Display driver nvlddmkm stopped responding and has successfully recovered
 

Okay, I'm at my wits end here so I'm going to reach out for any chance of sage knowledge about this issue, specifically related to overclocking a Core i7 CPU. I've only had this issue with Core i7's.

Is there some method to fix this or BIOS setting that might?

So far I get this Event log message from both games and F@H, although more often from Folding@home.

I've tested with two CPUs, two motherboards, two brands of RAM, three NVIDIA GPUs, multiple drives/OS installs, and two power supplies. I still get driver reset messages with all of them, and the only common factor seems to be my overclocks.

You guys have experience taking hardware to the limits and running 3D benchmarking programs, so I'm guessing this problem has been run into before and worked around, somehow?

I had this problem in two situations till now : crappy drivers and needed a reinstall or video card VRM problems (the video card did not display any artifacts, but the screen was going blank for a bit and when it was coming back I was getting that error; sometimes I was getting the error without the blank screen).

After replacing the video card with an identical one, the problem was gone.

What video card/cards are we talking about? If we are talking about the new GTXes, there are problems with BIOSes from multiple manufacturers with these cards, and some promised to release BIOS fixes.

Hey Stefan! Thanks for the reply.

EVGA GTX 260 (original + fully watercooled)
EVGA GTX 260 Core 216
EVGA GTX 285

The screen flips black when it resets. Causes a 2-5 second game freeze if a game is playing, or it corrupts the F@H work unit that was folding. I've had this problem ever since I built the system, but always thought it was for some different reason or another because it is intermittent. A few times it gets so bad I can't play a game, others it doesn't show up for days.

I would have tried an ATI card, but until Stanford fully switches to a universal GPGPU codebase ATI's folding performance is still underwhelming at best. And at this point I don't think it would fix it anyway...

Dear Kougar,

No problem, just trying to help. I know the folding applications stress a lot the video card and in many cases you can hear the so called "coil noise".

Have you also tried older video drivers to see if you have the same problem? Also, what OS do you use? Windows 7, Vista? If using windows 7, I would also try Vista to see if the problem manifests again.

On the non-watercooled cards, also try to run the fan at a higher speed, not AUTO, to see if its not an overheating problem.

Yes, this issue irritates me greatly. Out of the three cards I listed the only card that never made noise while folding was the card I custom watercooled... the analog power circuitry in the 285 and Core 216 are extremely loud under Folding@home loads, the digital VRM on the original GTX 260 seems to have made a difference there.

Yes, it's not OS specific. I use Windows 7 64bit mainly.

I tried maxing out the fan speed on the 285, but GPU temps are not the issue. Even with the fans on auto Furmark 1.8.0 doesn't overheat these cards or force the driver crash.

Have you also tried to install other older drivers? (by using also a driver cleaner prior to installation)

I've had this problem since the 180 drivers onwards. Yes I've tried driver cleaner and driver sweeper.

If you deactivate PhysX in the drivers, will the problem persist?

I deactivated PhysX. So far so good... if the problem doesn't reoccur within a week I can be sure your suggestion resolved it. I'll cross my fingers on this one.

I really hope the problem gets solved. Damn Nvidias' crappy driver programmers.

Please update me in this thread.

Okay, I can't take this. The sooner I tell you that fixed the problem, the sooner it will crash again. So.... it fixed the problem. :D

*waits for it*

You're funny :D

I wish I was!

Going by the event log warning, the driver crashed while I was away. Took it a little longer than usual.

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

Are you LinX stable with those CPU overclocks?

Yes. 25 runs with 4700 problem sizes. Also Prime stable. I do 24/7 Bigadv Folding@home which is sensitive to any sort of instability.

It is extremely rare, but I have experienced this problem when at stock, so I am sure it isn't the overclock. But the overclock does increase the frequency of the problem, so it is related somehow. I know it isn't a problem with F@H too because I sometimes experience it when playing a game.

I've been hitting my head against a wall trying to fix this problem for over a year... :??

Have you tried updating DirectX with this installer?

http://www.microsoft.com/downloads/d...displaylang=en

If this does not fix the problem, try the CUDA drivers:

http://developer.download.nvidia.com...13_general.exe

Hey Stefan. The DirectX installer didn't find anything to update, and the 197.45 drivers include CUDA 3.0.1 in them. I did try your drivers but they errored quickly and caused this (which I occasionally do see)

http://i111.photobucket.com/albums/n.../IMG_0889s.jpg

Those snowflakes are you talking about on the screen?

Yes. It's a rendering problem... if I take a screenshot of the page they don't show up, so I have to photograph the entire display to show them. When it occurs 3D applications error constantly until I restart the system.

Frankly it reminds me of failing GPU cores, you can get the exact same symptoms. These "snowflakes" appear and disappear randomly in random locations as the screen image changes

Do those 'snowflakes' appear on those other cards you own too?

I honestly do not remember. But the driver crashes are the same on all three.

This card has been the least "stable" in my system out of all three, if I could I would go back to my original watercooled card as it didn't have coil noise or high frequency whines/chirping. I refuse to use the GTX 285 as the noise is so loud and irritating I can hear it over music, the air conditioner, and even the TV.

NVIDIA uses the cheapest components possible on their PCBs to offset their higher silicon costs, starting by removing the digital VRM to save more on production costs... they have cut way to far. All they seem capable of building are crappy PCB designs anymore. Not even the GTX 480 uses a digital VRM I believe.

Because of those artifacts, I guess we cannot say that isn't something wrong with the video card too.

The next step would be to test this video card into another computer for a longer period of time in FAH for example, or simply borrow another video card to see if the problems manifests itself again. All the symptoms point now to the video card as being the problem.

I used my old Q6600 based computer to test the card in when I first received it because of all the problems I was having with it.

The video card works fine in it, which was the only reason I did not RMA the card as bad. I tested it with an hour of Furmark and it passed, tested it folding@home and it worked fine.

When it comes to computer equipment, I hate enigmas...

There seems that is a whole website dedicated to this problem : http://www.nvlddmkm.com/

Do you also have other cards installed into the computer?

Nope, I keep it simple with just one card. Not that it's helped anything... ;)

Very interesting link! That website never turned up in my googling, curious. I'll give it a good read through.

Anything to help!

I read through the suggestions, unfortunately most were not applicable (Vista specific patches, or mobile-specific NVIDIA power software). I don't have any tuners or other hardware, just a basic video card.

Tried stock speeds, screen flickered then the driver crashed again. I don't know what to do, I've switched each of the hardware components already. This problem is specific to this computer in some way as the cards work in a non-Core i7 PC.

Edit:

Working on a theory. There are quite a few accomplished overclockers here... about how likely is it that two Core i7 CPUs won't have a fully stable Uncore at 3,200MHz? I can verify the uncore is unstable at or somewhere above 3,600MHz.... as I run 1600Mhz memory it's one of the few common factors I've not explored. Are there any tests that would actually show uncore instability? Again the system passes Linx/Prime, so I don't know. But I am seriously running out of options here...

Edit2: Still getting the crashes at stock 2666Mhz uncore... that driver version ya recommended is really unstable, usually it takes a few hours to a few days for it to crash before I'd know...

I have a hunch.... I've never replaced the Dell DVI cable that came with the monitor. Anyone think that might be it? ;)

Not really...I guess best bet is to try with other video card, from a friend or from a store.

your PC crashing because of VGA cable? Never heard of that before;)

I've tried three. All three were new, can't all be bad. And again they WORK in a completely different PC. I can't afford to keep buying hardware randomly. ;)

Soon as PC Power's brand new Silencer Silver+ PSUs reach product availability I will get one, and be able to build a completely new Core i7 PC from spare parts I used to test THIS PC.

I did some googling, some users DO get this exact "nvlddmkm stopped responding" error and fixed it by changing their DVI or VGA cables, or removing VGA to DVI adapters. Same for ATI users "atikmdag stopped responding" errors. Even stranger things were reported in that website's list of suggestions that worked.

The problem has to be something, and I need it fixed. This mess has gone on long enough.

Well, wasn't the cable.

You guys do subzero overclocking loads of times... have you guys ever burned up or worn out a PSU before? I'm wondering if it's possible I've worn out both units I tested with from sustained loads, because increasing the power draw does increase the instability.

Considering the Cooler Master Ultimate Circuit Protection 1,100W PSU was proven to have gone bad, and the PC Power & Cooling 750W unit I am using now is a year older than the CM UCP 1.1kW unit....