UHD630

[qinjingliuan]

Now when using the Thunderbolt external display, only insert a type C data cable, which may cause the internal screen of the computer to appear screen, which will seriously lead to a direct crash, and can only force a restart. At present, the solution uses HDMI or dp interface and uses normally. Can you solve it, only use a data cable to connect the lightning interface to expand the display, so that it will be more convenient. Another point, I don't know what the situation is, there are no stripes on the left side of this screen, and there will be some small stripes under the right side, which were only discovered after the recent replacement of the boot. The screen is a 1080p screen at 300 Hz.

[qinjingliuan]

[qinjingliuan]

[qinjingliuan]

The screen appears striped, which seems to have something to do with the data cable I used the C port to connect the monitor. I found that if I plug in the C port of the monitor, there will be a stripe on the left side at the bottom. If it is removed, this provision will not disappear. This thin stripe will disappear after the restart, but the right one will not disappear, the right one will always be there. It will work better under Windows, and there will be no streaks on the left after insertion. But there are always very fine streaks on the right.

[qinjingliuan]

Using the type C connection, the monitor cannot be turned on when booting, it will be directly stuck on that boot interface for a long time, and finally it can only be forced to restart. After entering the system, you can use the Type C connection to use, but it will also force a crash after a while

[kingo132]

The type-c port monitor will use Intel Graphic and the HDMI output will use RX5700M. Therefore, it is not recommended to use both of them even in Windows. I disabled the type-c port video output by default. You can change the platform-id and framebuffer id in the config file to re-enable this but maybe a little buggy. Use the platform-id and framebuffer id I commented out there. I haven't tested much since I don't think it's a good usage. There're some HDMI to type-c video adapters in the market. I think using that will be more stable.

[qinjingliuan]

OK thanks

[TechProgenitor]

Hi, I’ve been trying to get down to the bottom of the iGPU issues over the past few days and was thinking of preallocating more DVMT memory to see if it helps. I’m not sure if it’ll fix the page fault error but at the very least it might allow us to boot the system without patching the stolenmem.

@kingo132 I see that since the first release of this repo you were able to dump the full BIOS options with what I'm assuming is Ifrextract and that there’s two hidden options related to DVMT within this BIOS.

I’m using UEFITool to read the latest BIOS revision and was wondering which specific body you extracted from the ROM in order to dump the settings with ifrextract? I’m assuming that it’s PE32 but it doesn’t seem to work... Whenever I use ifrextract on the PE32 body, the tool gives me an unknown protocol error. I'm not sure what I'm doing wrong, but I just wanted to dump the latest BIOS options in case any of the variable values have changed.

@qinjingliuan regarding your glitches, relaying what @kingo132 said, I personally use an HDMI to USBC cable and haven’t ran into any of your striping issues (except for the crashing when booting). For your internal display, I don't have a 300hz screen. I've got the 144hz version, however, I recently fixed the performance/stability for macOS Ventura by applying the following device properties. Give them a try to see if they fix anything for you. At the bare minimum you can apply the @(1,2,3)name, Force_Load_FalconSMUFW, and PP,PP_PhmUseDummyBackEnd device properties.

[TechProgenitor]

Nevermind, I got Ifrextract working now... I was using a ROM which I converted from Dell's website instead of my own ROM. Looks like the variables within the BIOS settings haven't changed.

[TechProgenitor]

Hi it’s been a while, Just wanted to update this thread and say that I couldn’t seem to preallocate more DVMT memory for the iGPU. To be honest though, it might be a red herring since this crash appears when using the headless framebuffers that don’t require any preallocation.

According to my research, the Intel framebuffer doesn’t cause a kernel panic if the stolenmem is set to 30MB or less. (The fbmem property isn’t required.)

Finally I believe I’ve discovered something that brings the iGPU fix an inch closer though you guys can confirm it for me: The boot argument igfxagdc=0 appears to make sleep/wake work properly with the iGPU enabled.

[qinjingliuan]

I changed DVMT memory for the iGPU in bios the DVMT memory changed 32M but igpu not working

TechProgenitor @.***> 于2023年7月4日周二 05:20写道：

Hi it’s been a while, Just wanted to update this thread and say that I couldn’t seem to preallocate more DVMT memory for the iGPU. To be honest though, it might be a red herring since this crash appears when using the headless framebuffers that don’t require any preallocation.

According to my research, the Intel framebuffer doesn’t cause a kernel panic if the stolenmem is set to 30MB or less. (The fbmem property isn’t required.)

Finally I believe I’ve discovered something that brings the iGPU fix an inch closer though you guys can confirm it for me: The boot argument igfxagdc=0 appears to make sleep/wake work properly with the iGPU enabled.

— Reply to this email directly, view it on GitHub https://github.com/kingo132/a51m-r2-5700m-hackintosh/issues/16#issuecomment-1619162822, or unsubscribe https://github.com/notifications/unsubscribe-auth/ASZDFCQANMVMHVPHMXUQVMDXOMZTFANCNFSM6AAAAAAXUS23HE . You are receiving this because you were mentioned.Message ID: @.***>

-- 柳岸青径工作室

[qinjingliuan]

I forgot to change the DVMT memory to 32M or 64M

[qinjingliuan]

How did you add the parameters? Do you have a screenshot? igfxagdc=0

[TechProgenitor]

You add the parameters to the boot-args section in the config.plist file (it's from the WhateverGreen kext). It doesn't seem to fix the crashing though, just sleep/wake.

[qinjingliuan]

I've tried it, same result as yours

[cechexl]

The TYPEC output is perfect, but the problem is that the dual graphics card output mode accelerates and causes instability. It is necessary to disable METAL at the core display position

Please refer to my evaluation or EFI for specific operations. Currently, the only issue is that the sound card cannot output HDMI, and the problem lies in the sound card encoding.

[qinjingliuan]

Today the system updated a small version, I tried it out, whether the c port can work normally with the external monitor under the 13 system, it turned out that the work is still not normal under the 13 system, but everything works normally under the 12 system , no kernel panics. I guess this may be a case of system versions targeting different processors, and 13 is targeting arm architecture processors.

cechexl @.***> 于2023年7月27日周四 10:18写道：

At present I am such a solution, can also stable output N card, add no N card shield patch, no impact on a card, stable operation, no a card still normal output TYPEC

— Reply to this email directly, view it on GitHub https://github.com/kingo132/a51m-r2-5700m-hackintosh/issues/16#issuecomment-1652815128, or unsubscribe https://github.com/notifications/unsubscribe-auth/ASZDFCQ2WJD6U6NQDFUCKIDXSHFXTANCNFSM6AAAAAAXUS23HE . You are receiving this because you were mentioned.Message ID: @.***>

-- 柳岸青径工作室

[cechexl]

https://bbs.pcbeta.com/viewthread-1975488-1-1.html

[TechProgenitor]

Hey everyone, Not sure if this fixes or breaks anything, but it looks like you can make the Activity Monitor show the activity of the AMD and Intel GPUs by assigning a PCI slot number to each of them.

This is done by adding the AAPL,slot-name device property to each GPU: For instance, I assigned the AMD GPU "Slot-0" and the Intel GPU "Slot-1".

This is the before and after: