PC instability: video freezes while gaming, LiveKernelEvent 141

[Etheles]

Hey folks,

 

I've been having instability while gaming on a newly-built computer, running a new installation of Windows 10. On several different games, but notably Genshin Impact and Satisfactory, the video freezes (you can often hear ambient game sound) and the game crashes, throwing a "Display driver nvlddmkm stopped responding and has successfully recovered" error. When checking in the Security and Maintenance panel, I see a critical error: hardware failure, code LiveKernelEvent 141, corresponding to each event. I can reproduce the problem with 100% success rate on a particular area in Genshin Impact (heavy rain, lots of water). The crashes seem to be most severe in 4K / high detail mode.

 

You can view the complete part list here: 

PCPartPicker Part List

Type	Item	Price
CPU	AMD Ryzen 9 5900X 3.7 GHz 12-Core Processor
CPU Cooler	Noctua NH-D15 CHROMAX.BLACK 82.52 CFM CPU Cooler
Thermal Compound	Thermal Grizzly Kryonaut 1 g Thermal Paste	$8.92 @ Amazon
Motherboard	Asus ROG Crosshair VIII Dark Hero ATX AM4 Motherboard	$728.95 @ Amazon
Memory	G.Skill Trident Z Neo 32 GB (2 x 16 GB) DDR4-3600 CL16 Memory	$228.99 @ Newegg
Storage	Western Digital Black SN850 2 TB M.2-2280 NVME Solid State Drive	$379.99 @ Newegg
Video Card	Asus GeForce RTX 3080 10 GB ROG STRIX WHITE OC Video Card
Case	Fractal Design Meshify 2 Compact White TG Clear Tint ATX Mid Tower Case
Power Supply	Corsair HX Platinum 1000 W 80+ Platinum Certified Fully Modular ATX Power Supply	$448.00 @ Amazon
Operating System	Microsoft Windows 10 Home OEM 64-bit
Case Fan	Corsair LL120 63 CFM 120 mm Fan	$39.99 @ Amazon
Case Fan	Corsair iCUE QL120 41.8 CFM 120 mm Fans 3-Pack	$121.27 @ Amazon
Case Fan	Corsair iCUE QL140 50.2 CFM 140 mm Fans 2-Pack	$118.99 @ Amazon
Monitor	Asus ROG SWIFT PG27UQ 27.0" 3840x2160 144 Hz Monitor	$3675.84 @ Amazon
Keyboard	Varmilo VA87MNN/EBW3 Wired Standard Keyboard
Mouse	Logitech G903 Wireless Optical Mouse	$114.50 @ Amazon
Speakers	PreSonus Eris E4.5 50 W 2.0 Channel Speakers	$199.95 @ Adorama
Custom	Noctua NA-HC4 chromax.white heatsink covers (for NH-D15, NH-D15S & NH-D15 SE-AM4)	$29.90 @ Amazon
	Prices include shipping, taxes, rebates, and discounts
	Total	$6095.29
	Generated by PCPartPicker 2021-04-14 09:01 EDT-0400

(I have no idea what's up with some of the prices on this list, neither the PSU, motherboard nor monitor cost nearly that much).

In addition to these, I'm also using two other M.2 drives (Samsung EVO 970): two of them plugged into the motherboard, and the remaining one onto an ASUS PCIe M.2 expansion card. Of these components, I've only brought the two Samsung drives and the monitor from an older computer, which did not have these issues.

 

I researched the issue extensively and have tried literally everything I could think of, and all the advice I found online. I'm hoping someone can see something I'm missing. If I'm still unable to diagnose, I'll likely have to RMA the motherboard and/or GPU.

 

Things I've already tried:

Physical solutions:

- Connecting the computer to a different power outlet

- Removing optional components like the M.2 PCIe expansion card, unplugging the wireless antenna, and the speakers

- Reseating the memory

- Reseating the GPU

- Purchasing and installing a Ryzen 3600x to test whether the 5900x was the problem (hope I can return it easily)

- Disassembling the computer and then building it all over again on a separate case (from Meshify 2 Compact to Lian-Li Lancool 205) using a completely different set of PSU cables, case fans and case fan controller (another, separate Corsair Commander Pro because I wanted a similar 5-fan array to eliminate inadequate cooling as a problem, and the motherboard only has 3 chassis fan terminals)

- Connecting only the PWM fan terminals, and leaving all the RGB cabling on the case fans unplugged

 

Software solutions:

- Reducing the speed of the memory from 3600 to 3200, and later disabling DOCP/XMP entirely

- Running memtest86 to check if the memory's working properly
- Disabling AMD's Precision Boost Overdrive and Core Performance Boost
- Changing Window's power settings to Balanced
- Disabling Windows Fast Boost
- Disabling all non-Windows services

- Underclocking the GPU first by 30, then by 100 MHz

- Disabling G-SYNC

- Uninstalling all ASUS software, including AURA
- Uninstalling iCue
- Booting in Safe Mode

- Uninstalling all GPU drivers with DDU, then reinstalling the newest GPU drivers

- Formatting C: drive and reinstalling Windows from scratch

- Installing the game on two other drives (including C: drive) and testing from those installations
- Rolling back to an older motherboard chipset software, then back to the newest one when that didn't seem to fix anything
- Installing the newest motherboard firmware

- Testing GPU on OCCT

- Testing Power on OCCT 

- Monitoring temperature spikes in CPU or GPU when the crashes happen

 

The system seemed more stable after rebuilding it on the Lancool 205: I could clear the encounter that would reliably crash my system 3 times in a row with no issue, but later I still experienced crashes while idling on that zone or in other parts of the map.

 

If you have any notion of what could be the problem or what else I could try, please help me.

[Etheles]

Update:

 

I've ordered replacement components for the parts I haven't tried to swap out yet, and ran into an interesting discovery.

First, I've done my best to crash the system today and haven't been able to. The only change I've made since yesterday was turning off HDR, decreasing the display's refresh rate to 98 Hz (highest that I can get 10 bit colour) and turning off G-Sync all together. When pushing things higher (120 Hz plus G-Sync) I'm getting periodic flashes of the screen going black.

 

I'm starting to wonder if a problem with the display's G-Sync module might be the cause of all of this, but I can't evaluate how likely that is. For one, games were crashing even while they were minimized. Does anyone know if that would be possible?

 

Edit:
Was able to crash with G-Sync enabled. I'll leave the game running in the background without G-Sync and see whether it crashes.

[Etheles]

Update:

 

It does look like G-Sync was the problem. Now you know: if you get display driver crashes, check that G-Sync isn't the culprit before you go to drastic solutions, like I did.

[Etheles]

Update:

Nope, it was the GPU. Got a replacement 6800 XT and there were no longer any issues. Sent the 3080 on RMA.