3080 Randomly No Display - Requires Restart

[Daeva]

SPECS -
CPU : AMD Ryzen 7 5800x 3.8GHz
GPU : AORUS Master RtX 3080 10GB
RAM : G.Skill TridentZ 3200Mhz CL16 32GB (2 x 16) XMP Profile on
MOBO : X570 Aorus Elite
PSU : Corsair HX1000W

PROBLEM -
While I have my RTX 3080 in my computer, I will randomly lose display to both of my monitors at the same time which will require me to turn the whole system off by holding the button then turning it back on again. On the rare occasion I will get a black screen with artifacts at the bottom and it will never even boot up.

These loses of display happen most often when I am trying to play a game. It doesn't matter how graphic intensive they are as I experience them in anything I play from Warzon 2 to Skyrim. I've also had multiple instances of no display while just browsing the internet or watching movies with VLC and even times where my computer is practically idle.
Each time this happens I often see multiple instances of "nvlddmkm Event ID 14" and "Application Error Faulting application name: dwm.exe".

WHAT I HAVE DONE -

• Checked for Windows updates.
• Ran the SFC /scannow command and found no integrity violations.
• Ran the DISM /Online /Cleanip-Image /CheckHealth command and found no component store corruption.
• Reinstalled Windows entirely.
• Installed the latested BIOS and CPU Chipset versions.
• Ran the CHKDSK command and found no problems.
• Used CrystalDiskInfo to check drives. My 250GB Windows drive Health Status: Good 94% and my 1TB SSD Health Status: Good 95%.
• Ran Memtest86+. Completed 11 Passes with 0 Errors.
• Ran Aida64 for 2 hours.
• Installed the latest Graphics Driver.
• Downgraded to the previous Graphics Driver.
• Tested 5 other Graphics Driver versions.
• Used DDU to remove the Graphics Drivers entirely and repeat the previous 3 things.
• Bought a new PSU (the HX1000W mentioned, the previous was a 850W from Corsair as well).
• Bought new Displayports and HDMI cables.
• Tried different Monitors to have the same problem.
• Ran 3DMARK Benchmark twice.
• Ran 3DMARK GPU Stress Tests Speedway, Timespy and Port Royal for 20 loops each, back to back.
• I've played some games while monitoring MSI Afterburner alongside it and didn't see anything in particular.

I've also sent my GPU in for an RMA which had it sent to Gigabyte to test. After 1 month of waiting I got the same GPU back with a "No Fault Found Report" slip which stated they tested the GPU in 3DMARK for 2 hours. A lot of these tests I did both before and after I sent my GPU in to them. During the month that they had my 3080, I was using my old 1060 and never had a single problem, even playing the same games, though not at the same settings except for Skyrim.

I just ran FURMARK and lost display after about 16mins into the stress test. This time I got no "nvlddmkm Event ID 14" but I did still get the "Application Error Faulting application name: dwm.exe" and additionally "NVIDIA OpenGL Driver A TDR has been detected".
I was running GPU-Z alongside FURMARK and here are some logs for that stress test:

GPU-Z Sensor Log.txt

[GOATWD]

4 minutes ago, Daeva said:

SPECS -
CPU : AMD Ryzen 7 5800x 3.8GHz
GPU : AORUS Master RtX 3080 10GB
RAM : G.Skill TridentZ 3200Mhz CL16 32GB (2 x 16) XMP Profile on
MOBO : X570 Aorus Elite
PSU : Corsair HX1000W

PROBLEM -
While I have my RTX 3080 in my computer, I will randomly lose display to both of my monitors at the same time which will require me to turn the whole system off by holding the button then turning it back on again. On the rare occasion I will get a black screen with artifacts at the bottom and it will never even boot up.

These loses of display happen most often when I am trying to play a game. It doesn't matter how graphic intensive they are as I experience them in anything I play from Warzon 2 to Skyrim. I've also had multiple instances of no display while just browsing the internet or watching movies with VLC and even times where my computer is practically idle.
Each time this happens I often see multiple instances of "nvlddmkm Event ID 14" and "Application Error Faulting application name: dwm.exe".

WHAT I HAVE DONE -

• Checked for Windows updates.
• Ran the SFC /scannow command and found no integrity violations.
• Ran the DISM /Online /Cleanip-Image /CheckHealth command and found no component store corruption.
• Reinstalled Windows entirely.
• Installed the latested BIOS and CPU Chipset versions.
• Ran the CHKDSK command and found no problems.
• Used CrystalDiskInfo to check drives. My 250GB Windows drive Health Status: Good 94% and my 1TB SSD Health Status: Good 95%.
• Ran Memtest86+. Completed 11 Passes with 0 Errors.
• Ran Aida64 for 2 hours.
• Installed the latest Graphics Driver.
• Downgraded to the previous Graphics Driver.
• Tested 5 other Graphics Driver versions.
• Used DDU to remove the Graphics Drivers entirely and repeat the previous 3 things.
• Bought a new PSU (the HX1000W mentioned, the previous was a 850W from Corsair as well).
• Bought new Displayports and HDMI cables.
• Tried different Monitors to have the same problem.
• Ran 3DMARK Benchmark twice.
• Ran 3DMARK GPU Stress Tests Speedway, Timespy and Port Royal for 20 loops each, back to back.
• I've played some games while monitoring MSI Afterburner alongside it and didn't see anything in particular.

I've also sent my GPU in for an RMA which had it sent to Gigabyte to test. After 1 month of waiting I got the same GPU back with a "No Fault Found Report" slip which stated they tested the GPU in 3DMARK for 2 hours. A lot of these tests I did both before and after I sent my GPU in to them. During the month that they had my 3080, I was using my old 1060 and never had a single problem, even playing the same games, though not at the same settings except for Skyrim.

I just ran FURMARK and lost display after about 16mins into the stress test. This time I got no "nvlddmkm Event ID 14" but I did still get the "Application Error Faulting application name: dwm.exe" and additionally "NVIDIA OpenGL Driver A TDR has been detected".
I was running GPU-Z alongside FURMARK and here are some logs for that stress test:

GPU-Z Sensor Log.txt 758.33 kB · 0 downloads

try using windows ctrl shift b next time it happpnes

 

[GOATWD]

1 minute ago, GOATWD said:

try using windows ctrl shift b next time it happpnes

 

also try reinstaling ur gpu drivers

 

[Blqckqut]

3 minutes ago, GOATWD said:

also try reinstaling ur gpu drivers

 

he already did, at this point it just seems like a bad gpu

[GOATWD]

2 minutes ago, Blqckout said:

he already did, at this point it just seems like a bad gpu

thats y i seaid to try windows ctrl shift b as it resets drivers if it breaks again so we can tell if its a driver isue

[191x7]

58 minutes ago, Daeva said:

SPECS -
CPU : AMD Ryzen 7 5800x 3.8GHz
GPU : AORUS Master RtX 3080 10GB
RAM : G.Skill TridentZ 3200Mhz CL16 32GB (2 x 16) XMP Profile on
MOBO : X570 Aorus Elite
PSU : Corsair HX1000W

PROBLEM -
While I have my RTX 3080 in my computer, I will randomly lose display to both of my monitors at the same time which will require me to turn the whole system off by holding the button then turning it back on again. On the rare occasion I will get a black screen with artifacts at the bottom and it will never even boot up.

These loses of display happen most often when I am trying to play a game. It doesn't matter how graphic intensive they are as I experience them in anything I play from Warzon 2 to Skyrim. I've also had multiple instances of no display while just browsing the internet or watching movies with VLC and even times where my computer is practically idle.
Each time this happens I often see multiple instances of "nvlddmkm Event ID 14" and "Application Error Faulting application name: dwm.exe".

WHAT I HAVE DONE -

• Checked for Windows updates.
• Ran the SFC /scannow command and found no integrity violations.
• Ran the DISM /Online /Cleanip-Image /CheckHealth command and found no component store corruption.
• Reinstalled Windows entirely.
• Installed the latested BIOS and CPU Chipset versions.
• Ran the CHKDSK command and found no problems.
• Used CrystalDiskInfo to check drives. My 250GB Windows drive Health Status: Good 94% and my 1TB SSD Health Status: Good 95%.
• Ran Memtest86+. Completed 11 Passes with 0 Errors.
• Ran Aida64 for 2 hours.
• Installed the latest Graphics Driver.
• Downgraded to the previous Graphics Driver.
• Tested 5 other Graphics Driver versions.
• Used DDU to remove the Graphics Drivers entirely and repeat the previous 3 things.
• Bought a new PSU (the HX1000W mentioned, the previous was a 850W from Corsair as well).
• Bought new Displayports and HDMI cables.
• Tried different Monitors to have the same problem.
• Ran 3DMARK Benchmark twice.
• Ran 3DMARK GPU Stress Tests Speedway, Timespy and Port Royal for 20 loops each, back to back.
• I've played some games while monitoring MSI Afterburner alongside it and didn't see anything in particular.

I've also sent my GPU in for an RMA which had it sent to Gigabyte to test. After 1 month of waiting I got the same GPU back with a "No Fault Found Report" slip which stated they tested the GPU in 3DMARK for 2 hours. A lot of these tests I did both before and after I sent my GPU in to them. During the month that they had my 3080, I was using my old 1060 and never had a single problem, even playing the same games, though not at the same settings except for Skyrim.

I just ran FURMARK and lost display after about 16mins into the stress test. This time I got no "nvlddmkm Event ID 14" but I did still get the "Application Error Faulting application name: dwm.exe" and additionally "NVIDIA OpenGL Driver A TDR has been detected".
I was running GPU-Z alongside FURMARK and here are some logs for that stress test:

 

GPU-Z Sensor Log.txt 758.33 kB · 1 download

Latest motherboard BIOS?

Latest GPU firmware (vBIOS)?

Running XMP (DOCP)?

ReBAR ON?

Tried disabling MultiPlane Overlay?

[Daeva]

Sorry for the late response. I was running FURMARK Stress Test again and this time I was able to run it for a full hour without issues.

1 hour ago, GOATWD said:

thats y i seaid to try windows ctrl shift b as it resets drivers if it breaks again so we can tell if its a driver isue

I'll try that next time it happens

36 minutes ago, 191x7 said:

Latest motherboard BIOS?

Latest GPU firmware (vBIOS)?

Running XMP (DOCP)?

ReBAR ON?

Tried disabling MultiPlane Overlay?

BIOS Version: F37

vBIOS Version: 94.2.71.80.6

Running XMP

ReBAR is off

I haven't tried disabling MultiPlane Overlay. I'll try that now.

 

Appreciate the responses!

[AlanPorter]

33 minutes ago, Daeva said:

[Truncated]

Have you happened to try disabling Hardware Acceleration/GPU Scheduling in Windows? I've heard tell that, in rare instances, Nvidia GPUs dislike to play politely with AMD CPUs because of some motherboard BIOS coding flub and this can help.

Have you tried using Nvidia's "High Performance" mode, located in the Nvidia Control Panel? 

Finally, have you tried modifying the TDR in Windows (an option Nvidia suggests in their support page if running into TDR errors)? 

Edited February 7 by AlanPorter
Clarity

[Daeva]

On 2/8/2023 at 1:47 AM, GOATWD said:

thats y i seaid to try windows ctrl shift b as it resets drivers if it breaks again so we can tell if its a driver isue

This didn't work, however the drivers were reset when I turned the computer back on.

On 2/8/2023 at 2:33 AM, 191x7 said:

Latest motherboard BIOS?

Latest GPU firmware (vBIOS)?

Running XMP (DOCP)?

ReBAR ON?

Tried disabling MultiPlane Overlay?

Still happened with MultiPlane Overlay disabled. I will turn ReBAR on.

EDIT: I've enabled ReBAR and still get the crashes.

On 2/8/2023 at 3:46 AM, AlanPorter said:

Have you happened to try disabling Hardware Acceleration/GPU Scheduling in Windows? I've heard tell that, in rare instances, Nvidia GPUs dislike to play politely with AMD CPUs because of some motherboard BIOS coding flub and this can help.

Have you tried using Nvidia's "High Performance" mode, located in the Nvidia Control Panel? 

Finally, have you tried modifying the TDR in Windows (an option Nvidia suggests in their support page if running into TDR errors)? 

Yup Hardware-accelerated GPU scheduling is off.

I will try NVIDIA High Performance mode.

EDIT: Still got problems with NVIDIA High Performance mode on.

I only ever got that TDR error once while running FURMARK. I think this will be fine.