Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Status
The first post of this thread is a WikiPost and can be edited by anyone with the appropiate permissions. Your edits will be public.
PSA: Don't attempt to use shiki bit 32 with the iMac Pro board-id on 10.15.2. You will get into a boot loop and won't even be able to boot into recovery with the keyboard shortcut. I learned the hard way. Luckily I was able to recover.
 
PSA: Don't attempt to use shiki bit 32 with the iMac Pro board-id on 10.15.2. You will get into a boot loop and won't even be able to boot into recovery with the keyboard shortcut. I learned the hard way. Luckily I was able to recover.
Do you have dual CPU's?
 
I do. Is that the issue?
Take a look at this
[automerge]1576517019[/automerge]
to avoid reboot loops in the future either enable :
PanicNoKextDump=True
and debug=0x144 keepsyms=1
But most importantly use a USB stick to load the OC. Once you remove it you go back to default Apple boot
 
Last edited:
  • Like
Reactions: h9826790
guys so things are by now mature?
I too, if possible, ask you to make a detailed step by step guide for newcomers if there are no problems.
A question then: is it necessary to install Catalina and you can stay with 10.14.6?
 
  • Like
Reactions: TheDevler
guys so things are by now mature?
I too, if possible, ask you to make a detailed step by step guide for newcomers if there are no problems.
A question then: is it necessary to install Catalina and you can stay with 10.14.6?
Mature? It's not 24hs yet from @startergo finally get it working with Safari…

Mojave works for encoding and decoding via AMD GPUs too.
 
guys so things are by now mature?
I too, if possible, ask you to make a detailed step by step guide for newcomers if there are no problems.
A question then: is it necessary to install Catalina and you can stay with 10.14.6?

The guide in post #594 still applicable.

And your question is already answered at post #910
 
1) Is an iMac Pro MLB (instead of a Mac Pro MLB) necessary?
Actually No. Just tested it.
2) Proper FaceTime and iMessage apparently depends on identifiers that include the MLB. However, now that we are changing the MLB and retaining some of the other identifiers, are we putting our machines at risk of being blacklisted?
If you put your own cMP5,1 valid MLB the iMessage keys are identical to the clean OSX. And the acceleration and the DRM playback are preserved. So this might be the preferred way. Use Hackintool to compare.

We have to figure out a way to enable all video ports on the card.
 
  • Like
Reactions: Chung123 and cdf
We have to figure out a way to enable all video ports on the card.
You mean apart from using the WEG boot argument agdpmod=pikera ?

Tested, in config.plist -> Kernel -> Patch, add the following code
Code:
<dict>
                <key>Base</key>
                <string></string>
                <key>Comment</key>
                <string>Patch AppleGraphicsDevicePolicy board-id to board-ix</string>
                <key>Count</key>
                <integer>0</integer>
                <key>Enabled</key>
                <true/>
                <key>Find</key>
                <data>Ym9hcmQtaWQ=</data>
                <key>Identifier</key>
                <string>com.apple.driver.AppleGraphicsDevicePolicy</string>
                <key>Limit</key>
                <integer>0</integer>
                <key>Mask</key>
                <data></data>
                <key>MaxKernel</key>
                <string></string>
                <key>MinKernel</key>
                <string></string>
                <key>Replace</key>
                <data>Ym9hcmQtaXg=</data>
                <key>ReplaceMask</key>
                <data></data>
                <key>Skip</key>
                <integer>0</integer>
            </dict>
 
Last edited:
  • Like
Reactions: startergo
Actually No. Just tested it.

If you put your own cMP5,1 valid MLB the iMessage keys are identical to the clean OSX. And the acceleration and the DRM playback are preserved. So this might be the preferred way. Use Hackintool to compare.

I also tested it with my original 5,1 MLB, everything still work.

So, it seems we can further narrow down the key factor to BoardProduct and SystemSKUNumber. Even though I also set the ChassisVersion and BID to Mac-7BA5B2D9E42DDD94, but they seems not as important as BoardProduct or SystemSKUNumber.
 
Last edited:
  • Like
Reactions: Chung123 and cdf
So, it seems we can further narrow down to the key to BoardProduct and SystemSKUNumber. Even though I also set the ChassisVersion and BID to Mac-7BA5B2D9E42DDD94, but they seems not as important as BoardProduct or SystemSKUNumber.
I only added:
Code:
<key>SystemProductName</key>
            <string>iMacPro1,1</string>
Different than cMP5,1. But looking at the Hackintool The iMac Pro board ID is automatically applied.
 
I only added:
Code:
<key>SystemProductName</key>
            <string>iMacPro1,1</string>
Different than cMP5,1. But looking at the Hackintool The iMac Pro board ID is automatically applied.

Oh, I forgot to mention that I am using manual mode. Which can keep SystemProductName as MacPro5,1, but only inject the board-id Mac-7BA5B2D9E42DDD94
 
Confirmed full HWAcell.

This is on Asus Gryphon Z97 (Haswell HD4600 Intel Graphics ) Hackintosh with MSI Armor Radeon RX580.
Maybe useful to someone in the future.

I didn't bother on my MacPro 5,1. Not much point using it, no boot screen, no AirDrop, generally just way way too slow.

Changed Product Model to: iMacPro1,1
Boot args:
shikigva=144
shiki-id=Mac-7BA5B2D9E42DDD94
keepsyms=1

In VideoPro, the reload icon MUST be clicked, simply opening does not do any checking.

VideoProc_HW_Acell_H264_HVEC.png


Clover_SMBIOS.png


Clover_Boot.png


Clover_Graphics.png
 
  • Like
Reactions: Warriors Photo
Confirmed full HWAcell.

This is on Asus Gryphon Z97 (Haswell HD4600 Intel Graphics ) Hackintosh with MSI Armor Radeon RX580.
Maybe useful to someone in the future.

I didn't bother on my MacPro 5,1. Not much point using it, no boot screen, no AirDrop, generally just way way too slow.

Changed Product Model to: iMacPro1,1
Boot args:
shikigva=144
shiki-id=Mac-7BA5B2D9E42DDD94
keepsyms=1

In VideoPro, the reload icon MUST be clicked, simply opening does not do any checking.

View attachment 883248

View attachment 883249

View attachment 883250

View attachment 883251
Hello there. I have a Z77 Hackintosh that I would like to start using over my Mac Pro. Are you using the HD4600 for QuickSync or the GPU? Thank you in advanced for this input. :)
[automerge]1576612238[/automerge]
All 6,1 native GPU won’t work. Need at least Polaris to provide HWAccel.

However, 6,1 + eGPU may work. Still waiting for someone to confirm it.
I like and don't like the sound of this. I was hoping to move on to the newer generation of Mac Pro. At least I can look at the bright side. The 5,1 Mac Pro, is still a very powerful PC. :)
 
Hello there. I have a Z77 Hackintosh that I would like to start using over my Mac Pro. Are you using the HD4600 for QuickSync or the GPU? Thank you in advanced for this input.

Hey mate,

Okay, there's some peculiarities or myths surrounding HWAcel.

I had the iGPU disabled, and Handbrake using VideoToolBox H265 encoder.

First run I noticed 60%+ CPU core utilisation, Radeon GPU usage was noticicible but I don't have any numbers because the GPU history meter in Activity Monitor doesn't have any reference measurements. I don't know if the meters being half the window vertical can be assumed to be 50% use or more or less.

Then I realised the Handbrake profile had Video Filters enabled, it is known the filters, decomb, de interlace etc are not processed on the GPU, these are purely CPU bound, so I had to do another run, disable the filters to get amore accurate understanding of resource usage, sure enough CPU use dropped to about 20% use, that is explained because Audio encoders only use the CPU, rest of resource use was on the Radeon.

Then I enabled the IGPU in the UEFI firmware (BIOS) and expected Handbrake to default to using Intel Iris iGPU but it didn't.

Ran same tests in that ugly VideoProc app, H265 encoding in this app showed the Intel doing the encoding not the Radeon. Then I read VideoProc knowledge base that Radeon RX580 does not accelerate H265, only h264.

This was rather pointless because h264 on the GPU is no faster than on the CPU. The big gains, according to Handbrake tech info is that h265 on GPU is where the big advantage is seen over CPU encoding.

When I ran H264 on VideoProc it only used the CPU, but I went so fast I was unable to make an assessment because it is trial version and the second encode probably encoded only the first minute of the video source, and I wouldn't pay for that junk windows ported software anyway. Probably there are other apps that people regularly use that would be more suited to testing to get useful metrics.

As far as other source material and apps are used, such as Quicktime playing movies etc I didn't have the iGPU enable to observe Intel Quicksync because at that stage I was primarily interested in if the Radeon was doing the processing.

In my observations when the iGP was enabled it appeared to be ignored in favour of the Radeon, but I can't rule out the possibility that my Clover configuration is having an effect and preventing the iGPU from functioning.
 
  • Like
Reactions: Warriors Photo
Hey mate,

Okay, there's some peculiarities or myths surrounding HWAcel.

I had the iGPU disabled, and Handbrake using VideoToolBox H265 encoder.

First run I noticed 60%+ CPU core utilisation, Radeon GPU usage was noticicible but I don't have any numbers because the GPU history meter in Activity Monitor doesn't have any reference measurements. I don't know if the meters being half the window vertical can be assumed to be 50% use or more or less.

Then I realised the Handbrake profile had Video Filters enabled, it is known the filters, decomb, de interlace etc are not processed on the GPU, these are purely CPU bound, so I had to do another run, disable the filters to get amore accurate understanding of resource usage, sure enough CPU use dropped to about 20% use, that is explained because Audio encoders only use the CPU, rest of resource use was on the Radeon.

Then I enabled the IGPU in the UEFI firmware (BIOS) and expected Handbrake to default to using Intel Iris iGPU but it didn't.

Ran same tests in that ugly VideoProc app, H265 encoding in this app showed the Intel doing the encoding not the Radeon. Then I read VideoProc knowledge base that Radeon RX580 does not accelerate H265, only h264.

This was rather pointless because h264 on the GPU is no faster than on the CPU. The big gains, according to Handbrake tech info is that h265 on GPU is where the big advantage is seen over CPU encoding.

When I ran H264 on VideoProc it only used the CPU, but I went so fast I was unable to make an assessment because it is trial version and the second encode probably encoded only the first minute of the video source, and I wouldn't pay for that junk windows ported software anyway. Probably there are other apps that people regularly use that would be more suited to testing to get useful metrics.

As far as other source material and apps are used, such as Quicktime playing movies etc I didn't have the iGPU enable to observe Intel Quicksync because at that stage I was primarily interested in if the Radeon was doing the processing.

In my observations when the iGP was enabled it appeared to be ignored in favour of the Radeon, but I can't rule out the possibility that my Clover configuration is having an effect and preventing the iGPU from functioning.

This is the most accurate way to check if your AMD GPU is doing any hardware decode / encode.

https://forums.macrumors.com/thread...n.2180095/page-11?post=27484973#post-27484973
 
  • Like
Reactions: Warriors Photo
Hey mate,

Okay, there's some peculiarities or myths surrounding HWAcel.

I had the iGPU disabled, and Handbrake using VideoToolBox H265 encoder.

First run I noticed 60%+ CPU core utilisation, Radeon GPU usage was noticicible but I don't have any numbers because the GPU history meter in Activity Monitor doesn't have any reference measurements. I don't know if the meters being half the window vertical can be assumed to be 50% use or more or less.

Then I realised the Handbrake profile had Video Filters enabled, it is known the filters, decomb, de interlace etc are not processed on the GPU, these are purely CPU bound, so I had to do another run, disable the filters to get amore accurate understanding of resource usage, sure enough CPU use dropped to about 20% use, that is explained because Audio encoders only use the CPU, rest of resource use was on the Radeon.

Then I enabled the IGPU in the UEFI firmware (BIOS) and expected Handbrake to default to using Intel Iris iGPU but it didn't.

Ran same tests in that ugly VideoProc app, H265 encoding in this app showed the Intel doing the encoding not the Radeon. Then I read VideoProc knowledge base that Radeon RX580 does not accelerate H265, only h264.

This was rather pointless because h264 on the GPU is no faster than on the CPU. The big gains, according to Handbrake tech info is that h265 on GPU is where the big advantage is seen over CPU encoding.

When I ran H264 on VideoProc it only used the CPU, but I went so fast I was unable to make an assessment because it is trial version and the second encode probably encoded only the first minute of the video source, and I wouldn't pay for that junk windows ported software anyway. Probably there are other apps that people regularly use that would be more suited to testing to get useful metrics.

As far as other source material and apps are used, such as Quicktime playing movies etc I didn't have the iGPU enable to observe Intel Quicksync because at that stage I was primarily interested in if the Radeon was doing the processing.

In my observations when the iGP was enabled it appeared to be ignored in favour of the Radeon, but I can't rule out the possibility that my Clover configuration is having an effect and preventing the iGPU from functioning.
This was very much appreciated and worth the read. Thank you. :)
[automerge]1576623080[/automerge]
This is the most accurate way to check if your AMD GPU is doing any hardware decode / encode.

https://forums.macrumors.com/thread...n.2180095/page-11?post=27484973#post-27484973
This was also very much worth looking into. I am glad I've looked into this deeper. I am mostly going to be shooting with H.264 and HEVC codecs and like having the Hardware Acceleration working for when I edit. If FCPX can't do HEVC encode/decoding with my current set up then that's what PROXY and Optimized media is for. :)
 
On a clean install of macOS 10.15.2, I can now unfortunately confirm @startergo's troubles with dual processors and spoofing.
What exactly is the issue?
Can OC fake this out? (Something like pretend it’s actually a single socket 12 core CPU with 24 threads)
if you boot cpus=1 (single core / thread) is the issue gone?
 
I believe the offending kext loads in the prelinked kernel stage so it won't help, but he can try.
Also:
Mac-7BA5B2D9E42DDD94 and the 4 known processors are:

Intel Xeon W-2140B @ 3.2GHz ( 8 cores and 16 threads)
Intel Xeon W-2150B @ 3.0GHz (10 cores and 20 threads)
14-Core (28 threads) Xeon W-2170B
18-Core (36 threads) Xeon W-2191B

How I fixed it?
I moved AppleIntelMCEReporter.kext to the kexts folder in my ESP. Although to tell you the truth I don't see this kext loaded when the operating system is booted so it may be as good as just (backup) and remove it from its original location.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.