Hello!
 
At the start of this year I bought a 3060ti xc for a sff build. Sadly, from the beginning I've been experiencing issues with it.
I've finally ran out of all options and I hope someone here can identify my issue.
 
When playing games or using other graphics intensive apps, the application will crash. This is no BSOD crash, but a crash to desktop.
Usually this is within the first 5-10 minutes of playtime, causing a lot of frustration. The error I get always seems GPU related. They are the following:
 
Unreal engine games:
Unreal Engine is exiting due to D3D device being lost. (Error: 0x887A0006 - 'HUNG')
DXGI ERROR DEVICE HUNG
 
Microsoft flight simulator
Faulting application name: FlightSimulator.exe, version: 0.0.0.0, time stamp: 0x60c0b1e3
Faulting module name: VCRUNTIME140.dll, version: 14.29.30035.0, time stamp: 0x608dc4a3
 
3D mark:
Fault bucket LKD_0x141_Tdr:6_IMAGE_nvlddmkm.sys_Ampere, type 0
Event Name: LiveKernelEvent
Response: Not available
Cab Id: 834ff0c4-5447-4a9d-88a6-ea5ef3ffbe57

 
I've tried the following to fix the issue:

1. Remove drivers in safe mode with the DDU tool, install latest official nvidia drivers
2. Enable re-bar technology
3. Increase TdrDdiDelay using regedit
4. Install all latest updates
5. Reseat hardware
6. Check temps (max. 70 degrees, hotspot 80-85 degrees), Also forced fans to run 100% once, which brought the temps even lower, but still crashed.
7. Enable debug mode in nvidia control panel
8. Underclock GPU
9. Ran Memtest86 for multiple passes, multiple times (0 errors)
10. Try different graphics settings in games, and different refresh rates for monitor
11. Try different display port outputs
12. Validate game files, reinstall games

My hardware is the following:
Motherboard: MPG B550I GAMING EDGE WIFI
Memory: Corsair Vengeance LPX CMK64GX4M2E3200C16
PSU: Corsair sf600
CPU: AMD 5600x
 
Additional, pc is small form factor, gpu is connected over a pcie gen3 riser cable. pcie is set to gen3 in motherboard.
When doing office work, or programming or any other heavy cpu workload, the pc is absolutely fine.
 
I hope someone has any more suggestions to resolve the issue. I really don't want to go to the struggle of RMA.
Thanks in advance for your read and help!
 
Greetings,
Jelle
 
 

vschaefer
I'm thinking bad gpu, but I would urge you to try another riser cable for the hell of it.   Or breadboard everything and bypass the riser and put it into the actual slot.
 
How old is the psu?

Thanks for the suggestion, ill try to put my gpu directly onto the motherboard soon, it takes some effort to "unbuild" everything. Everything in the pc is new, I started from sratch, no component is older than 2021.

FoTacTix

jelknab96
 
Thanks for the suggestion, ill try to put my gpu directly onto the motherboard soon, it takes some effort to "unbuild" everything. Everything in the pc is new, I started from sratch, no component is older than 2021.

Let us know how it turns out...I'm curious

Sadly, plugging the GPU directly into the motherboard results in the same errors happening. Tested it with PUBG and 3d mark, nothing changed. The suggestion however was solid, thanks for that.
Just changed to pcie gen4, also noticed amd curve optimizer was on in the bios and disabled it, clocked my memory down to 3200 -> 3000MT/s. So far so good, will keep testing and keep you updated.
 
The gpu has completed a 20 minute 3d mark stress test, a full pubg session and completed a full cuda memory test.

 
Really hoping this is not just good luck, will try to return settings as they were to find the fault and post it here.