Jump to content
1504 posts in this topic

Recommended Posts

On 1/17/2020 at 6:40 PM, vit9696 said:

Should be shikigva=80, but it is not supported before 10.15, and RX580 drivers have a bug with it at the moment as of 10.15.2. No option for you at this step.

Hi @vit9696

 

First of all a big thanks to the team for WEG.

 

Although I’m using a Vega 56 I still have my RX 580 Pulse that I prefer, but not using it since with Vega I do have Netflix on my headless iMac19,1 with Safari; first world problems, I know.

 

You are saying that the DRM issue with Polaris is driver related, is this something that is expected to be resolved by Apple and given up by the WEG team? If not, how can users assist into making it work?

 

- In terms of logs

- In terms of testing

 

...etc?

 

I myself am a software engineer but I specialize on iOS and MacOS development with the known frameworks. I haven’t touched the surface of what you guys are doing but if there is some way I could assist, I would be more than happy to do so.

 

 

  • Like 1
On 1/18/2020 at 10:17 AM, vit9696 said:

The only way for IGPU to reach high clocks is to use GuC. This claim looks false to me as WhateverGreen does nothing to IGPU performance by default.

I had overclocked iGPU on my MSI past 1.3GHz and no i didn't feel like hit with WhateverGreen.

 

 

Anyway I moved now backwards to Sandy Bridge.. with HD3000 and iTunes shows blackscreen but plays movies if not downloaded. IS this again limitation of my hw? like i had with earlier MSI with HD4600 which i never got to it work and it was reported.

 

 

Sorry i know there is mention of Sandy Bridge but i am not 100% sure has it been able to play DRM.

 

 

Thanks of your time!

On 1/18/2020 at 9:17 AM, vit9696 said:

The only way for IGPU to reach high clocks is to use GuC. This claim looks false to me as WhateverGreen does nothing to IGPU performance by default.

Another thing is driving me crazy (I cannot understand how this could happen) is that many users are reporting boost in performance when _not_ using WeG, either using the GPU completely natively (without any additional setups in the config.plist file) or by using Inject ATI + device-id / framebuffer patches and so on. 

 

Here if you want you can use the chrome plug-in to translate the page from Italian to English/russian (its very basic italian, nothing poetic) Topic here

 

Their improvement impacts on GeekBench for example, experiencing a 15-20% improvement on benchmarks (55K metal score for a Vega56 with WeG to 71K without WeG)

 

I cannot understand how this could happen, those improvements are very frequent (as users report them here also, again very basic italian translation so no worries)

 

Thanks in advance for the clarifications, I'm very curious to know how this could happen (basing on your knowledge/opinions) :)

On 1/21/2020 at 4:46 PM, XantheFIN said:

I had overclocked iGPU on my MSI past 1.3GHz and no i didn't feel like hit with WhateverGreen.

 

 

Anyway I moved now backwards to Sandy Bridge.. with HD3000 and iTunes shows blackscreen but plays movies if not downloaded. IS this again limitation of my hw? like i had with earlier MSI with HD4600 which i never got to it work and it was reported.

 

 

Sorry i know there is mention of Sandy Bridge but i am not 100% sure has it been able to play DRM.

 

 

Thanks of your time!

You can try to check here

https://github.com/acidanthera/WhateverGreen/blob/master/Manual/FAQ.Chart.md

 

Also note that HD3000 isn't supported on Mojave, as it's not metal compatible

 

8 hours ago, 1Ale1 said:

You can try to check here

https://github.com/acidanthera/WhateverGreen/blob/master/Manual/FAQ.Chart.md

 

Also note that HD3000 isn't supported on Mojave, as it's not metal compatible

 

Ah its then NO if i don't go eGPU way in future. Anyway thanks.

 

Yeah but i not interested move to Mojave in any time soon. Good if i get myself to upgrade to High Sierra. :hammer:

On 1/18/2020 at 9:17 AM, vit9696 said:

The only way for IGPU to reach high clocks is to use GuC. This claim looks false to me as WhateverGreen does nothing to IGPU performance by default.

If I use WEG then igpu (UHD630) works in FCPX at about 30% of power. 1080p multicam jams a lot. Just remove WEG and everything works perfectly. (Asus Prime Z390A; i9 9900k; MSI RX580 Armor)

Regarding GPU performance, we believe there is no difference in general, and so far were unable to observe any. However, it is possible to use different AGDP profile for testing purposes. Basically that is the only thing that can affect GPU power management when using WEG. Currently this can be configured with a device property or a boot argument — agdpmod. By default it is working in vit9696,pikera mode. You can try others:

  • agdpmod=vit9696 disables check for board-id (or add agdpmod property to external GPU).
  • agdpmod=pikera replaces board-id with board-ix.
  • agdpmod=ignore disables AGDP patches (vit9696,pikera value is implicit default for external GPUs).

Please report the results with all these variants.

2 hours ago, Najrad said:

None of the arguments helps. VidoProc shows the maximum iGPU load of 600MB during the h264 test. Without WEG it is 1.2GHz

IGPU? The only thing that affects IGPU performance is the framebuffer id. Most likely your setup is broken, simply, and you do not compare equivalent configurations (i.e. connector-full and connector-less).

You are definitely right. Not using WEG, I have everything working perfectly, as soon as I throw in WEG- iGPU gets stuck. It is definitely the fault of my poorly made configuration. On Mojave, the same WEG configuration works perfectly.

Edited by Najrad
On 1/17/2020 at 5:40 PM, vit9696 said:

Should be shikigva=80, but it is not supported before 10.15, and RX580 drivers have a bug with it at the moment as of 10.15.2. No option for you at this step.

@vit9696 that does include the RX570 too, right?

 

I'm using imac19,1 with a z390 aorus master, i5-9600k and a msi rx570. Maldon created a DSDT that's working brilliant.

I have internal GPU enabled and VDA decoder is fully supported

Only thing not working is Amazon Prime. 

 

shikigva=80 causes a system freeze when opening a prime video webpage (not even the video itself).

shikigva=16 enabled AppleTV app and Apple TV web content in Chrome.

 

If you have something I could test or in other way how I can help, please let me know.

 

Thanks

Eddie


 

edit: updated my profile but it doesn’t show with this post yet

Edited by Benjamin Franklin

Soooo

 

CPU: Intel Core i9-9900K

GPU: Vega 56

 

iMac19,1 SMBIOS with iGPU set as headless

macOS Catalina 10.15.2

WEG 1.3.6

 

Since a lot of people are complaining about Netflix working but Amazon Prime Video Crashing but as far as I've seen no one is uploading crash reports so here is one:

 

This is the WebKit crash when trying to watch an Amazon Prime Video:
com.apple.WebKit.WebContent_2020-01-31-181054-1_iMac.crash

On 1/28/2020 at 10:53 PM, Benjamin Franklin said:

shikigva=80 causes a system freeze when opening a prime video webpage (not even the video itself).

shikigva=16 enabled AppleTV app and Apple TV web content in Chrome.

ok I did some further testing. shikigva=80 ONLY freezes the screen, I can still enter keyboard commands (like Ctrl-Cmd-Power to restart).

Amazon Prime video works in Firefox!

Edited by Benjamin Franklin

I'm lost here. Hopefully someone here can share some light After upgrading to 10.15.3 my display image is shrunken at the center of the monitor. Like letterboxing but on all sides (windowboxing).

IMG_3568.thumb.jpeg.5284bf59f4d6433bdabadc1d2cb33a02.jpeg

My monitor is 2560x1440 resolution and MacOS System Information shows it is correctly displaying that resolution, but somehow the monitor shrink the image at the center of the screen. It is a very simple korean monitor with only one displayport input and NO overscan/underscan options. It worked flawlesly until 10.15.2.

My config:

Monitor 1440p connected through DisplaPort

i7-6600k

iGPU in headless configuration with Platform-ID Ox19120000

Radeon 5700 XT

Whatevergreen 1.3.6 with pikera boot option

1861536755_ScreenShot2020-02-01at11_39_15AM.thumb.png.de81ae1ac5d070d6d498818bbb2a27f5.png

 

 

Hi there,

 

I successfully switched from clover to open core but now I get an Error-Message in the Hackintool Logs from WEG

 

Timestamp                       (process)[PID]    
2020-02-04 17:30:59.555467+0100  localhost kernel[0]: (kernel) WhateverGreen       weg: @ console buffer allocation failure

 

Does anybody got an idea what that means - or even better - how to solve it? 

 

Thanks in advance. 

Edited by Maddeen
21 hours ago, Maddeen said:

Hi there,

 

I successfully switched from clover to open core but now I get an Error-Message in the Hackintool Logs from WEG

 


Timestamp                       (process)[PID]    
2020-02-04 17:30:59.555467+0100  localhost kernel[0]: (kernel) WhateverGreen       weg: @ console buffer allocation failure

 

Does anybody got an idea what that means - or even better - how to solve it? 

 

Thanks in advance. 

 

I am not sure if this is the case but make sure that Lilu is loaded before WEG on your OC config.plist.

Hey @WizeMan,

 

thanks but sadly that can't be the solution. Lilu is the first entry in my config and WEG and AppleALC doing their job. I got the correct display options, sounds and airplay.

Nevertheless WEG telling me this error when I start Hackintool - going to LOGS - set "WhatEverGreen" and click on generate.

I cant see any problems - but I hate error messages and want to get rid of them :)

Edited by Maddeen

Anyone tried 10.15.4 beta 1? Rumors have it that DRM issue is resolved with Polaris Ellesmere cards. Unfortunately I cannot test because I'm on a production Hackintosh right now, but just letting you know in case someone has the time / courage to do it.

Hello, on my dell lns7590 I found that "division-by-zero" caused the kernel to crash. The division-by-zero "WhateverGreen driver does not seem to solve this problem.
This is my problem report

Code:
panic(cpu 2 caller 0xffffff80042652fa): Kernel trap at 0xffffff7f8796816c, type 0=divide error, registers:
CR0: 0x0000000080010033, CR2: 0x00007fff8f2b3d48, CR3: 0x0000000856678022, CR4: 0x00000000003626e0
RAX: 0x017d68fdc0000000, RBX: 0x017d68fdc0000000, RCX: 0x0100000100000000, RDX: 0x0000000000000000
RSP: 0xffffff83b4213330, RBP: 0xffffff83b4213360, RSI: 0xffffff83b4213158, RDI: 0xffffff837d42a000
R8: 0x00000010ea7007f7, R9: 0xffffff837d435ca8, R10: 0x0000000000000000, R11: 0x0000000000000000
R12: 0x000000001fc8bfd0, R13: 0x0000000000000000, R14: 0xffffff83b4213380, R15: 0x0000000009a7ec80
RFL: 0x0000000000010246, RIP: 0xffffff7f8796816c, CS: 0x0000000000000008, SS: 0x0000000000000010
Fault CR2: 0x00007fff8f2b3d48, Error code: 0x0000000000000000, Fault CPU: 0x2, PL: 0, VF: 0

Backtrace (CPU 2), Frame : Return Address
0xffffff83b4212d90 : 0xffffff800413bb2b
0xffffff83b4212de0 : 0xffffff80042734d5
0xffffff83b4212e20 : 0xffffff8004264f4e
0xffffff83b4212e70 : 0xffffff80040e2a40
0xffffff83b4212e90 : 0xffffff800413b217
0xffffff83b4212f90 : 0xffffff800413b5fb
0xffffff83b4212fe0 : 0xffffff80048d2aa9
0xffffff83b4213050 : 0xffffff80042652fa
0xffffff83b42131d0 : 0xffffff8004264ff8
0xffffff83b4213220 : 0xffffff80040e2a40
0xffffff83b4213240 : 0xffffff7f8796816c
0xffffff83b4213360 : 0xffffff7f87965c6e
0xffffff83b42137e0 : 0xffffff7f8793bcff
0xffffff83b42139a0 : 0xffffff7f877fed33
0xffffff83b4213a30 : 0xffffff7f877feb90
0xffffff83b4213ac0 : 0xffffff800486739b
0xffffff83b4213b10 : 0xffffff7f87805256
0xffffff83b4213b60 : 0xffffff8004870443
0xffffff83b4213ca0 : 0xffffff8004222d12
0xffffff83b4213db0 : 0xffffff80041419d8
0xffffff83b4213e10 : 0xffffff8004118635
0xffffff83b4213e70 : 0xffffff800412f0e5
0xffffff83b4213f00 : 0xffffff800424b575
0xffffff83b4213fa0 : 0xffffff80040e3226
Kernel Extensions in backtrace:
com.apple.iokit.IOGraphicsFamily(569.4)[1F9B5D88-52DB-3A16-8373-4F608A3CB2D8]@0xffffff7f877e2000->0xffffff7f87832fff
dependency: com.apple.iokit.IOPCIFamily(2.9)[ADD485B5-3EF8-37C4-B3C5-F86326E497A4]@0xffffff7f84cf8000
com.apple.driver.AppleIntelCFLGraphicsFramebuffer(14.0.4)[A30AC04F-FB51-328A-8942-A2BE0CEB5C3D]@0xffffff7f8791f000->0xffffff7f87b4cfff
dependency: com.apple.iokit.IOPCIFamily(2.9)[ADD485B5-3EF8-37C4-B3C5-F86326E497A4]@0xffffff7f84cf8000
dependency: com.apple.iokit.IOACPIFamily(1.4)[3D78401B-5D2D-33BC-9E41-DD2164EA874D]@0xffffff7f84ce6000
dependency: com.apple.iokit.IOAcceleratorFamily2(438.3.1)[66992525-3204-3CB0-8F03-4B70031B1CF2]@0xffffff7f87852000
dependency: com.apple.iokit.IOReportFamily(47)[988360A2-2E10-3014-A119-BE81BC045A10]@0xffffff7f84c28000
dependency: com.apple.AppleGraphicsDeviceControl(4.7.2)[2F63196D-03C6-3E49-BE5D-574F4AADED1A]@0xffffff7f87915000
dependency: com.apple.iokit.IOGraphicsFamily(569.4)[1F9B5D88-52DB-3A16-8373-4F608A3CB2D8]@0xffffff7f877e2000

As well as EFI and issues     

Edited by cookiies

Just confirming that with WEG 1.3.6, iMac19,1, RX 580, i9-9900K iGPU set in headless mode, shikigva=80 behaves as with the Vega 56. Which means that we have Apple TV, Netflix on Safari but prime video crashes Webkit.

Edited by WizeMan
On 2/2/2020 at 1:11 AM, antiflash said:

I'm lost here. Hopefully someone here can share some light After upgrading to 10.15.3 my display image is shrunken at the center of the monitor. Like letterboxing but on all sides (windowboxing).

IMG_3568.thumb.jpeg.5284bf59f4d6433bdabadc1d2cb33a02.jpeg

My monitor is 2560x1440 resolution and MacOS System Information shows it is correctly displaying that resolution, but somehow the monitor shrink the image at the center of the screen. It is a very simple korean monitor with only one displayport input and NO overscan/underscan options. It worked flawlesly until 10.15.2.

My config:

Monitor 1440p connected through DisplaPort

i7-6600k

iGPU in headless configuration with Platform-ID Ox19120000

Radeon 5700 XT

Whatevergreen 1.3.6 with pikera boot option

1861536755_ScreenShot2020-02-01at11_39_15AM.thumb.png.de81ae1ac5d070d6d498818bbb2a27f5.png

 

 


LoL,, 0x19120000 is not headless / connector-less ig-platform-id. It should be 0x19120001

On 2/1/2020 at 9:11 PM, antiflash said:

I'm lost here. Hopefully someone here can share some light After upgrading to 10.15.3 my display image is shrunken at the center of the monitor. Like letterboxing but on all sides (windowboxing).

IMG_3568.thumb.jpeg.5284bf59f4d6433bdabadc1d2cb33a02.jpeg

My monitor is 2560x1440 resolution and MacOS System Information shows it is correctly displaying that resolution, but somehow the monitor shrink the image at the center of the screen. It is a very simple korean monitor with only one displayport input and NO overscan/underscan options. It worked flawlesly until 10.15.2.

My config:

Monitor 1440p connected through DisplaPort

i7-6600k

iGPU in headless configuration with Platform-ID Ox19120000

Radeon 5700 XT

Whatevergreen 1.3.6 with pikera boot option

1861536755_ScreenShot2020-02-01at11_39_15AM.thumb.png.de81ae1ac5d070d6d498818bbb2a27f5.png

 

 

This is overscan/underscan problem.

 

1. Open terminal.
2. You want to open /var/db/.com.apple.iokit.graphics file. To do that, enter following command in terminal: sudo open -a TextEdit /var/db/.com.apple.iokit.graphics
3. TextEdit will open with the file. You’ll notice that the file has following structure:
751716839_!DOCTYPEplistPUBLIC-AppleDTDPLIST1.0ENhttpwww.apple.comDTDsPropertyList.thumb.png.b809f50792e6de989a76b27e34e77e69.png
4. Starting with line 5, every key tag is a separate display and following dict is list of preferences for it.
5. Note: You may need to do a bit of trial and error to find the correct display. In my case, I only use a limited number of displays, so I reset the settings for every display, restarted and it was fixed.
6. Find the lines that say <key>pscn</key>. This is the preference that controls underscan/overscan.
7. Next line after every pscn key will look like this: <integer>10000</integer> (line 34 above).
8. 10000 means that there is no underscan. Value less than 10000 means that display will be underscanned, larger value means it is overscanned.
9. For values less than 10000, change them back to 10000.
10. Save the file, you may need to enter your admin password.
11. Restart.

The issue should be fixed.

  • Like 1

Hello everyone.

 

I Have an R9 380 Graphics card (MSI ARMOR 2x) its the 2048 Vram model.

 

It should work OOB with WEG, but it boots to black screen always.

 

I am using a VFIO VM with passthrough of the entire PCI card, and I have tried with clover and open-core with and without WEG.

 

Without WEG it spams VRAM bugs, and with WEG it does not, but even though it detects the GPU, it seems the VRAM is 0mb framebuffer (According to when I can see it with the default qemu display connected.

 

Without the default display, I get no screen.

 

I am not sure if there is a problem with the Connectors or something with this card, lots of people have reported success with the 4gb version of the card, but mine is 2GB, can it be fixed?

×
×
  • Create New...