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

[CaseySJ]

Test 3 works for me with the same conditions in my other tests. I might try to compile OcQuirks down to as few options as needed on my machine. I didn't need very many on OpenCore, although no combination on OpenCore let me boot with iGPU unless I reset NVRAM, and then eventually I'd end up not being able to boot with iGPU anyway.

Are you able to install this EFI setup on a USB drive and then boot an NVME drive, or does NVME fail to boot no matter where Clover is booted from?

I should clarify that in each case I’m still booting macOS Catalina from the NVMe SSD. Only location of EFI folder is being varied.

 

[svandive]

I'm still on 10.14.6 and I've never been able to get HEVC working with dGPU and 19,1. I doubt this is a Vega 64 vs VII issue (but then maybe it could be), but do you remember what you did to get the HEVC support or was it just always working?

Also, just a quick question, but I had to find this out for myself... Have you clicked that refresh button between opening/closing VideoProc? When playing with various configurations I found it wasn't automatic.

I do click the refresh after each open/close of the application. I know it caches the proviso test run. You have to hit the refresh to get it to tests the current setup.

And to what I did, I'm running the latest Whatevergreen, SMBIOS iMac19,1, iGPU disabled, made sure to remove everything from Devices->Properties for the iGPU in my config.plist, and I'm running AptioMemoryFix-64.efi for my memory driver.

Not being in the Graphic arts or video work industries I've not tested this setup with Final Cut Pro, I've relied on the test from "VideoProc" and the output from Hackintool "VDA Decoder" = Fully Supported as signs that my rig was using with the headless iGPU or the dGPU for video work offload.

Are their other tests I should run to verify my setup is truly using the dGPU for video processing work?

 

[CODYQX4]

I should clarify that in each case I’m still booting macOS Catalina from the NVMe SSD. Only location of EFI folder is being varied.

That's what it sounded like, it's just odd that you have that issue.

It boots perfectly fine from NVME as long as you boot Clover from a non NVME drive. Maybe you'll find a fix for that issue, though I can boot direct from NVME. If I couldn't I'd probably just dedicate one of the spare USB flash drives I've got sitting around. They're good for if I want to make dramatic changes to EFI and test the results.

 

[gundogg9999]

Hi guys, on start up of my machine I need to select the OS X partition, it used to automatically go to the right partition and unless I touched the keyboard would count down for a few seconds and start the OS X boot.

Any ideas how to fix this? Thanks a lot

 

[djx8605]

Oh the mistake was nothing to do with Bluetooth. My backup had become corrupted due to sleep not Ejecting my backup drive correctly, while testing the theory that Catalina had resolved the problem. It hadn’t so I set my memory speed back to 2666

Hey @kwalter, what memory are you using? Brand/native speed? Your post just reminded me that after the FWRuntimeServices/Quirks test, we should probably all test sleep/USB disks lol.
I've been running Catalina using Crucial memory 3200 (2 sticks, 16GB each), place 1 on each bank for dual channel performance. On Mojave 10.14.6 I would always get the dreaded disk not ejected properly whenever I put the system to sleep or it slept on its own with USB drives plugged in. However, with Catalina 10.15.1, the machine sleeps on its own, it does power napping, I can put it to sleep manually, and it always wakes up fine... I've done a fresh install a few times, and the behavior is still the same...
Am I the only one with this rare blessing?

I'm only asking because I'm curious to see if we can find what (if something else) could also be causing this issue, assuming that there are others besides me who are not having the disk eject issue with above 2666 speeds on their RAM... Could it be specific RAM? Could it be the way Clover detects the RAM? hmmm...

 

[CaseySJ]

Hi guys, on start up of my machine I need to select the OS X partition, it used to automatically go to the right partition and unless I touched the keyboard would count down for a few seconds and start the OS X boot.

Any ideas how to fix this? Thanks a lot

• In the Boot section you can select either LastBootedVolume or a specific volume.

 

[JimSalabim]

@djx8605 Some folks have stated that disk ejection bug has been fixed in Catalina. But on my machine (3600 MHz RAM) the bug is still there on Catalina. The rest of the sleep behaviour and power nap works fine here on Mojave and Catalina, too.

 

[CaseySJ]

Speculative digression...

If a new 16-inch MacBook Pro is released tomorrow, will it be accompanied by the new Mac Pro? Or if Mac Pro production is shifting to Austin, TX there may be a longer delay.

16-inch MacBook Pro Announcement Expected by Tomorrow

Apple plans to announce its widely rumored 16-inch MacBook Pro this week, according to a tweet from Bloomberg's Mark Gurman. In another tweet,...

www.macrumors.com

 

[digumo]

I do click the refresh after each open/close of the application. I know it caches the proviso test run. You have to hit the refresh to get it to tests the current setup.

And to what I did, I'm running the latest Whatevergreen, SMBIOS iMac19,1, iGPU disabled, made sure to remove everything from Devices->Properties for the iGPU in my config.plist, and I'm running AptioMemoryFix-64.efi for my memory driver.

Not being in the Graphic arts or video work industries I've not tested this setup with Final Cut Pro, I've relied on the test from "VideoProc" and the output from Hackintool "VDA Decoder" = Fully Supported as signs that my rig was using with the headless iGPU or the dGPU for video work offload.

Are their other tests I should run to verify my setup is truly using the dGPU for video processing work?

The best way I know of to determine when an app is using the iGPU is to watch the Intel Power Gadget output. You can see in this image that I've got spikes when playing a video in the timeline of FCP. You could test other apps by just watching what happens.

Right now I'm still doing testing using SMBIOS 19,1 and how much of a difference there is between using dGPU only vs. iGPU and dGPU. Obviously, you get the encoding/decoding benefits of hardware based operations using 19,1 with the iGPU but even with 1024MB/MAX on DVMT settings for the iGPU, things still don't seem "quite" as speedy as when I use dGPU only.

I still need to test switching to iMacPro1,1 with iGPU active as was suggested by JimSalabim a few pages up.

 

[digumo]

...we should probably all test sleep/USB disks lol.
I've been running Catalina using Crucial memory 3200 (2 sticks, 16GB each), place 1 on each bank for dual channel performance. On Mojave 10.14.6 I would always get the dreaded disk not ejected properly whenever...

I'm only asking because I'm curious to see if we can find what (if something else) could also be causing this issue, assuming that there are others besides me who are not having the disk eject issue with above 2666 speeds on their RAM... Could it be specific RAM? Could it be the way Clover detects the RAM? hmmm...

Even with FwRuntimeServices.efi and OcQuirks.efi running 10.14.6(18G1012) I'm still getting ejected USB disks after coming out of sleep.