New PC display loses connection to GPU, no signal

[RichardStones]

Hello everyone!
Hoping someone can help me determine what is the root cause of my issue.

Brand new build ~2 weeks ago https://cz.pcpartpicker.com/list/dnMR28
I'm running this monitor with a DPI cable. 34" Dell S3422DWG
System is running Win 10 Pro.

A few days ago suddenly the screen went black, there was no signal from the gpu. Restarted the PC and the GPU was disabled in the device manager with code 22.
When I enable it, it says there is some driver error code 31. I tried downloading the AMD adrenaline software to update the drivers. So I disabled the gpu, installed new drivers and activated it again, and after restart it works.
However yesterday this happened 3 times in a row, within 10 min intervals.

I tried troubleshooting by adding gpu device IDs in the gpedit.msc, to not allow Windows to automatically update the drivers, but it did not help at all.
Maybe it's worth noting that the system detects 2 gpus (I assume one is attached to the integrated MoBo gpu), and the 6800XT is the one not working.
So in the device manager there is:
AMD Radeon RX 6800 XT
AMD Radeon(TM) Graphics
 

I have the whole hardware under warranty, but I just want to determine what’s the problem before I contact the shop.
Is it a software or hardware issue.

Thank you all in advance!
Cheers!

[ChrisLoudon]

2 hours ago, RichardStones said:

I have the whole hardware under warranty, but I just want to determine what’s the problem before I contact the shop.
Is it a software or hardware issue.

Thank you all in advance!
Cheers!

Hi buddy,

 

So firstly, yes, the 2nd AMD GPU entry in the device manager is the iGPU on the 7800X3D.

 

With the PC up and running are you able to have a poke around the Windows Event Viewer to see whats what? In the search bar at the bottom of the screen type 'Event Viewer' then in the window that pops up, expand 'Windows Logs' then click on 'System' This will take a few seconds to load all the entries. In the right hand pane, you will see a list of time stamped entries. Scroll down to when the last time your PC black-screened and see if you can see anything that will shed some light on what is happening. There's loads of Gobbledygook but hopefully you will see some Red (Critical) or Amber (Warning) messages relating to your GPU driver / hardware.

 

While the above route might only point to the obvious, at least you know where to focus your efforts.

 

Assuming its a GPU issue then its either a software / driver issue or a hardware problem.

 

If its software, the only real 2 ways of attempting to fix it would be to use a tool called DDU and completely remove any and all traces of the AMD driver then reinstall. Make sure you have the correct AMD driver for your GPU ready and downloaded direct from AMD so Windows Update doesn't try and beat you to it.

 

Whether or not you find an 'Event' you should probably download and install a hardware monitoring utility like HWInfo. It will show you CPU / GPU load and temps and might alert you to something like overheating or throttling that can be the cause of issues like this.

 

If you have any Overclocks enabled the set everything back to stock, at least until you can get to the bottom of the issue.

 

Make sure your motherboard BIOS and chipset drivers are up-to-date.

 

If that doesn't work, a fresh install of Windows never hurts. Its a bit of a ball-ache if you've got loads of files, games and applications that need redownloading and installing but there's nothing better than a fresh install of Windows and its normally the only real way to rule out a software problem. DON'T do a Windows Repair! A proper 'delete all partitions / everything' and start again.

 

If none of the above methods help then the chances are its the XT6800 that's the problem child. Take it out and boot from the iGPU. You're unlikely to be gaming with the iGPU but at least now, you know what to say to the shop when you go their 1st line of support. If you have another PC (or access to another PC) then whack the XT6800 in there and see if you can replicate the problem.

 

If the Mobo has more than 1 PCIe x 16 slot then change slots. If the PC has more than 1 stick of RAM then run it with a single stick. If the PC has the RAM running in XMP/DOCP/EXPO mode then run it at stock.

 

Check, the PCIe power cables from the PSU. If the PSU has more than one PCIe power socket then try another one.

 

Does the fan on the GPU kick in when gaming / under load?

 

Keep us posted.

 

P.S. the above list isn't necessarily in any particular order, just a few standard procedures I'd go though.

[RichardStones]

14 minutes ago, ChrisLoudon said:

Hi buddy,

 

So firstly, yes, the 2nd AMD GPU entry in the device manager is the iGPU on the 7800X3D.

 

With the PC up and running are you able to have a poke around the Windows Event Viewer to see whats what? In the search bar at the bottom of the screen type 'Event Viewer' then in the window that pops up, expand 'Windows Logs' then click on 'System' This will take a few seconds to load all the entries. In the right hand pane, you will see a list of time stamped entries. Scroll down to when the last time your PC black-screened and see if you can see anything that will shed some light on what is happening. There's loads of Gobbledygook but hopefully you will see some Red (Critical) or Amber (Warning) messages relating to your GPU driver / hardware.

 

While the above route might only point to the obvious, at least you know where to focus your efforts.

 

Assuming its a GPU issue then its either a software / driver issue or a hardware problem.

 

If its software, the only real 2 ways of attempting to fix it would be to use a tool called DDU and completely remove any and all traces of the AMD driver then reinstall. Make sure you have the correct AMD driver for your GPU ready and downloaded direct from AMD so Windows Update doesn't try and beat you to it.

 

Whether or not you find an 'Event' you should probably download and install a hardware monitoring utility like HWInfo. It will show you CPU / GPU load and temps and might alert you to something like overheating or throttling that can be the cause of issues like this.

 

If you have any Overclocks enabled the set everything back to stock, at least until you can get to the bottom of the issue.

 

Make sure your motherboard BIOS and chipset drivers are up-to-date.

 

If that doesn't work, a fresh install of Windows never hurts. Its a bit of a ball-ache if you've got loads of files, games and applications that need redownloading and installing but there's nothing better than a fresh install of Windows and its normally the only real way to rule out a software problem. DON'T do a Windows Repair! A proper 'delete all partitions / everything' and start again.

 

If none of the above methods help then the chances are its the XT6800 that's the problem child. Take it out and boot from the iGPU. You're unlikely to be gaming with the iGPU but at least now, you know what to say to the shop when you go their 1st line of support. If you have another PC (or access to another PC) then whack the XT6800 in there and see if you can replicate the problem.

 

If the Mobo has more than 1 PCIe x 16 slot then change slots. If the PC has more than 1 stick of RAM then run it with a single stick. If the PC has the RAM running in XMP/DOCP/EXPO mode then run it at stock.

 

Check, the PCIe power cables from the PSU. If the PSU has more than one PCIe power socket then try another one.

 

Does the fan on the GPU kick in when gaming / under load?

 

Keep us posted.

 

P.S. the above list isn't necessarily in any particular order, just a few standard procedures I'd go though.

Hello, thanks for the reply.

So, there are some logs in the event viewer. There are errors and warnings.

 

This seems to happen whenever the issue occurs, however sometimes it's labeled as warning, sometimes as error, as you can see in the image above.
Everything else should be completely up-to-date.
There was also a crash of the Service Control Manager "The AMD Crash Defender Service service terminated unexpectedly.  It has done this 1 time(s)."
And lastly there e2fexpress warning as well, but that seems unrelated.

I'll try with the DDU (is it from wagnard?), and if that doesn't work I guess I'll go for a clean install of Windows. 
If that doesn't fix it, maybe it's worth to send it to support, before removing parts and pieces of the PC, as it's under warranty.

Thanks for the help!

[ChrisLoudon]

8 minutes ago, RichardStones said:

Hello, thanks for the reply.

Hi buddy,

 

The DistributedCOM errors are unrelated, I have a million of these. The Amber warning for the 'Display' is interesting. in the grey box underneath the list, does that particular error message say anything enlightening?

 

If my GPU (Nvidia) was having issues, I would typically see a load of similar error messages reporting that the display driver had 'stopped responding and had successfully recovered'

 

On that note, when your PC black-screens, has the PC properly crashed or is it just the screen that's turned off?

 

You can usually tell when you hit the num-lock, scroll-lock or caps-lock key on the keyboard and see if the little green led turns on and off. If its does then the PC is still running and responsive but the screen has just turned off.

 

GPU drivers are designed to be able to recover from a crash but most of the time they don't fully recover and the screen goes black until you've done a reset / reboot.

 

Link for DDU below.

 

https://www.guru3d.com/download/display-driver-uninstaller-download/

[RichardStones]

1 minute ago, ChrisLoudon said:

Hi buddy,

 

The DistributedCOM errors are unrelated, I have a million of these. The Amber warning for the 'Display' is interesting. in the grey box underneath the list, does that particular error message say anything enlightening?

 

If my GPU (Nvidia) was having issues, I would typically see a load of similar error messages reporting that the display driver had 'stopped responding and had successfully recovered'

 

On that note, when your PC black-screens, has the PC properly crashed or is it just the screen that's turned off?

 

You can usually tell when you hit the num-lock, scroll-lock or caps-lock key on the keyboard and see if the little green led turns on and off. If its does then the PC is still running and responsive but the screen has just turned off.

 

GPU drivers are designed to be able to recover from a crash but most of the time they don't fully recover and the screen goes black until you've done a reset / reboot.

 

Link for DDU below.

 

https://www.guru3d.com/download/display-driver-uninstaller-download/

The display error is just: "Display driver amduw23g stopped responding and has successfully recovered."
But it has not fully recovered, because the gpu is disabled in the device manager. If I enable it, I get code 31, that something is wrong.
In fact, I can't even open the AMD adrenaline software in current state.

When I have black screen, the screen is completely running fine, and you can even hear programs/videos if they are on, it's just that the display is busted, and you can't see anything. Rebooting helps in a way that the iGPU gest called, and you can see stuff, but can't open games, or the Adrenaline software.

 

[ChrisLoudon]

6 minutes ago, RichardStones said:

The display error is just: "Display driver amduw23g stopped responding and has successfully recovered."
But it has not fully recovered, because the gpu is disabled in the device manager. If I enable it, I get code 31, that something is wrong.
In fact, I can't even open the AMD adrenaline software in current state.

When I have black screen, the screen is completely running fine, and you can even hear programs/videos if they are on, it's just that the display is busted, and you can't see anything. Rebooting helps in a way that the iGPU gest called, and you can see stuff, but can't open games, or the Adrenaline software.

Coolio,

 

Well its almost certainly a GPU issue then. That looks like the AMD version of what I would expect from an Nvidia card.

 

Work down the list I gave you in whatever order is the easiest and see if anything helps.

[RichardStones]

1 hour ago, ChrisLoudon said:

Coolio,

 

Well its almost certainly a GPU issue then. That looks like the AMD version of what I would expect from an Nvidia card.

 

Work down the list I gave you in whatever order is the easiest and see if anything helps.

So, I tried ti reinstall windows deleting all files, and now my PC is stuck in this weird loop crash, and I don’t know what to do.

 

The Republic of Gamers logo is on screen, and a window pop-up called Install Windows comes up with decription.

The computer restarted unexpectedly or encountered an unexpected error. Windows installation cannot proceed. To install windows, click “OK” to restart computer and resume installation.

 

but after clicking OK, the same thing happens again…

any ideas?

Edit: Solved it, with shift + f10 and regedit, nice tutorial on YT. I reinstalled it clean, and will download new programs. Here's to hoping it works.

[ChrisLoudon]

46 minutes ago, RichardStones said:

but after clicking OK, the same thing happens again…

any ideas?

Ooft, doesn't sound good.

 

When you tried to reinstall Windows, how did you go about it?

 

Did you create an install USB media from Microsoft's website and boot from the USB?

 

During the initial setup process make sure you select ' Custom.....' NOT 'Upgrade....'

 

The next page headed 'Where do you want to install Windows' Highlight ALL partitions on Drive 0 and click 'delete'. There are usually 3 or 4 partitions. Click Next (No need to create or format any new partitions as they will be created for for you.

 

 

[RichardStones]

1 minute ago, ChrisLoudon said:

Ooft, doesn't sound good.

 

When you tried to reinstall Windows, how did you go about it?

 

Did you create an install USB media from Microsoft's website and boot from the USB?

 

During the initial setup process make sure you select ' Custom.....' NOT 'Upgrade....'

 

The next page headed 'Where do you want to install Windows' Highlight ALL partitions on Drive 0 and click 'delete'. There are usually 3 or 4 partitions. Click Next (No need to create or format any new partitions as they will be created for for you.

 

 

No, I went with the option to "Reset PC" and I deleted all files and apps, and downloaded it from the cloud.

[ChrisLoudon]

Just now, RichardStones said:

No, I went with the option to "Reset PC" and I deleted all files and apps, and downloaded it from the cloud.

Ahh, ok.

 

That's not an ideal way to do it. Resetting kinda sometimes works but I've never had much luck with it.

 

The best way to do it would be to Google 'windows usb installation tool' and follow the instructions. You will need a usb stick with 8GB or more of capacity to make one.

 

https://support.microsoft.com/en-gb/windows/create-installation-media-for-windows-99a58364-8c02-206f-aa6f-40c3b507420d

 

Select Windows 10 (Or 11 if you prefer)

 

Once the tool has been downloaded, run it and when asked, click 'Create Installation Media.....'

 

The next page will then ask you to select what version of Windows you want. Win 10 Pro 64bit...... Be careful of this as if you are downloading on another PC, the tool will default to selecting the version on Windows on THAT PC.

 

After clicking 'next' you will then be asked where you want the media to be installed to. Click the USB Flash Drive.

 

Once this has been created, you will need to make sure the drive is plugged in and your ready to restart your PC and boot from the USB media.

 

This can either be done by hitting an 'F' key as the system boots similar to hitting 'Del' to get into the BIOS. This key is different depending on Mobo manufacturer so you'll need to find this out but sometimes it flashes up on the BIOS flash screen before the Windows logo displays.

 

Alternatively, you can force the issue by going into the BIOS and adjusting the BOOT Priorities to boot from the UEFI USB first.

 

*VERY IMPORTANT* Boot from UEFI USB (there can be a UEFI or NON-UEFI entry for older systems)

 

After the initial install process has completed, you will get 20 seconds to remove the USB stick before the system reboots and continues the install process.

 

If it doesn't get much past this process then report back.

[ChrisLoudon]

It could very well be your GPU is borked and causing Windows to fail on install.

 

If you can't get Windows installed with the dGPU plugged in, try again after removing the 6800XT and use the iGPU.

[RichardStones]

Just now, ChrisLoudon said:

It could very well be your GPU is borked and causing Windows to fail on install.

 

If you can't get Windows installed with the dGPU plugged in, try again after removing the 6800XT and use the iGPU.

Right, so I solved the installation, and got Windows running.

But the same issue (black screen) happened again after resetting of the PC. I can only guess it’s a HW issue now. 

[ChrisLoudon]

1 minute ago, RichardStones said:

Right, so I solved the installation, and got Windows running.

But the same issue (black screen) happened again after resetting of the PC. I can only guess it’s a HW issue now. 

Yeh,

 

it sounds like it.

 

Just to make sure it is the 6800XT, remove it and run the PC for a while on the iGPU.

 

If you can make it through a day or so without it black screening then you can be sure its a faulty GPU.

 

You can then cut through all the nonsense you'll likely get when trying to RMA the system or take it back to the shop.