I seem to learn the hard way, and found that I was creating problems by
not following your advice precisely. I'm sorry if I caused any confusion here.
No problem, I always try to be as relevant as possible but recently so many things have been changing in the Hackintosher scene where it has been hard to keep up the guides updated with the recent changes.
With regard to AptioMemoryFix, I tried substituting OsxOptioFix3Drive, which
you felt was a mistake. Although I never ran into any obvious problems after
I booted, I noticed that it would incur a few second delay in the initial boot,
and once in a blue moon my Clover would fail to find the macOS disk, and I'd
have to restart. When I went back to AptioMemoryFix, I didn't have these problems.
The reason why AptioMemoryFix is not available on Clover Installer Packages anymore is because the
AptioFixPkg repository has been archived as they move to FwRuntimeServices.efi which these new drivers are now available on AppleSupportPkg which is pretty much OpenCore only, we have to download the drivers manually from AptioFixPkg and add the AptioMemoryFix again until Clover team updates Clover to use FwRuntimeServices or get a workaround for that.
Keep in mind AptioMemoryFix still works fine it's just that it will not be updated anymore which means it may break on a future update or something similar.
I had a similar experience with VirtualSMC. I noticed that you switched to VirtualSMC
in your Catalina guide, and decided to give it a try. Although it seemed to work fine,
I noticed syslog messages that it could not find a compatible chip for the SMCSuperIO.kext.
The larger problem was that the 10.14.6 updates incurred 4 stalls of four minutes each
in ACPICPU. When I went back to FakeSMC, these stalls went away.
I did notice that, the problem with VirtualSMC is that it doesn't have a proper implementation for reading non Apple SMC keys which explains why many sensors like fans etc are not showing up on HWMonitor and other monitoring apps.
All sensors that were available on FakeSMC are available into VirtualSMC & SMCSuperIO but like I said they have not implemented a way to read the non SMC keys yet.
So it may be fine to stay on FakeSMC and the plugins for a bit longer, although I have no issues with VirtualSMC
Yes I do have that message and I removed the SMCSuperIO.kext and left VirtualSMC.kext and SMCSuperIO.kext for now.
I write to give feedback about using your methods with an i5-3470 with HD2500
graphics and to ask some questions. The questions involve optimizing IGPU/DGPU
performance with an i5 8300 and either an Nvidia 730 2 GB Kepler DGPU or an
MSI RX560 LP 4 GB DGPU.
With the Nvidia DGPU, it seems the best combo is with the IGPU enabled. With the
IGPU enabled in BIOS, I found that setting the DGPU as the primary in BIOS would
cause clover/Lilu/WEG to automatically inject a 0x01620007 headless IGPU framebuffer
if undefined in device/properties. This seems to be the recommended method in
the current WEG documentation. The IGPU appears in the Clover options Graphics
menu, but not in About this Mac - System Report. The current WEG documentation
also seems to recommend an iMacX,Y definition in SMBIOS for the HD2500 graphics.
If I wanted to inject a 0x0166000A or 0x01620005 as recommended for a desktop
computer with a MacMini SMBIOS, I had to set the IGPU as the primary in BIOS
or it would reboot halfway through the macOS boot, and define the desired framebuffer
with device/properties. The IGPU appears as a separate device in About this Mac - System Report
under Graphics/Displays. If a slotname is also supplied, the IGPU appears in About this Mac
PCI devices.
I also experimented with injecting a 0x01660003 or 0x01660004 as recommended for
a mobile computer with a SMBIOS MacBookPro9,2. I can measure performance for all the
connector-full framebuffer definitions using Geekbench4, which shows the IGPU as a
separate compute device. The IGPU always gets a metal score of zero, as it fails to
pass the Particle Physics test, but it passes all the other tests.
I've also tried disabling the GPU in BIOS and using the RX560 as a standalone DGPU.
I basically get the best overall performance in this mode, and see a degradation of
the Siri startup / graphics with any of the connector-full IGPU solutions.
Yes that is correct, I will add a note into the guide for i5 models or make a config_i5.plist file in the repo with this headless Framebuffer so it enables Hardware Encoding/Decoding, OpenCL QuickLook and other features.
I have some testing ideas about USB ports and will soon test if that works to override the PRTx ports being injected on iMac13,x SMBIOS.
When I'm ready with the files for testing, I will quote you and upload the files for testing so you can give them a try and see if solves your problem/improve the situation.
But I have some problems with Safari and iTunes streaming video, and am curious
if the other users of i5-3470 or i7-3770 systems have the same problems. When
I run Safari or iTunes streaming video on my real Apple MacMini with an i5-3210M
with HD4000 graphics, I can play Hulu video on Safari or stream iTunes video directly from
the Apple store.
On my hackintoshes, I can only play Hulu video successfully on Firefox or Chrome browsers.
On Safari Hulu, I can typically play the network intro, and maybe a commercial, but it fails when
the body of the program starts. Depending on the configuration, it will sometimes fail to
start streaming at all, or it will stream with a red screen and I can hear the audio play.
On my hackintoshes, I can download iTunes TV shows and movies and play them after the
download finishes, but I cannot stream a video prior to download. Sometimes I get a black
screen with no action at all, or I get a red or green with time indicators ticking down, which
sometimes flashes artifacts.
As far as I know this is an issue with improper IGPU setup which we discussed above, see if the problem get's solved when you test the new files.
There's another topic here dedicated to using the AMD graphics cards as standalone GPUs. It seems
to indicate that since 10.14.6, the only solution for H.264, HVEC, and DRM operability is to
define the SMBIOS as an iMacPro1,1 with IGPU off in BIOS. I've tried this, but don't see a big
difference in the iTunes, Safari problems. I can verify in Hackintool and VideoProc that H.264,
HVEC, and VDA decoder are fully supported.
I also tried the various configurations with macOS Catalina AppleTV, and see the similar problem of
having to download a program first in order to play the video.
My first question is: Do others here see the same problems in streaming videos in iTunes and/or
playing Hulu videos in Safari?
My second question is: Is it possible to modify the hotpatches to be compatible with other models?
My understanding is the the USB in particular is more specific to the MacMini model.
Thanks for any feedback you can provide.
Yes, the reason to pick that SMBIOS is that it's the only one without IGPU if I'm not mistaken and that's why they use it by default.
Like I mentioned above, this comes due to not having a correctly configured IGPU.
As far as I know I don't have any problem with DRM content in iTunes etc, I believe I did try a while ago and it was working fine because HD4000 is natively supported.
Yes it's on my To-Do list to add better compatibility and will soon do the changes and cover everything what is missing.
First: thanks for covering these problems, I'm aware for the most of them but good thing you brought the i5 Problems with DGPU in surface, you and others with i5 definitely need a better solution which I will provide soon.
Second: Apologies for last time, I may have given a harsh answer which it wasn't completely your fault as that turned out to be another problem and documentation conflict between different forums as everyone has it's own approach.
Also extra note, OpenCore support for this machine will be added probably September/October, it's working excellent but I need more testing and need to make the Open Core General guide before updating any of my guides to support it.
