Intermittant Driver failures

[bear2790]

I'm seeing what is becoming a common failure amongst GTX260-216 cards (I do not know if it happens to GTX275/280/285/295 cards). I'll be gaming (doesn't seem to matter which game I'm playing) and suddenly the screen will go black, and the audio will continuously loop and then I need to start Task Mamager and manually end the game program. Once that's done I see a yellow triangle with a ! in it waiting for me to see it in the taskbar. If I restart the game the fps will be half the frames I had before the crash until I reboot the computer. I've seen this happen with at least three differnt GTX260-216 cards by three different subvendors.

What's causing this driver failure and what's the fix? Thanks

tomana

 

[Mussels]

its usually the card overheating/going bad.

i also had this happen on my 8800GTX and 8800GT as they got older.

 

[bear2790]

hmmm, was maufactured early 2010, has a massive heatsink plate for the VRm/Memory/NVIO chips and the GPU wasn't oner 55C ??? Was playing Half Life 2 Episode 1 when it failed but it also fails (randomly) on other games as well, and perhaps a program like Paintshop Pro once in a while. Very spuratic failure. Card can work perfect for days on end and then, bam, the drivers crash, or the card is locking up somehow. For the rest of tonight I haven't seen a failure. Temp failures are almost always repeatable when you apply heat to the circuit or component(s) [I used to be a component level repair tech] so if it's temp related I should be able to run FurMarks and then the failure ought to definately show itself. But it doesn't. It keeps working perfect even though the card is running at a temp 15C higher than when the failure occured.

I have to say that I doubt it's a temp failure, and since the card is almost new, it's going bad could be the case, though I have to say probably not. I'm guessing that something is locking up the circuitry and the card blanks the screen until the game is shut down. Then , when I restart the game, the frame rates are half what they were before the card failed. It happens this way every time. Has to be something else besides temps.

I've read online about a lot of folks who had the same exact failure:

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

 

[Mussels]

GPU temp is only one part of the card. VRM's or RAM could be 50C higher for all you know.

 

[1nf3rn0x]

I've had this problem but it was 5 months ago, it seemed to disappear with newer drivers.

 

[yogurt_21]

I've had this problem but it was 5 months ago, it seemed to disappear with newer drivers.

second only 9 months ago with my 9600gt and win7 beta

@op what drivers are you using?

 

[bear2790]

Tried v197.45 thru v257.21 and I just checked the driver downloads page at the subvendor's website. The latest release they have posted is v191.07 so I'm going to give those a try. I had the same failure occur with a different subvendor brand of GTX260-216 and they replaced the card under warranty. The replacement card failed the same way, once, soon as I ran it in my rig - while using a Windows XP based Paintshop Pro graphics program - but after that I never saw any failures, regardless what driver version I tried. The card that spuradically failed was the 55nm GPU model, as is this card that is failing now. The replacement that the subvendor sent me was the 65nm model. I don't know if the GPU die size/design or the subsequent component layout, or the PCB itself, or .....

I have a used GTX275 I bought on ebay coming tomorrow. If that card fails the same way it will be RMA time for both of them. Problem solved

 

[overclocking101]

happens with both my cards seprate and in sli, its a driver not hardware issue. For me though it doesnt happen in game it only happens using folding@home. I can game for hours no problem but turn on f@h for 3 minutes and wham! its a driver problem for sure i just dont know whats causing it, funny thing is in win vista and winXP it doesnt happen.

 

[bear2790]

I believe you are mostly right, but I have Vista 32 and it is happening to me as well. I'm beginning to think there is a possibility that it's a inherent flaw in the nVidia design that went out to the subvendors and if that proves true there could be a class action lawsuite against nVidia in the future. If not, they have noting to worry about. It is a wide spread problem, based on all the postings I've seen online.

 

[overclocking101]

yeah im not sure what actually causes it, it seems for me anyway it only happens at 100% cpu load for long periods of time, even underclocked and overvolted it happens

 

[bear2790]

yes, I thought it was due to OC but it's not. I'm running factory OC speed. It's a known problem but it seems those who may be "in the know" ain't saying ? No way to know if the new GTX 400 series will share the same problem. Need more info.

 

[Mussels]

yes, I thought it was due to OC but it's not. I'm running factory OC speed. It's a known problem but it seems those who may be "in the know" ain't saying ? No way to know if the new GTX 400 series will share the same problem. Need more info.

its a known problem, we say why, we get ignored.

either your system is unstable and feeding bad data to the card, or the card itself is unstable (due to heat/just being dodgy).


just because you are at stock clocks on your system and video card, doesnt mean its stable...

 

[bear2790]

ok, I just got a GTX275 today, and I decided to test it for stability and proper operation by using rthdribl and MSI KomBuster (a much gentle stresser utilty than FurMarks) and guess what? As soon as I launched KomBuster it opened to the test window and then the screen went blank and I have to start Task Manager and shut Kombuster down and yes, the yellow Yield sign with the ! in it was there - and - same driver failure is posted:

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

so there's no way it's a temperature problem because the GPU never went past 60C (the utility blanked the screen and the drivers stop responding almost immediately). If it's a dodgy problem then in these cases I have witnessed, that's the same as saying it's a failure of the product, the drivers (which are put out by the same nVidia Corp. as the cards that are sent to the subvendors) or both.

To date, I had three different brands/models of GTX260-216's experience the same driver failure, and now, a very high end GTX 275

The GTX260's spuradically failed (only on about three separate occasions) during some games, but very often with Kombuster and Paintshop Pro; so far the GTX 275 has been running about an hour and no game driver failure yet.

 

[{JNT}Raptor]

I'd have to say the Issue is in your system and not the cards given the fact that your on the 4th card now.....you must at least entertain the Idea that something in your rig is causing this other than the vid card....IE bad PCI-E slot...hot NB......memory going bad.....PSU crapping out......theres a host of possibilities to entertain.

It would help out ALOT if you filled in your system specs in your profile so we could see what your already looking at.

Hope it helps.

 

[bear2790]

I sent a month old 55nm GPU GTX260-216 card to one of the subvendors for the same failure, after I received it back from an ebay buyer I sold it too - for the same failure. I thought it was co-incidence until I went online and used keywords:

Display driver stopped responding and has recovered

Go ahead, give it a whirl on Google and see how many hits you get. Personally, I was quite surprized.

Here's my hardware:

- GIGABYTE GA MA790GP UD4H (rev 1.0) 790GX/750SB
- AMD Athlon X2 7850 AM2+ Black Edition (OCed to a stable 3.0GHz)
- PC2-6400 ( 2GB OCZ Reaper 4-4-4-12 @1.9V / 1GB G.Skill 4-4-4-12 @ 1.9V)
- Western Digital Caviar HDD SATA II 7200rpm 150GB
- OCZ ModXtream PRO 600 Modular (dual rail ... 2 x +12V @ 26A)
- Thermaltake V9 Black Edition Case (2 x 230mm fans, 2 x 120 fans)
- Palit GTX 275 HDMI model (DVI, VGA, HDMI outputs)
- MSI N260GTX Twin Frozr OC (RMA'ed back to the egg today)

 

[overclocking101]

yeah its not a hardware problem. its a driver problem end of story, there must be a specific driver that doesnt have it and I will find it. I have tried 3 different GTX2XX cards with the same problem. bear2790 download the latest furmark (V1.8.2), run it but click on "furmark 1.7.0 mode" see if your driver still fails, mine no longer fails in furmak after doing that but with folding it does.

 

[bear2790]

Alright, I'll try V1.8.0 in 1.7.0 mode as well as Furmark 1.7.0 (they have all the versions listed for download) and I'll let you know. It's been half a day and so far, I've seen no failures while gaming or using any other program except Furmarks and Kombustor (which has the same interface window layout as Furmark).

Wow, got something here. I can make it fail repeatedly and consistanly to cause the following error:


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

Here's what was what was what. What? Oh, ok ... just the facts:

GTX 275: factory clock 633MHz set to 680MHz/shader clock linked to GPU clock/memory factory 1040MHz set to 1200MHz and fan at 75%

Furmarks 1.8.0 ran on 640x480, stability mode, FX turned off, windowed, no AA, just to get the GPU to run at 100%

I pre-positioned the GPU clock slider to 700MHz (a clock speed I know this card works well at) and as the temp rose past about 58C I clicked the APPLY button in MSI Afterburner and Furmark 1.8.0 froze, and the driver stops running, the yellow yield sign appeared ... you know the story by now.

Interesting thing is, if I start at 633MHZ GPU factory clock setting and increase to 670MHz then the drivers do not fail. It seems the higher the card gets OCed the probability of failure increases ???

 

[Mussels]

yeah its not a hardware problem. its a driver problem end of story, there must be a specific driver that doesnt have it and I will find it. I have tried 3 different GTX2XX cards with the same problem. bear2790 download the latest furmark (V1.8.2), run it but click on "furmark 1.7.0 mode" see if your driver still fails, mine no longer fails in furmak after doing that but with folding it does.

its hardware.


the driver stopped responding error is a GENERIC error. nvidia drivers come up with that one error when ANYTHING goes wrong - be it an unstable system feeding bad data, overheating (STOP saying it cant be overheating based on just the GPU temp - there is more than just a GPU on your card!), or PSU problems. My guess at this point, is your PSU.

 

[segalaw19800]

Ati card will do the same ,,.... try updating all the driver for your system and update hdd firmware and dvd cd dirve firmware

 

[segalaw19800]

if no joy it hardware.

 

[segalaw19800]

Any good news

 

[bear2790]

Understood. No more mention of it.

I have an extra PSU in my hardware box of goodies, has lots of amps on the +12V lines and the other lines as well. Not as good quality as the OCZ but it worked well in the past. I keep it as a back up, especially for summertime months (T-storms).

thanks for all the input folks, I do read what you post and I take it into consideration. Now I just need to learn to apply it as well and not think I always know it all (a challenge area for me).

I'll try the PSU tomorrow and let you know how it went.

 

[bear2790]

I decided to pass on the PSU swap-out, but I did notice there was alot of studder dueing gameplay, especially with Half-Life 2 series, so I removed drivers v257.21 and installed 197.45 and the studder disappeared. Everything works great right now so I'll just use it as it is. Thanks for all the input and see you in other Tech Powerup threads