Display driver atikmdag/nvlddmkm stopped responding and has recovered

[taltamir]

could it be a synching issue with dual channel? maybe your two sticks of ram are each individually ok. but somehow out of synch, so that dual chanel mode introduces errors...

Do they pass a FULL run of memtest+ without a single error? (it can take many hours to do a full suite of tests! so leave it on overnight, and verify that the "wall time" matches how long it was supposed to have ran)

 

[Chaotic42]

This only happens to me in Sins of a Solar Empire and Silent Hunter IV. Very annoying.

 

[Iarwain]

Yes, a full run, I've let it go overnight, many times. I've let the windows memory diagnostic go, overnight, many times. Never get BSODs or any other issues from XP, it just doesn't happen, there is nothing wrong with the memory, which I think is verified when I actually try to underclock the RAM, and the situation isn't helped at all.

I've done those steps in Nvidia's answer now 7 times. I've been over this, apparently it's just not being read. And it isn't as if I'm the only person having this issue.

My video card is -not- factory OCed.
It runs at 60-65 degrees in Vista
I have tried every voltage for every setting
I bought a new hard drive and installed Vista on it, and that hard drive is specifically for trying to figure this out.
I've tried RCs of service pack 1, but as far as I know the official sp is not out, but of course I will be giving it a shot then as well.
There's nothing else I can think to try to prove to you guys this isn't a hardware fault.

I think it may also be important to note that Aero will even crash it, its not just an intensive game. I mean, if I remove whatever amount of RAM, my GPU usage on the desktop (aero enabled) is 0-1% at all times. You mean to tell me that World in Conflict on XP can use 100% of my GPU, and nothing happens, but if Vista uses nothing at all, it crashes? How is that believable?

I have, over and over, acknowledged that a lot of the problems surfacing are just bad hardware. And yes, Vista may exacerbate a problem that already existed in XP, but was managable in XP. The point is, I have no problems at all in XP, and all my hardware passes any test I can throw at it. It even passes the same tests in Vista, provided I lower the RAM amount.

You say that this is a hardware fault, and yet there is a update to attempt to fix it. Now, it possibly fixed the issue for some people, but is it so unbelievable that it didn't fix it for everyone?

 

[taltamir]

Originally posted by: Chaotic42
This only happens to me in Sins of a Solar Empire and Silent Hunter IV. Very annoying.

that's a game issue... sins of a solar empire has some bugs. Turn off all the "effects" and sins should stop crashing (thats what fixed it for me on the 7900GS... but then I got a 8800GTS 512 and it doesn't crash even with all the effects on)

 

[zorrt]

Originally posted by: Iarwain
Nevermind that fact that, if what you say it true, Vista can only be sensitive to hardware changes/faults when there's more than 2gb of RAM in the system, which just doesn't make sense at all. Because otherwise, I can do anything I want, and play any game I want.

If thats targetted at what I said I didn't say anything regarding how much RAM is installed. I have 2gb of RAM and had the same problem.

Yes taltamir XP can run orthos for hours with no issues, people blame vista because once they get into vista it crashes just like that, be it hardware or software fault. Mine ran 100% fine in XP, only when I went into Vista then had problems.

Iarwain, I know its fustrating but don't give up yet. You've said you had already loaded up BIOS default settings. Can I just ask if you have tried increasing your NB voltage and see if it does anything? If you already had then I'll just stay out of this one.

I only have 1 computer and 1 videocard so I can't do different tests to say if the cause of my problem was just an isolated one but hey, doesn't hurt to suggest something that fixed my problem to others right?

 

[magreen]

Originally posted by: Iarwain
Yes, a full run, I've let it go overnight, many times. I've let the windows memory diagnostic go, overnight, many times. Never get BSODs or any other issues from XP, it just doesn't happen, there is nothing wrong with the memory, which I think is verified when I actually try to underclock the RAM, and the situation isn't helped at all.

I've done those steps in Nvidia's answer now 7 times. I've been over this, apparently it's just not being read. And it isn't as if I'm the only person having this issue.

My video card is -not- factory OCed.
It runs at 60-65 degrees in Vista
I have tried every voltage for every setting
I bought a new hard drive and installed Vista on it, and that hard drive is specifically for trying to figure this out.
I've tried RCs of service pack 1, but as far as I know the official sp is not out, but of course I will be giving it a shot then as well.
There's nothing else I can think to try to prove to you guys this isn't a hardware fault.

I think it may also be important to note that Aero will even crash it, its not just an intensive game. I mean, if I remove whatever amount of RAM, my GPU usage on the desktop (aero enabled) is 0-1% at all times. You mean to tell me that World in Conflict on XP can use 100% of my GPU, and nothing happens, but if Vista uses nothing at all, it crashes? How is that believable?

I have, over and over, acknowledged that a lot of the problems surfacing are just bad hardware. And yes, Vista may exacerbate a problem that already existed in XP, but was managable in XP. The point is, I have no problems at all in XP, and all my hardware passes any test I can throw at it. It even passes the same tests in Vista, provided I lower the RAM amount.

You say that this is a hardware fault, and yet there is a update to attempt to fix it. Now, it possibly fixed the issue for some people, but is it so unbelievable that it didn't fix it for everyone?

Iarwain, I think you have one of the highest knowledge to post count (KPC) ratios I've ever seen. But the more you post the lower that ratio will fall

 

[Iarwain]

No I know guys, I'm not trying to be a prick, but it is frustrating. NB voltage has been on auto, normal, high, and very high (those are the options I have)

 

[zorrt]

Oh ok then. Not much control then but so long as you've given it a go.

If its a software fault maybe there'll be a fix for it one day, I wouldn't count on it though as I dont think tthe same issue on XP ever got resolved. But yea most times it is a hardware issue. But not all of us has the money to go spending into swapping parts in and out so we're stuck with what we have. The way I see things, its pointless stress testing your current hardware to see if it craps out. We all know stress testing is 100%. But if you ever get the chance try your parts in other machines, including everthing except the motherboard then that would be the quickest test rather than stress testing for 48hours at a time.

 

[mkivsuptt]

Just wanted to add my experiences with this issue. I started experiencing this issue when I started playing Company of Heroes and World in Conflict about a couple of weeks ago. My setup is as follows:

Asus Striker Extreme 680i
Q6600 OC'd to 3.2 @ 8x400
4x1GB Geil Ultra PC6400 Ram @ 800
2 x BFG 8800GTX OC2 in SLI

I had never really experienced this issue until I started playing these two games. After doing a ton of research, I tried multiple things trying to find a solution. I tried bumping up the vcore voltage, the NB voltage, the RAM voltage. Even after I stopped overclocking the CPU, I was still having this issue. Sometimes, these changes would show small signs of improvement, in that I could play sometimes for an hour before this issue would occur. Sometimes though, it would take less than a minute. It only seemed to happen when I was playing games. Finally, I found someone mention that they fixed their issue by reducing the OC of their videocards. I really didn't think this could be it since my cards came from the factory overclocked, and I figured that they should be stable at these settings, but since I had exhausted all of my other options, I was willing to give it a try. So, I reduced my overclock on the video cards, and so far, I have not had this issue anymore. It's worth a try to see if underclocking your videocard fixes this issue, if anyone that is facing this issue has not tried this yet.

 

[magreen]

Originally posted by: Iarwain
No I know guys, I'm not trying to be a prick, but it is frustrating. NB voltage has been on auto, normal, high, and very high (those are the options I have)

Not at all. I really just meant it as a compliment.

 

[rizorith]

I just built a new rig with a 8800gt overclocked by MSI and I'm running Vista 64. I started getting this same issue from the start and it's not only while playing a game. I get it just sitting there doing nothing in my desktop sometimes.

I've gone back to stock speeds on both the proc and video card and no difference - it still does this. I've heard ATI cards have this problem as well. If that's the case then it's obviously not just a driver/hardware flaw with the card.

 

[Midloo]

I came across this thread while researching the graphical issues I've been having on my brand new machine. Specs in the signature below.

Greatly appreciate the comments from all- especially taltamir, Iarwain, Zorrt, and eyeballkid- as it seems after months of hardware shelf life and a service pack update to vista, there are many with continuing problems.

I'm running vista 64 bit and after a clean install, all driver updates, and SP1 for vista, I'm facing the same "Display driver nvlddmkm stopped responding and has recovered" error. I plan on following Nvidia's support advice to the letter this evening after my third clean install, but I don't hold out much hope as it seems about the same process as I have done previously.

Have run memtest86 will all 4 sticks of 1gb ram at length with no errors, ran seatools for the hard drive, crunched the bejeezus out of Prime95 and nothing is out of order. Try installing 3DMark06 and I can't get past the 600 frame of the first test. Without a "recovered" error.

Funny thing is, this rig ran perfectly for a week before all this nonsense started. Don't have any games installed- just focused on benchmarking with OCCT, Prime95, and 3DMark06. Even when I try to refresh my Windows Experience score, it hangs during the Aero test every single time. On some restarts I get as many as 7 "recovered" errors a minute just sitting on the desktop.

I've RMA'd my GPU, Mobo, and PSU without any improvement.... oh yeah... running a 700w OCZ gamestream which has more than enough juice on the 12v rails to take care of my system. My CPU does have the highest possible VID in the range given by Intel for a Q6600- 1.3250.

Played with PCI-E latency, NB volt, CPU volt, Memory Volt and timings, SLI-enabled memory- the works... to no avail.

Suppose I could keep RMAing until I'm blue in the face or have a 1 year old computer I've yet to use (at day 36 now... when will I be compensated for my depreciable asset?) Will try XP tonight as well if there's a chance I could at least get started using this pc.

I have a hard time believing it is entirely hardware related after the extensive testing and replacing I've done over the last month. Also tried an MSI Neo2 FR board with all the same hardware at one point with the same results. It's disappointing- especially after the high hopes I had for progress after the service pack.

 

[taltamir]

that sounds like quite a harrowing experience. I would recommend you try replacing the motherboard to that of another manufacturer. There might be issues with your specific model. Even from a relatively solid brand like MSI it can still happen. Weather the "fault" lies entirely with vista, nvidia drivers, or hardware incompatibility is nearly impossible to tell, from your intensive tests and various replacements it sounds like hardware isn't the issue. But if there is some sort of compatibility problem then replacing the motherboard with some other model is the most surefire way to fix it.

While the error is an indication of vista power cycling your video card, terminating the driver, and reloading it. It could be caused due to a real video card crash (faulty video card, lack of power, defective ram, or CPU errors). Or it can be caused through some sort of software bug (usually in specific games) or it can be caused potentially through a driver/OS bug, which while the rarest of causes for this error, it seems to be what you are suffering from midloo. Swapping the mobo might do the trick for you.

Or it could be that MSI factory OC video cards consistantly have crashing errors. (or ... maybe they have problem OCing on YOUR specific motherboard, or your power, while sufficient, might be "unstable" power due to a defect in your specific PSU!). Hard to tell without more swapping. Good luck on your quest for stability midloo.

 

[rizorith]

Just a quick update. I installed the latest nvidia driver which is a beta and I haven't seen this bug pop up since. Of course the last driver was good for a few days before it came up but it's something to look into.

I have a 8800GT and am running vista 64 for reference.

 

[chizow]

Originally posted by: Midloo
I came across this thread while researching the graphical issues I've been having on my brand new machine. Specs in the signature below.

Greatly appreciate the comments from all- especially taltamir, Iarwain, Zorrt, and eyeballkid- as it seems after months of hardware shelf life and a service pack update to vista, there are many with continuing problems.

I'm running vista 64 bit and after a clean install, all driver updates, and SP1 for vista, I'm facing the same "Display driver nvlddmkm stopped responding and has recovered" error. I plan on following Nvidia's support advice to the letter this evening after my third clean install, but I don't hold out much hope as it seems about the same process as I have done previously.

Have run memtest86 will all 4 sticks of 1gb ram at length with no errors, ran seatools for the hard drive, crunched the bejeezus out of Prime95 and nothing is out of order. Try installing 3DMark06 and I can't get past the 600 frame of the first test. Without a "recovered" error.

Funny thing is, this rig ran perfectly for a week before all this nonsense started. Don't have any games installed- just focused on benchmarking with OCCT, Prime95, and 3DMark06. Even when I try to refresh my Windows Experience score, it hangs during the Aero test every single time. On some restarts I get as many as 7 "recovered" errors a minute just sitting on the desktop.

I've RMA'd my GPU, Mobo, and PSU without any improvement.... oh yeah... running a 700w OCZ gamestream which has more than enough juice on the 12v rails to take care of my system. My CPU does have the highest possible VID in the range given by Intel for a Q6600- 1.3250.

Played with PCI-E latency, NB volt, CPU volt, Memory Volt and timings, SLI-enabled memory- the works... to no avail.

Suppose I could keep RMAing until I'm blue in the face or have a 1 year old computer I've yet to use (at day 36 now... when will I be compensated for my depreciable asset?) Will try XP tonight as well if there's a chance I could at least get started using this pc.

I have a hard time believing it is entirely hardware related after the extensive testing and replacing I've done over the last month. Also tried an MSI Neo2 FR board with all the same hardware at one point with the same results. It's disappointing- especially after the high hopes I had for progress after the service pack.

It sounds like you've gone through a lot, but I didn't see these mentioned anywhere although they border on obvious:

1) Update your BIOS. I have the little brother version of your board which would not run more than 2 dimms or 2GB+ without a BIOS update (documented fix also).

2) Try running with only 1 or 2 sticks of RAM installed. No this isn't a long-term solution, but many of these TDR errors are RAM/mobo related problems and this would help isolate the problem.

3) Run everything at stock clock speeds, fail-safe defaults loaded in BIOS. If you find you're not getting errors than you know a tweak or overclock is causing the problem and you can isolate the issue.

 

[patgwashere]

try a better power supply.
i had the driver stopped working and has recovered error. i changed the video card to a different rail on power supply and now all is well.

 

[patgwashere]

power supply was my problem

 

[Arkons]

I have a BFG Tech 7800GT which is factory overclocked. It is definitely overheating when playing video games. I've set the fan speed up to 100% using RivaTuner but to no avail. This started happening after I recently updated my drivers. Originally the problem was occurring under XP so I initially suspected the drivers were faulty. I used this as an excuse to upgrade to vista which I had been meaning to do for sometime now for some reason. Under vista the problems continued so I downloaded SpeedFan and found out my GPU was running 65C just running the desktop and well over 70 when I even think about running a game.

I can't find how to change the settings suggested on page 3 (posted below). How can I do this with RivaTuner, I can't find the settings? What could be causing my card to be overheating ... I've had it for 3 years without any prior problems.


"- Some of these computers had factory overclocked videocards (mainly 8800 GTs). On these machines all problems disappeared when I set the the card to default reference speeds in 3D-mode with Rivatuner. If this is the case for you I'd recommend to RMA, or if that's not possible to flash your card with a default bios (without OC).

6. If you have a factory overclocked graphics card use Rivatuner to set '3d performance mode' clocks to default Nvidia reference clocks. (i.e. default clocks for 8800 GT: core 600 MHz, shader 1500 MHz, memory 900 MHz)."

 

[Arkons]

Figured it out actually. A little confused because it told me I needed to reboot to determine the default clock frequencies. Not sure if I just have to re-run the program after I finish the reboot to detect the default frequencies. I rebooted then ran the program again immediately after restart and the frequencies were set at 425 and 1050 for the core and memory respectively. I set the core to 400 and the memory to 999 but it didn't have any effect on my overheating.

What could cause a graphics card to just start to overheat after 3 years of flawless operation?

 

[taltamir]

Originally posted by: Arkons
Figured it out actually. A little confused because it told me I needed to reboot to determine the default clock frequencies. Not sure if I just have to re-run the program after I finish the reboot to detect the default frequencies. I rebooted then ran the program again immediately after restart and the frequencies were set at 425 and 1050 for the core and memory respectively. I set the core to 400 and the memory to 999 but it didn't have any effect on my overheating.

What could cause a graphics card to just start to overheat after 3 years of flawless operation?

is the fan still spinning or did it fail? is the heatsink clogged up with gunk?

 

[Arkons]

fan is still spinning. Heat sink was filled with gunk that I just blew out. GPU is running cooler now but is still spiking to nearly 70 degrees when in a game.

edit: heat is now reaching around 76C but isn't currently causing a crash.

 

[taltamir]

excellent. 76c load is completely within tolerance, and I am glad to hear your crashes stopped.

 

[Mr Fox]

Originally posted by: taltamir
excellent. 76c load is completely within tolerance, and I am glad to hear your crashes stopped.

Not For 7800 GT He needs to clean out the damn card still...

mine ran no higher than 50C

The Same Card.

 

[Arkons]

Originally posted by: taltamir
excellent. 76c load is completely within tolerance, and I am glad to hear your crashes stopped.

So after I posted that a little later I had a little hicup that concerned me. Whatever happened, I saw a yellow blip on the screen for a second (was just running desktop) and then it recovered but went completely black a split second later. Monitor said there was no video, I tried restarting but it didn't pop back up. I may have accidentally loosened the DVI cable with my foot but otherwise I think it could be a problem. I wish I would have been more scientific about this but the way I got the video back was shutting down the entire machine, unplugging the psu, flipping the power supply switch on the back, unplugging the dvi cable, plug it back in and then switched the power on and plugged the psu back in. When I turned it on after that the monitor came right up and I haven't had an issue yet. I'm hoping I just loosened the cable.

 

[BLHealthy4life]

Originally posted by: Arkons


What could cause a graphics card to just start to overheat after 3 years of flawless operation?

easy.....

Your TIM has gotten old and has "dried out" and/or migrated out from between the GU/HSF.

If I wer you, I would reapply a good TIM and see what happens.