that briefly but after a few restarts lost the ability to do so. I have only ever shorted the CMOS on F8 in all my testing. With the default 64MB DVMT I get 14580 pages and have never successfully booted like that.

I don't see DVMT settings anywhere in F8 so I can't replicate that test. I would like to test and see if clearing the CMOS consistently gives a high enough memmap to boot with iGPU enabled for the first few reboots before the memmap settles.

Things I have found to have no effect on my memmap so far (I have not cleared the CMOS before and after but across multiple reboots, shutdowns, and unplugs I end up stabilizing at the same values).
1. CSM Support Enabled/Disabled (I have only very briefly tested this and have almost always had it disabled and I don't see why enabling it would give me more memory)
2. VT-d Enabled/Disabled
3. Onboard Wi-Fi Enabled/Disabled
4. Audio Controller Enabled/Disabled
5. PCH Onboard LAN Enabled/Disabled (Annoyingly we can only disable one of two NICs in BIOS)
6. Installing an Aquantia 10G Gaming NIC (this didn't lower my memmap at all on several checks on F7)

Things that definitely effect my memmap (and do so immediately).
1. iGPU Disabled/Enabled
2. DVMT Preallocated (setting seems to be removed in F8)

Again, I have only ever cleared CMOS once (yesterday) and I don't know if clearing and then setting specific settings in a specific way can lead to different values persistently (my experience clearing CMOS lead to higher than normal available pages for a check or two and then settling), but on F7 I got consistent values changing the things that I confirmed have an effect and then changing them back.

What I would like to know is if the difference between 16580 pages on F7 and 19269 pages on F8 is due to the BIOS change itself, or the CMOS reset I did on F8 making something stick. I only had my 10GB NIC for a week on F7 but had it from the beginning on F8. I also had to clear CMOS upon flashing F8 because doing fast flash (which isn't supposed to persist anything) left me with a lower than normal memmap.

Another change I made last night (nothing related to memmap) was to go back to iMac19,1 SMBIOS and add the shiki values in, which allow previews to work without iGPU enabled. I don't know if there are any negative effects of that, if Sidecar works like that (I don't have an iPad), or if that breaks DRM.

Thanks for this research @CODYQX4. Just wanted to add that after second boot DVMT settings show up in BIOS. I don't see them there right after flushing to F8.