4090 PC Crashes when the display goes to sleep and wakes up.

[burakksglu]

 

Issue:
I’m experiencing crashes and display issues with my PC under certain conditions:

1. When the display goes to sleep and wakes up, the PC sometimes crashes.
2. Using the profile buttons on my Samsung G9 Neo causes the display to go black and lose signal occasionally.
3. If I put the PC to sleep from the Windows 11 menu, it sometimes restarts from the BIOS instead of resuming normally.

Setup:

• Asus RTX 4090 Strix OC
• Ryzen 7800X3D
• MSI X670E Tomahawk (BIOS version 1.E0 with AGESA Combo 1.2.0.0a)
• 6000MHz CL30 DDR5 RAM
• 1000W 80+ Gold PSU (FSP Hydro)
• Windows 11 24H2
• Samsung G9 Neo 49" (DP) + Samsung CHG70 (DP/HDMI)

Errors:
I’ve noticed these errors in the Windows Event Viewer:

• Event ID 14 (nvlddmkm):

  The description for Event ID 14 from source nvlddmkm cannot be found. Either the component that raises this event is not installed on your local computer or the installation is corrupted. You can install or repair the component on the local computer.  
  
  If the event originated on another computer, the display information had to be saved with the event.  
  
  The following information was included with the event:  
  \Device\000000d5  
  02176a74 00000000 00000000 00000000 00000000 00000000 00000000 00000000  
  
  The message resource is present but the message was not found in the message table.  
  

• Event ID 153 (nvlddmkm):

  The description for Event ID 153 from source nvlddmkm cannot be found. Either the component that raises this event is not installed on your local computer or the installation is corrupted. You can install or repair the component on the local computer.  
  
  If the event originated on another computer, the display information had to be saved with the event.  
  
  The following information was included with the event:  
  \Device\000000d5  
  UCodeReset TDR occurred on GPUID: 100  
  
  The message resource is present but the message was not found in the message table.  
  

• Another Event ID 153 (nvlddmkm):

  The description for Event ID 153 from source nvlddmkm cannot be found. Either the component that raises this event is not installed on your local computer or the installation is corrupted. You can install or repair the component on the local computer.  
  
  If the event originated on another computer, the display information had to be saved with the event.  
  
  The following information was included with the event:  
  \Device\000000d5  
  Error occurred on GPUID: 100  
  
  The message resource is present but the message was not found in the message table.  
  

Troubleshooting so far:

1. Cleanly uninstalled/reinstalled NVIDIA drivers using DDU.
2. Reseated GPU and PSU cables, cleaned gold contacts.
3. Removed USB hub cable between G9 Neo and PC.
4. Reset GPU to default settings (removed overclocks/undervolts).
5. Changed display connections (spaced out inputs, switched CHG70 to HDMI).
6. Switched out display cables, reducing screen blinking on G9 Neo.
7. Adjusted BIOS settings (SOC voltage to 1.2V, curve optimizer for CPU voltage).
8. Tested both optimized RAM timings from "Actually Hardcore Overclocking," default EXPO profiles, and MSI X670E Tomahawk's suggested timings.
9. Disabled hardware acceleration in Windows 11 settings.

Despite these efforts, the issues persist. Any insights or advice would be greatly appreciated!

 

[freeagent]

Did you test your curve? Did you test your memory settings?

 

Your system isn't stable, and that is why your driver keeps crashing.

[burakksglu]

14 minutes ago, freeagent said:

Did you test your curve? Did you test your memory settings?

 

Your system isn't stable, and that is why your driver keeps crashing.

I don't even crash under any other condition, though. I've no issues with any games. Also, I remember testing these settings for hours with CoreCycler, y-cruncher, 3DMark, and Furmark. Right now, I'm using loser timings with default EXPO and lower curve.

 

Thanks for the response.

[GOTSpectrum]

So I had a similar issue with one of my 4070tis when I first got it and it turned out to be a manufacturing defect that caused issues when it moved to lower power profiles. I got an RMA for it and never had the issue again 

 

Not saying that is the issue here, but that's the only time I've had this happen other than unstable overclocks on the GPU 

[burakksglu]

So I've tested my memories again as it is suggested. In windows it looks alright.

 

Start:


Later:

[freeagent]

I use TM5, with the Absolut, or Extreme profile and no swap file.

[burakksglu]

So there's an update. I might be on to something.

I have deactivated Process Lasso from startup and turned off Idle Saver mode to prevent my power plan from switching when the PC is idle.

I removed the second USB hub from the CHG70, which previously connected my old Android tablet that acts as a stream deck using TouchPortal. I connected the tablet directly to the PC via the motherboard's USB ports on the back.

So far, there have been no crashes or issues when waking up from idle state or sleeping displays. As a result, I can now hear the "G-Sync compatible device is connected" notification from Nvidia without any crashes.

Tomorrow morning, I plan to test the idle saver or power saving mode to determine if the issue reoccurs. The PC has never crashed within 5-10 minutes; it typically crashes after being idle for some time with the display asleep.

[burakksglu]

Leaving the PC on overnight resulted in no crashes. The display was even turned off by itself, but I was able to turn it back on and continue the Windows session without issues.

 

It seems the problem arises only when the power saving mode is on or when Process Lasso's Idle Saver feature is activated.

[burakksglu]

On 12/1/2024 at 1:21 PM, burakksglu said:

Leaving the PC on overnight resulted in no crashes. The display was even turned off by itself, but I was able to turn it back on and continue the Windows session without issues.

 

It seems the problem arises only when the power saving mode is on or when Process Lasso's Idle Saver feature is activated.

Well, the issue came back again.

I updated the latest bios for my motherboard.
Removed any RAM overclocks just use the 4 EXPO values and others are all set to auto. 
RAM voltages are set to 1.360V.
GPU is always set to default values.

Started to crash into black screen in Marvel Rivals also. It crashes to black screen and fans go crazyyy... Never happened before Mobo BIOS update.

Contacted ASUS support to see what they'll suggest after all these information and crash dumps. I also collected minidump and debugged them. The issue seems to be GPU or RAM access/management for sure.

Attaching them here for information purposes.

CrashDumps.zip

[burakksglu]

UPDATE:

It looks like it was a driver issue with Nvidia. I'm not having any issues with the hotfix driver 566.36.

Steps I followed:
- DDUed older drivers, from 572.16 to 572.85 all were problematic.

- Removed the Nvidia App as well.

- Changed PCI-E gen version from Auto to Gen4 in X670E Bios

 

Right now, there are no issues so far from waking up from the display's sleeping

[jamyjet]

Currently having this issue on my 5090 build with the latest drivers (obviously I can't roll back to 566). PC is completely stable while in use but if I swap video inputs or turn my monitor off, the pc will sometimes freeze up with no input or just restart on it's own when turning the monitor back on or changing inputs to the gaming pc again.