Windows nvlddmkm erros on startup, ONLY ON COLD BOOTS. Stable all other times.

[pagusas]

I've got a very strange problem happening. Basically I built a brand new system, 6700k, Maximus Impact VIII, 32GB 3200hz Gskil Ram and, for the time being, using an old 770 GTX for the GPU (until I can upgrade it to a 1080TI whenever they come out).


So the system works great, perfectly stable in all stress test (in both stock mode and OC'd). However one issue keeps cropping up randomly:

On cold boots (meaning its been off for a few hours) the system will load into windows fine then completely crawl to a near halt (mouse will work sluggishly and things will open, just at 1/1000th the speed of normal, until it eventaully crashes completely a few minutes later. On a reboot the system will be completely 100% stable and fine, all day long, tons of gaming and video editing, no issues. I turn the system off every night, some days she'll boot up fine without issue, some days she'll have this cold booth issue. Its driving me a bit batty.

I checked the event viewer after every cold boot issue, the constant in all reports is "nvlddmkm" errors. Specifially these ones:

"Display driver nvlddmkm stopped responding and has successfully recovered."

and

"The description for Event ID 14 from source nvlddmkm cannot be found. Either the component that raises this event is not installed on your local computer or the installation is corrupted. "


Things I have done so far to investigate/troubleshoot this issue:

Returned all system settings to stock - Still have the same cold boot issue

Reflashed the bios - just in case, no change

Upped the memory voltage by a small margin - Incase the large memory I have is starving

Ran Memtest overnight - no errors, rock stable (also ran one with it overclocked and on XMP profile, perfectly stable)

Reseated the memory and GPU - just incase anything was loose, nothing was

Changed all powermanagment settings like PCI Link stuff to disabled - just incase their was some power policy causing issues

Checked Windows for errors, reinstalled all graphic drivers (and did a full cleaning of them before reinstalling - no change


Things I still need to try:

Removing the 770gtx and running of the igpu for awhile just to see if the 770gtx is causing issues (wouldnt make sense as it was in my previous PC for years with no issue, and on anything but a cold boot it is rock stable!)

Change out Power Supplies - I really dont think my PSU is the issue, the voltages all look fine and I'd imagine the PC wouldnt even boot if it was having legit issues. Its a Lian Li 750watt SFX PSU.

Change out the memory - Again I can't imagine it being the issue given it passed memtest.

Could it be a defective board? Why would it only have issues with cold boots? Why would Windows still load (though extremely slow)?

Thanks everyone for any thoughts or input.

[Chino]

pagusas wrote:

Removing the 770gtx and running of the igpu for awhile just to see if the 770gtx is causing issues (wouldnt make sense as it was in my previous PC for years with no issue, and on anything but a cold boot it is rock stable!)

Change out Power Supplies - I really dont think my PSU is the issue, the voltages all look fine and I'd imagine the PC wouldnt even boot if it was having legit issues. Its a Lian Li 750watt SFX PSU.

Report back when you've tried the above. 🙂

[Subsider]

Open a command as admin Run Windows System File Checker ("sfc /scannow") report back
can also try:
1. Perform a clean installation of the NVIDIA graphics driver in “Safe Mode with Networking”:
341.74 NVIDIA graphics driver:
http://www.nvidia.com/download/driverResults.aspx/87988/enus
Steps on how to perform a clean installation of the graphics driver:
http://nvidia.custhelp.com/app/answers/detail/a_id/2900/
2. Restart the computer into normal mode and install the PhysX and GeForce Experience
from the links below:
NVIDIA PhysX System Software:
http://www.nvidia.com/object/physx-9.14.0702-driver.html
GeForce Experience:
http://www.geforce.com/geforce-experience/download

[pagusas]

Subsider wrote:
Open a command as admin Run Windows System File Checker ("sfc /scannow") report back
can also try:
1. Perform a clean installation of the NVIDIA graphics driver in “Safe Mode with Networking”:
341.74 NVIDIA graphics driver:
http://www.nvidia.com/download/driverResults.aspx/87988/enus
Steps on how to perform a clean installation of the graphics driver:
http://nvidia.custhelp.com/app/answers/detail/a_id/2900/
2. Restart the computer into normal mode and install the PhysX and GeForce Experience
from the links below:
NVIDIA PhysX System Software:
http://www.nvidia.com/object/physx-9.14.0702-driver.html
GeForce Experience:
http://www.geforce.com/geforce-experience/download

So far I havnt had a lot of time to work on this due to work, but I'll give an update:

I did the recommendations above, no change. Its clear to me its not a driver issue, rather a hardware issue thats likely causing the driver to get corrupt on a cold boot when loaded to memory.

So last night I started playing around in the bios some more, and instead of doing everything manually, selected the TPUII setting and let the board overclock it self. So far no boot issues based off its setting (suprised how high it took the vcore, but still staying under 70c during stress test so I can live with it). It changed a lot of system voltages higher than I did so I'll see if maybe the system just needed more juice on cold boots to stay stable. It will take a few days of testing to confirm.

[pagusas]

So ive discovered something crazy.

First, after a full day of testing i found the system fully stable, even on cold boots. So I went in to the bios to turn on some features I need enabled for quality of life, specifically the bios "wake from" area. Specifically I turned on the wake from lan and the wake from timer settings.

Here's the crazy part, the computer started experiancing the issue again!!! I turned off the wake from timer setting and the issue hasnt re-appeared! Something about the wake system in the bios is causing the computer to have cold boot issues.

Just so people understand, I use the wake feature to turn my computer on every morning at 6am so I can remote into it when needed while I'm on shoots. It turns it self off at night around midnight automatically. I'll get around this for now by using the wake on wan settings to just remote wake it when I need it.

Anyone know where I can submit my research and findings to Asus for review?