Help with nvlddmkm stopped responding windows 10 Inno3d ichill x3 1080ti

May 21, 2018
3
0
10
Good Morning,
I'm having some problems with the GPU that I bought last year in October an inno3d ichill x3 1080TI
Could somebody please help me to further troubleshoot the problem.

The GPU driver keeps crashing to desktop while playing games, after 5 to 45 seconds.
There is no error message on screen and the game freezes, audio continues for about 2-4 seconds and than, most of the times the game exists to desktop. There are times when a ctr+shift+escape is needed to force stop the game's exe.
On rare occasions the system just freezes while the game's audio keeps going and a hard reset is needed.

Via windows event viewer I found this error:

"

Display driver nvlddmkm stopped responding and has successfully recovered.

"

In a normal non-stress environment the GPU works just fine and the system is stable
Under stress while trying to play games, the game crashes to desktop. The GPU temp reported by MSI Afterburner before crash is around 80 degrees Celsius
The games that I have now and tested with are:
Witcher 3 (worked fine after buying the card, now crashes)
Wolfenstein 2, The new colossuls (worked fine after buying the card, now crashes)
Far Cry 5 - crashes constantly
Assassins Creed Oringin - crashes constantly
COD - WW2 - crashes less often but still crashes and gives this message: DirectX Encountered an unrecoverable error . In window's event viewer there still is the nvlddmkn event.

System is:
GPU inno3d ichill x3 1080TI
win 10 pro edition updated to the latest updates.
aorus gaming7 rev 1.0 gigabyte mb
i5 8600k cpu
2x HyperX Predator, 8GB DDR4, 3000Mhz, CL15, 1.35v
Corsair CS750M PSU
latest nvidia geforce drivers

Steps taken to remedy the problem:
- updated the nvidia driver via geforce experience
- uninstalled the nvidia driver in windows safemode, disabled windows driver auto-install using DDU, than installed the latest driver via nvidia installer
- uninstalled the nvidia driver in windows safemode, disabled windows driver auto-install using DDU, than installed the latest driver via nvidia installer without geforce experience
- uninstalled the nvidia driver in windows safemode, enabled windows driver auto-install using DDU, windows downloaded and installed a driver
- taken all the steps above regarding the video drivers using older drivers (rolled back as far as december 2017)
- replaced the PSU and all the power cables with a new PSU (FSP HYDRO G 750, 750W, 80+ Gold)
- unmounted the GPU, cleaned all the dust inside the unit with compressed air, mounted back the GPU
- updated the MB to the latest bios
- reset - ed all the bios setting to "Load factory defaults"
- run windows sfc command to repair system files
- reinstalled all direct x files
- reinstalled all Viusal C++ redistributables
- changed all the power settings (both in windows and nvidia control panel) to hi performance
- changed the system's registry values for TdrDelay to 8, than 10, than 20 (HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\GraphicsDrivers)
- formatted system drive and installed a fresh windows
- down clocked the GPU clock and memory clock using MSI Afterburner as far as -700 mhz
- uninstalled MSI Afterburner altogether
- selected way lower settings in-game or with geforce optimize, disabled anti-alising, sync, lowered the resolution from 4k to 2, or 1k

All of the actions above resulted in no change in the system's behavior while trying to play the games mentioned in the beginning.

Also I stress tested the GPU with Frumark but everything is working smoothly with no crashes. I noticed that FurMark is stress testing using OpenGL while my problems come from d3d11

My next step will be to take out the GPU and install it in my work system that uses the same windows and also to memtest the ram. If the problem persist I'll need to RMA the GPU since it seems to be a hardware issue.

Are there any other steps to take? Can somebody help me with debugging windows errors error files(dumps) or even send me in the right direction to fix this?

Thank you !!!
 
May 21, 2018
3
0
10


Thanks! I never checked this! This seems to be a good point. I'll need to identify the rams properly (not at home now) and check with gigabyte qvl list.
I'll post back regarding the memory compatibility
 
May 21, 2018
3
0
10
OK. Finally, after hours of testing, fiddling and cursing I found a solution that works in my case:

The core clock of the GPU was under-clocked by 11 mhz (started from -50 and than raised it until it crashed at -10) BUT the trick here, that makes no sense to me, was to use EVGA Precision X, since a week ago I did the same thing using MSI Afterburner with no success.
Also increased power limit to 120% and temp limit to 90 deg C (just to be sure). Now it runs fine at around 1936mhz 40 -55 fps 82 deg C in ACO 4k, all on ultra. No crashes.

I have no idea why precision x does the trick and afterburner does not. Not one second I thought to switch from afterburner to xoc So yeah ... there you have it !!! God bless the scrambled ways of programming !!!