Display Crash. nvlddmkm, Event ID 4101, 14. GPUID: 100.

[Mojash]

Hi, first post here.

 

Have had this problem since January when I put together this computer.

This is my second self built computer over the past 10 years, so I am a novice however would like to think I have done some reasonable due diligence trying to diagnose the problem.

 

The computer was built with new components (minus the GPU) in January. I have consistently had Black Screens and program crashes when running semi demanding/demanding games.

 

System Specs:

Motherboard: Asus ROG STRIX X670E-E
CPU: AMD Ryzen 9 7950X (AM5)
RAM: G.Skill Trident Z5 2x16 GB DDR5-6000, installed in A2/B2 DIMM slots
Storage: Samsung 980 Pro 2 TB M.2
GPU: NVIDIA Geforce RTX 2070
PSU: Corsair RM850x 850 W

 

See attachment with the common error that occurs, also attached event files if useful.

 

I have spent the past 3 months reviewing various nvlddmkm error solutions over the internet, and not one has worked.

 

A list of solutions I have tried, but not limited to:

- Factory Default overclock/bios

- Turning XMP off

- Underclocking RAM

- Power Plan (GPU max performance)

- Turning off Hardware Accelerated GPU Scheduling

- Underclocking GPU (somewhat works, but still occurs less frequently)

- DDU reinstall of graphics drivers

- Updated graphics drivers (NVIDIA)

- Updated chipset

- Ran Graphics Benchmark tests, all run fine no problem. No stutters, no crashing, all on max settings.

- Ran Windows Memory Diagnostic Tool, no problems detected.

 

Running non-intensive graphic games poses no problem.

 

Games that have crashed:
Valheim

Hell Let Loose

Last Epoch

 

I have logged 2 example crashes using GPU-Z's text logger, and uploaded/formatted in Google Sheets if there is any useful information there.

GPU-Z reports in Google Sheets, at the time of crashes. <- CRASH LOCATION IS BOLDED IN SHEET

 

I am at a bit of a deadend when it comes to self-diagnosis and troubleshooting, any suggestions welcome.

 

Thoughts that haven't been attempted yet:
- Anything related to buying new components

- Taking apart GPU to check thermal paste etc

 

Kind regards,

 

Memory Diagnostic Event Report.evtx nvlddmkm Event Report.evtx

[Mojash]

Bump

[Mojash]

[Mojash]

Anyone have any ideas? I have no where to turn to.

[Mojash]

For anyone that finds this post.

 

I found a solution that worked, don't know why I had never considered comparing.

 

My GPU Spec states that base clock is 1400 MHz, with boost up to 1650 ish MHz. I noticed my GPU was default overclocking to 1950 MHz, along with the voltage rise expected due to the default curve f/V.

 

It must have been the default overclock settings combined with the reference voltage making the system unstable to the point the video driver/gpu cut out.

 

I have since forced undervolt, and capped voltage at 950 mV at the spec defined boost of 1650 ish MHz.

 

Since then, I have had no crashes. Please see below attachment showing f/V curve before/after.