980ti NVLDDKM error and driver crash during games

Apathy

Gawd
Joined
Oct 29, 2004
Messages
890
I just built this new pc.

Specs:
i7 6700k
evga 980ti hybrid
corsair vengeance 16gb
dell u3011
corsair 850hx psu

I am getting this NVLDDKM error 4000 times in the last 24 hours. The computer works fine browsing youtube etc. When I try to play a game, the screen sometimes turns black for a second then will come back. Other times the driver will crash and recover. Other times the monitor will go into standby and tell me the refresh rate is not supported by this monitor. I tried using 2 different HDMI cables, 1 new, 1 old. I tried an HDMI to DVI. Same problem.

Current resolution (until my displayport cable arrives) is 1600x1200. Refresh rate is set at 60hz. I tried running Unigine Heaven and driver crashed within 5 seconds. I then tried down-clocking the GPU clock to -90mhz in Precision X and Unigine ran for about a minute or two.

The same thing happens with DayZ standalone, world of tanks, world of warships, smite, and battlefield bad company 2.

I tried a clean install of the latest EVGA current driver 358.87 with no luck.

Has anyone found a fix to this issue? I did a quick search on [H] and google and found several people have experienced this issue. I am going to try reinstalling the driver without the HD Audio driver and 3d Vision drivers selected.

*edit
Fixed title for correct error name
 
Last edited:
Wait until the test with a new DP cable. until then my two options point to PSU as first, and GPU at second option..
 
Not stable in unigine with the -90mhz gpu downclock from stock settings. Not stable with latest nvidia drivers installed with hd audio and 3d vision drivers unchecked from installation.

Forgot to mention I'm on Windows 10 Pro

Whole slew of different errors from the gfx card it looks like.
Here are some:
The description for Event ID 13 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. You can install or repair the component on the local computer.

If the event originated on another computer, the display information had to be saved with the event.

The following information was included with the event:

\Device\Video3
Graphics Exception on BE 26: CRD_CACHE_HIT_FROM_OTHER_GPC

And more:
\Device\Video3
Graphics Exception on BE 26: Graphics is hung, FATAL!!
Graphics Exception on BE 26: CRD_DEAD_TAGS\Device\Video3
Graphics Exception: ESR 0x416470=0xffffffff
Graphics Exception on BE 25: _L2_NIL_RETURN_THREAD_MISMATCH
Graphics Exception on BE 25: _L2_IL_RETURN_THREAD_MISMATCH
Graphics Exception on BE 25: _HUNG_ON_NIL_THREAD_RDAT
Graphics Exception on BE 25: ZWRRAM_ORPHANS
Graphics Exception on BE 25: Graphics is hung, FATAL!!

IRhw5Dz.png




Going to try to use the driver from last month.
 
Last edited:
No dice with older driver. Is it possible this is being caused by not using a DisplayPort cable?
 
Not stable in unigine with the -90mhz gpu downclock from stock settings. Not stable with latest nvidia drivers installed with hd audio and 3d vision drivers unchecked from installation.

Forgot to mention I'm on Windows 10 Pro

Whole slew of different errors from the gfx card it looks like.
Here are some:
The description for Event ID 13 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. You can install or repair the component on the local computer.

If the event originated on another computer, the display information had to be saved with the event.

The following information was included with the event:

\Device\Video3
Graphics Exception on BE 26: CRD_CACHE_HIT_FROM_OTHER_GPC

And more:
\Device\Video3
Graphics Exception on BE 26: Graphics is hung, FATAL!!
Graphics Exception on BE 26: CRD_DEAD_TAGS\Device\Video3
Graphics Exception: ESR 0x416470=0xffffffff
Graphics Exception on BE 25: _L2_NIL_RETURN_THREAD_MISMATCH
Graphics Exception on BE 25: _L2_IL_RETURN_THREAD_MISMATCH
Graphics Exception on BE 25: _HUNG_ON_NIL_THREAD_RDAT
Graphics Exception on BE 25: ZWRRAM_ORPHANS
Graphics Exception on BE 25: Graphics is hung, FATAL!!

IRhw5Dz.png




Going to try to use the driver from last month.

Don't. The latest driver is stable for me, and I'm also on Windows 10 Pro. First try completely uninstalling with DDU then install the driver.
 
Don't. The latest driver is stable for me, and I'm also on Windows 10 Pro. First try completely uninstalling with DDU then install the driver.

Well, the newest driver with -90mhz on gpu clock allowed me to get the furthest into unigine heaven. I'll go back to the newest driver and try down-clocking the gpu and mem.

*edit*
With -90mhz on the gpu clock and -200 mhz on the memory clock, the computer doesn't black screen and lock up forcing a restart, instead it just crashes unigine heaven after about a minute in.

Any ideas?
 
Last edited:
I'd wait on the new DP cable, see if that fixes it. If not, then the usual testing:

1. Set everything to stock with a clean Windows install and then test. If it still crashes step #2
2. Take one stick of ram out and test the system. If it still crashes, take the one out and put the other one back in and test. Also run memtest.
3. Test with alternate PSU if you have one
4. If it still fails, it could be the graphics card is defective, send it in for RMA or exchange it.
 
I'd wait on the new DP cable, see if that fixes it. If not, then the usual testing:

1. Set everything to stock with a clean Windows install and then test. If it still crashes step #2
2. Take one stick of ram out and test the system. If it still crashes, take the one out and put the other one back in and test. Also run memtest.
3. Test with alternate PSU if you have one
4. If it still fails, it could be the graphics card is defective, send it in for RMA or exchange it.

I have a feeling the cable is the problem even though I've tried two different HDMI cables and an HDMI to DVI. I remember my u3011 having an issue with anything but a displayport cable in the past. Unfortunately my old 6870 used a mini displayport to display which the 980ti doesn't have a port for.

I just tried using DDU and reinstalling drivers (no hd audio / 3d vision drivers) along with removing precision x and oc scanner from evga. Made it further into unigine heaven demo but it still crashed and brought me back to desktop.

Why do you think the DP cable is the solution just out of curiosity?
 
As I see it you got 2 ways to go.

Either bios mod and bump the voltage for each bin, thats what got my 970 stable. EVGA were too aggressive on the voltages.
Or RMA the card.

In my case the card was basically faulty in the way it was sold it was not high enough quality to be stable at the clock speeds vs voltages EVGA set to. It is fine now tho on my custom bios.
 
Ok... so I just reinstalled precision x and turned up the voltage slightly which did nothing at stock clock settings. Then I turned the stock clocks down as much as I could and set the power target at the lowest possible setting and I made it through a full loop of unigine heaven.

Any idea what that means?

tDNZFI9.png
 
modding the voltage via software sadly is nowhere near as good as in the bios.

But your tests confirm what I said, the card isnt stable at the stock settings, so either it needs RMA'ing or you can fix it yourself in the bios.

If bios editing isnt something you want to do then RMA it as faulty, unstable at stock configuration.
 
modding the voltage via software sadly is nowhere near as good as in the bios.

But your tests confirm what I said, the card isnt stable at the stock settings, so either it needs RMA'ing or you can fix it yourself in the bios.

If bios editing isnt something you want to do then RMA it as faulty, unstable at stock configuration.

Heading to the bios... be right back.

*edit*
can't find voltage control for gpu in bios. what would it be named other than gpu voltage?
 
Last edited:
Heading to the bios... be right back.

*edit*
can't find voltage control for gpu in bios. what would it be named other than gpu voltage?

You have to use Maxwell BIOS Editor, edit the GPU bios you obtain via GPU-Z, make sure you keep a backup copy, then flash using NVFlash.

I suggest you RMA the card instead of doing this. It's gonna cause you less troubles and you'll actually be able to overclock with the new card (hopefully).
 
You have to use Maxwell BIOS Editor, edit the GPU bios you obtain via GPU-Z, make sure you keep a backup copy, then flash using NVFlash.

I suggest you RMA the card instead of doing this. It's gonna cause you less troubles and you'll actually be able to overclock with the new card (hopefully).

Yea i was just reading that I'd have to flash the GPU bios. I think I'm going to wait for the displayport cable, and if that doesn't fix it, and at this point I don't think it will, then I'll RMA the card.

Ah the joys of building from scratch. At least it's not the mobo.

It's interesting, when I turn the clocks down on the card and adjust the power target to 85% in precision x, the frame rates are 100 - 200 fps in battlefield bad company 2. But if i switch that power target just 1% below that to 84%, the frame rates drop to 15-25 fps. Strange.
 
at this point as i've said early, just wait for the new DP cable. if the issue persist then the issue it's the card or the PSU, a faulty or degraded PSU that can't handle the load (which its a lot of power on a high clocked 980TI) can point to your issue..

I would just RMA the card.
 
I have a feeling the cable is the problem even though I've tried two different HDMI cables and an HDMI to DVI. I remember my u3011 having an issue with anything but a displayport cable in the past. Unfortunately my old 6870 used a mini displayport to display which the 980ti doesn't have a port for.

I just tried using DDU and reinstalling drivers (no hd audio / 3d vision drivers) along with removing precision x and oc scanner from evga. Made it further into unigine heaven demo but it still crashed and brought me back to desktop.

Why do you think the DP cable is the solution just out of curiosity?


I've had issues with HDMI cables in the past but not the kind of crashes you mention so that's not really high on my list of probabilities.

Do you have GeForce Experience installed? If so, there's an issue I ran into with the NVIDIA game streaming service that forces my Titan X's voltage to stock so when they hit 3D clocks, games crash. To get around this, I had to disable coop game streaming in GFE (just uncheck the box) and then restart your PC or kill the service manually. If you don't have GFE installed, then do what I listed above once you have a DP cable handy.
 
I picked up a displayport cable. No change to crashing behavior. I disabled streaming in GFE, tested with both sticks of ram pulled individually, played with precision x by increasing voltage and down-clocking. Still no change. The PSU is brand new and yes I know it can be the source of the problem, but I have a strong feeling it's the gpu.

Time to RMA.
 
I picked up a displayport cable. No change to crashing behavior. I disabled streaming in GFE, tested with both sticks of ram pulled individually, played with precision x by increasing voltage and down-clocking. Still no change. The PSU is brand new and yes I know it can be the source of the problem, but I have a strong feeling it's the gpu.

Time to RMA.

If you're confident it's not the PSU or any other software conflict, then RMA seems prudent.
 
This is a clear case of EVGA selling you a card they claimed would run with an OC at a set voltage... and it won't. Email them, and tell them to go to hell. But nicely.
 
This is a clear case of EVGA selling you a card they claimed would run with an OC at a set voltage... and it won't. Email them, and tell them to go to hell. But nicely.

that's not the problem if not then he would have fixed it with the Underclock and overvolting as he did tried already and the problem persist "played with precision x by increasing voltage and down-clocking. Still no change." it's a matter of a faulty card, those things just happens.
 
Card is in a box, dropped off to UPS, and on it's way home and hopefully to the graveyard. Thanks for all the replies [H]. I'll post back and resurrect this thread once the new card arrives.

p.s.
Thank you silent-circuit for all the build help in the past with the previous pc's in my sig.

*edit*
Two passes with zero errors in memtest86.
 
Last edited:
yeah I probably wouldnt have tried the bios mod if I had a normal card, but my FTW has multiple bios's so if I screwed it up all I would have had to do is flick a switch to get it working again.

But for sure bios modding opens up all sorts of stuff, it makes software overclocking obselete in my book.

In my view tho if its not stable at stock and needs a bios mod to make stable, its faulty so RMA would be valid.
 
In my view tho if its not stable at stock and needs a bios mod to make stable, its faulty so RMA would be valid.

I'm just hoping it's not a software issue or a hardware conflict. Going to run prime 95 (v27.9) to check cpu now.

Is there a recommended software to test psu?
 
that's not the problem if not then he would have fixed it with the Underclock and overvolting as he did tried already and the problem persist "played with precision x by increasing voltage and down-clocking. Still no change." it's a matter of a faulty card, those things just happens.


my card I couldnt get fully stable using precision X, he is right.

What I did to fix my problem is go into the bios and I basically did the following.

1 - cap the boost speed (the evga cap is really high effectively only limited by power and thermal limits).
2 - I raised the power limit by increasing the power input limits over the 2 pci-e power cables, this more than compensated for the increased voltages required for each clock speed.
3 - I raised the voltage that each clock speed needs, so each individual bin I raised the voltage limit (there is dozens of them), note I only touched 3d clocks not 2d clock voltages. I raised each bin by about 0.13mv and that was enough to make the card stable.
4 - I also raised the base 3d clock from 1215 to 1291 which also means the base 3d voltage is higher as well, this meant less voltage fluctuation between min and max 3d speeds.

Software management gave me the following issues.

1 - when I raised the voltage limit the card was more stable in games but still having issues.
2 - when I raised the voltage limit for some reason netflix wasnt stable in full screen mode, I had artifacts and other issues. (the voltage control in precision X includes 2d clocks)
3 - raising the power limit just made it boost higher it didnt make the card any more stable, although lowering it to 90% made it almost completely stable but the clocks were pretty bad.

I also found in the stock state the card throttled so much on high load that the speed went below the advertised evga speed. these cards seem over aggressive on power limits and voltages to the point not all cards work properly.
 
I'm just hoping it's not a software issue or a hardware conflict. Going to run prime 95 (v27.9) to check cpu now.

Is there a recommended software to test psu?

by the way if you willing to do the bios dump part and the flash part I can mod your bios for you.
 
by the way if you willing to do the bios dump part and the flash part I can mod your bios for you.

I'm very curious to know whether this would have resolved my problems or if my card was just defective. If the situation repeats itself, I suppose a bios mod may be in order if the PSU is determined to be in good working order.
 
ahh yeah I just noticed you already actioned the RMA, probably the best choice, good luck on the replacement.
 
ahh yeah I just noticed you already actioned the RMA, probably the best choice, good luck on the replacement.

Thanks dude. Fingers crossed. I got in touch with newegg live chat and they put a "rush order" on the replacement process which switched my 3-5 day return shipping to next day delivery. Still love newegg.
 
p.s.
Thank you silent-circuit for all the build help in the past with the previous pc's in my sig.

...and here I thought no one would recognize me after all this time. No problem man, hope they've been working out for you.
 
Good news. Too impatient to wait for RMA. Went to local b&m and picked up an EVGA 980ti Classified. Made it all the way through unigine heaven without error. Going to test some games now. Looks like it might have been the card.
 
Good news. Too impatient to wait for RMA. Went to local b&m and picked up an EVGA 980ti Classified. Made it all the way through unigine heaven without error. Going to test some games now. Looks like it might have been the card.

Glad to know, it was pointed to PSU or GPU since the first post. :D keep testing and report.
 
Back
Top