Question Screens go off, fan speed to max, Event ID 14 from source nvlddmkm

[cachirro]

I never had any issues playing any game, but of couple of weeks ago my RTX2060 died on one of many daily gaming sessions, the computer would simply not boot anymore, long story short, i put my old GTX760 and pc was back to normal and sent the RTX to warranty which arrived back today (but i don't want to install it back until i figure this out).

After this episode i started to have issues while gaming, i could play fine for hours (counter strike) and suddenly on one more of many rounds the screens go black, gpu fan maxes out but i still hear the game playing in the back. Checking the event viewer i see the event ID 14 from source nvlddmkm or the similar one "Display driver nvlddmkm stopped responding".

I did all the possible and imaginable things to try to debug the issue, including reading many pages on the web having the same issue but nothing worked for me, until i found something that did make a difference.

When i removed all USB peripherals and left only mouse and keyboard i left CS playing alone with bots the whole night and can even run furmark to 6 minutes without crashing, while before it crashes in 3mins, 1min, depends. (it's not about the ports, i tried the removed peripherals in other ports and still crashes)

The peripherals taken out are an asus wifi antena, razer webcam and kingston pendrive, all of them work fine, this issue only occurs randomly in games, the quickest way i find to make it happen is running furmark, and i tried changing them from port (i have 6 usbs) and tried adding them one by one as well. (I managed to connect a 3rd device, razor kraken x, but if i connect the wifi or the webcam the issue will come back.)

Now i am a bit lost, does this behavior make any sense? is there a valid reason behind this? Is the problem the PSU or the MB? I wouldn't mind buying new ones, but buying without knowing if it will solve it is risky.

Hope you guys can help out. Thanks.


CPU: Intel Core i7 4790
Motherboard: Gigabyte GA-Z97P-D3 (rev. 1.1)
Ram: Dual 16Gb HyperX DDR3 800Mhz
Main SSD: Samsung 850 PRO 256Gb
Second SSD: Samsung 870 EVO 1Tb
GPU: Nvidia RTX 2060 died (GTX 760 temporarily)
PSU: Enermax Triathlor ECO 800W
OS: Win 10 64bits
Monitor: Dual Samsung SyncMaster 226BW

 

[JohnBonhamsGhost]

i see the event ID 14 from source nvlddmkm or the similar one "Display driver nvlddmkm stopped responding"...

I did all the possible and imaginable things to try to debug the issue...

this is a direct failure of Nvidia drivers.
whether you have a corrupt .dll or other data, it is likely the culprit.

there are hundreds+ ways a user could have tried diagnosing/fixing this issue.
so there is no way you actually, "did all the possible...".

my first step would be to remove all graphics related data by using DDU in Windows Safe Mode.
then reinstall the latest available driver package directly from Nvidia.

Is the problem the PSU or the MB?

no one has any clue what hardware you may be using or what state it may be in.

first rule of forum hardware diagnostics;
always include your complete system specs with make & model of all components, their time in use, and any modifications that may have been done to them.

 

[cachirro]

this is a direct failure of Nvidia drivers.
whether you have a corrupt .dll or other data, it is likely the culprit.

Hi John,
i have removed the drivers with DDU and reinstalled the lastest ones to no change on the behavior.

As for the state of the hardware, i have 0 issues in heavy intensive use on a daily basis, only instance was few weeks ago when the RTX2060 died (came back yesterday with chipset replaced in warranty) and since then i started having this issue.

Just today i played for hours while having live tv stream on the second screen, if i would plug the webcam or the wifi antena the gpu driver would have crashed.

The MB is a Gigabyte GA-Z97P-D3 (rev. 1.1) and the PSU is a Enermax Triathlor ECO 800W, but my question was more generic, does this issue that i am facing make any sense? is this a thing?

Just a note that before the RTX2060 died i had all USBs plugged and never had any issues, after the incident despite having wiped the drivers and using a new (old) GPU the issue would not go away and after dozens of things done to troubleshoot it, the only impactful one found was when USB peripherals were removed.

 

[cachirro]

update:

I plugged all USBs back on and gave another go at Furmark and to my surprise i managed to hit 5 minutes, so i stop the stress test, i close the program and few seconds later.... pufff, same issue

I unplug the webcam and antenna, run same test and after 7 minutes the temps seem to have maxed out, all looked good, i stop stress test, i run a benchmark, it finishes, no problems, all looks good, close furmark and some seconds later..... pufff

I'm starting to not know what else to try, how can this happen when the GPU is no longer on load, it's like the fact that i stopped pushing for it caused the issue to happen because while tests were running all looked good, this is so weird.

Hope someone can help somehow.

 

[JohnBonhamsGhost]

few weeks ago when the RTX2060 died (came back yesterday with chipset replaced in warranty) and since then i started having this issue...

the fact that the chipset malfunctioned could've caused some damage to the motherboard.
if there is another PCIe x16 slot, try moving the card to that open slot and see if the symptoms continue.

 

[cachirro]

i've ordered a new MB and case to replace these, yesterday i started having problems with usb sticks, when i plugged them windows said something like there's nothing connected, or damaged, while they were working fine (checked after in the laptop)

I just wonder if my windows will accept a new MB with a weaker chipset, going from Z97 to B85, but it's hard to find LGA1050 boards now, so i hope windows boots fine.

Any advice any cleanup activieties i should do on windows after changing the board?

 

[JohnBonhamsGhost]

I just wonder if my windows will accept a new MB

anytime you are changing motherboards you should be reinstalling the OS.

due to the Windows license being tied to that particular system you will need to verify the hardware update through MS support and get a new license activated for the new build.

 

[cachirro]

It's no problem to re-activate, i have all the data needed for my license, my question was more technical, if it would crash or something.
I will uninstall anything i recognize from the old motherboard and install stuff from the new one.

 

[JohnBonhamsGhost]

my question was more technical, if it would crash or something

anytime you are changing motherboards you should be reinstalling the OS.

reactivating the Windows license has nothing to do with why i stated you should reinstall the OS.
that is just a secondary statement.

there are many hardware specific drivers and settings tied to your specific motherboard that are embedded in the OS.