[muboa]

Video card: GTX570
Mobo: ASRock Z68 Extreme4 Gen3
CPU: i7 2600k
1000+ watt PSU

Apparently when I leave my computer on for more than like 2 days, the latest Nvidia drivers will crash. "Driver has stopped working and has restarted," or something to that effect.

At first I suspected overheating, but:
It idles at about 40C on average, according to MSI Afterburner. It gets up to 80C sometimes when I'm in games, and I have no problems then, and from what I've read those are normal temps for the gtx570. So, I assume temperature isn't the issue.
I've even monitored the temps as the driver crashes happen, and it's always around 40-45C during those times - not much higher than when I first turn the computer on.

I have never had a full system crash. The drivers just crash and restart (which looks like a black screen for 2-3 seconds, then everything comes back up). At its worst, this will happen about once every 45 seconds to 1 minute.
Something that has happened once is that the display in Firefox was really wonky - most of it was black, and bits of text were all at the top of the screen, scrambled into each other. Nothing else looked weird at the time - just Firefox.

And again, this only happens when I've had my PC on for a few days. If I turn it off for about half-an-hour, it's fine again.

Let me know if you need any more info, and thanks for reading.

 

[makinu1der2]

What brand is the Power Supply? What are the temp and voltage readings from the BIOS?

I would suggest uninstalling the driver and trying a previous version of the NVIDIA driver

-Download the Driver Sweeper installer from here and install it. Don't run the program yet
- First, Uninstall the NVIDIA driver/software from Control Panel > Uninstall a program and restart
-Now run Driver Sweeper, select NVIDIA- Display and then select Clean
-After complete restart the PC then install the previous version of the graphics driver.

 

[muboa]

Sorry for taking so long to get back to this thread.

Thanks for your response. I don't know why I didn't just try different driver versions in the first place, seems like a pretty obvious possible solution.
I actually found a newer version of drivers for me, which are in beta so that's why I hadn't seen them before. I'm going to give those a try for a few days and see what happens, then if those don't work (which could very well be the case since they're in beta) I'll try some older ones for a few days.

Thanks again.

 

[makinu1der2]

Okay let us know how the other version of the drivers work.

Just remember to use the process I posted above when changing drivers. You want to make sure the previous version is completely removed.

 

[muboa]

Just reporting that everything seems fine now, after several days.
Sorry for making a thread in the first place when the solution was pretty obvious.
And thanks for the suggestion makinu1der2

 

[makinu1der2]

No problem and glad you got it sorted.

 

[muboa]

Re: [SOLVED] Nvidia driver crash when computer on for long periods

I hate to bring the topic back up when it's been marked solved, but apparently it's not solved. Started happening again today, after over 10 days of the PC being on. I tried a few other driver versions, and the problem persisted, even immediately after installing them.

Interestingly, I couldn't find voltage info for my card/pci slot in bios, but here's everything speccy has to say, including voltage:

GPU GF110
Device ID 10DE-1081
Revision A2
Subvendor EVGA (3842)
Series GeForce GTX 500
Current Performance Level Level 1
Current GPU Clock 50 MHz
Current Memory Clock 135 MHz
Current Shader Clock 101 MHz
Voltage 0.913 V
Die Size 520 nm˛
Release Date Dec 07, 2010
DirectX Support 11.0
OpenGL Support 5.0
Bus Interface PCI Express x16
Temperature 42 °C
ForceWare version 295.73
BIOS Version 70.10.17.00.70
ROPs 40
Shaders 512 unified
Memory Type GDDR5
Physical Memory 1280 MB
Virtual Memory 992 MB
Bus Width 64x5 (320 bit)
Filtering Modes 16x Anisotropic
Noise Level Moderate
Max Power Draw 219 Watts
Count of performance levels : 3
Level 1 - "Default"
GPU Clock 50 MHz
Memory Clock 135 MHz
Shader Clock 101 MHz
Level 2 - "2D Desktop"
GPU Clock 405 MHz
Memory Clock 324 MHz
Shader Clock 810 MHz
Level 3 - "3D Applications"
GPU Clock 732 MHz
Memory Clock 1900 MHz
Shader Clock 1464 MHz
OpenGL
Version 4.2.0
Vendor NVIDIA Corporation
Renderer GeForce GTX 570/PCIe/SSE2
GLU Version 1.2.2.0 Microsoft Corporation

I'm too dumb to know the significance of that.
Any ideas?

 

[muboa]

Re: [SOLVED] Nvidia driver crash when computer on for long periods

Oh also, apparently turning my computer off for even 10 seconds completely eliminate's the problem. I'd left it off for at least 30 minutes before.

I guess that makes it hardly a problem at all, but I'd still like to know what's going on, because that's very very weird.

 

[Tomken15]

Re: [SOLVED] Nvidia driver crash when computer on for long periods

Have you checked the Event Viewer for about the times these failures occur ?

I'm not sure if running this Novabench test NovaBench - Free Computer Benchmark Software will pick up on anything when it does the graphics test, as I think the test would have to fail before it would produce an error message.

Perhaps running it around the times when you may expect a fail could be more productive(?)

 

[Tomken15]

Re: [SOLVED] Nvidia driver crash when computer on for long periods

After digging around a bit more, a more promising fix is to increase the voltage.

Using gpu-z GPU-Z - Download to check its current voltage then with msi afterburner MSI Afterburner you can up the voltage.

I've never used the latter so it will be up to you to work this.

I take it that you will have already used the driver autodetect Drivers - GeForce for the latest drivers for your card, but from the various links that I've visited, this is no guarantee of success as your problem is widespread throughout Nividia and came across this narrative from another forum (which I wasn't sure that it wasn't a wind up), which I think is worth a read but so far, those using the voltage increase fix hadn't reported any further instances of the problem.

Posted 08 May 2012 - 03:20 AM
Hello,

I work for a semiconductor research and development organization. I work with a few former Intel engineers and have spoken about the ongoing "Display Driver stopped responding" error (with black screens/recovery, screen freezing, and blue screens). Come to find out, the former Intel engineers have informed me that the error is well known by Intel. They have received internal memos on the subject stating the issue and the root cause. The cause of the error is an architectural flaw in the Intel processors that spans multiple generations (core 2 duos - LGA 2011, includes mobile processor line also). The error occurs during intense gaming scenarios and/or high resolution graphics settings. This is why, in many instances, turning your graphics down may relieve the error or simulate a fix for a period of time until the components degrade. I have personally researched this problem as a gamer and have spent 12k of my own money troubleshooting the issue over that past 3 years.

The engineers have informed me that the internal Intel memos have stated that when the error is encountered, the video card and processor attains physical damage. In addition, a damaged video card and processor can and will cause additional damage to other hardware components (motherboard, memory, and even power supplies). If any of these components are transferred to another or new system, the new system may become damaged during intense graphical applications. Intel has not released this information to the public due to the potential lawsuits that would cost the company millions of dollars.

With this information, the engineers have informed me that there is a third party company that has released a software work around for this error. Unfortunately, Intel is the only company that possesses this work around and will not release the software due to its admittance to their design flaws. While many in the gaming industry have said, in forums, that they have not encountered this issue, eventually the error will occur in there systems as the components age (many gamers will upgrade within a year or two time before the error will occur). I have seen this happen on a water cooled system after a 1.5 years of use. After turning the graphics down, the error went away on the water cooled system. Considering that most gaming rigs are based on the Intel processor, this error is rampant. After the error occurs, even video streaming can be affected by it. Replacing the video card can resolve your issue for a period of time but the damaged processor and components will eventually damage the video card. The only known way to prevent or slow the occurrence of this error is by not playing games or by playing on a small monitor with very low resolution.

The only current fix for this issue is to either attain the confidential software fix from Intel or purchase/build a AMD processor based rig. If you purchase/build an AMD, you cannot use any components that have been plugged into your Intel machines. If you do, the AMD will have permanent issues. I have been in the system building and semiconductor industry for over 20 years. I have seen many companies cover up there flaws in the same manner (examples: 1st generation DDR memory and AMD’s K6 series processors). You may find people that have made driver changes, software changes, and/or components changes. By doing this you may fix the issue for a period of time but it can reoccur any time. The way people use, setup, and/or play games (types of games and DX10) will determine how fast or delayed the error will occur with an Intel processor. Until Intel is forced to fix this problem, you will continue to see forum entries on this error for future generations of Intel processors. In addition, the Intel processor flaw is not just limited to NVidia video cards but also AMD video cards. If you search the web you will see many AMD video cards with the same black screen and driver stop error but with a different Windows error (type of error notification is OS, hardware, and driver dependent). Intel must fix this architectural problem and revert back to the quality company they based their name and reputation on.

Good luck!!!!

 

[Tomken15]

Re: [SOLVED] Nvidia driver crash when computer on for long periods

Just noticed that you already have the current voltage via Speccy.

 

[Wrench97]

Re: [SOLVED] Nvidia driver crash when computer on for long periods

After digging around a bit more, a more promising fix is to increase the voltage.

Using gpu-z GPU-Z - Download to check its current voltage then with msi afterburner MSI Afterburner you can up the voltage.

I've never used the latter so it will be up to you to work this.

I take it that you will have already used the driver autodetect Drivers - GeForce for the latest drivers for your card, but from the various links that I've visited, this is no guarantee of success as your problem is widespread throughout Nividia and came across this narrative from another forum (which I wasn't sure that it wasn't a wind up), which I think is worth a read but so far, those using the voltage increase fix hadn't reported any further instances of the problem.

Posted 08 May 2012 - 03:20 AM
Hello,

I work for a semiconductor research and development organization. I work with a few former Intel engineers and have spoken about the ongoing "Display Driver stopped responding" error (with black screens/recovery, screen freezing, and blue screens). Come to find out, the former Intel engineers have informed me that the error is well known by Intel. They have received internal memos on the subject stating the issue and the root cause. The cause of the error is an architectural flaw in the Intel processors that spans multiple generations (core 2 duos - LGA 2011, includes mobile processor line also). The error occurs during intense gaming scenarios and/or high resolution graphics settings. This is why, in many instances, turning your graphics down may relieve the error or simulate a fix for a period of time until the components degrade. I have personally researched this problem as a gamer and have spent 12k of my own money troubleshooting the issue over that past 3 years.

The engineers have informed me that the internal Intel memos have stated that when the error is encountered, the video card and processor attains physical damage. In addition, a damaged video card and processor can and will cause additional damage to other hardware components (motherboard, memory, and even power supplies). If any of these components are transferred to another or new system, the new system may become damaged during intense graphical applications. Intel has not released this information to the public due to the potential lawsuits that would cost the company millions of dollars.

With this information, the engineers have informed me that there is a third party company that has released a software work around for this error. Unfortunately, Intel is the only company that possesses this work around and will not release the software due to its admittance to their design flaws. While many in the gaming industry have said, in forums, that they have not encountered this issue, eventually the error will occur in there systems as the components age (many gamers will upgrade within a year or two time before the error will occur). I have seen this happen on a water cooled system after a 1.5 years of use. After turning the graphics down, the error went away on the water cooled system. Considering that most gaming rigs are based on the Intel processor, this error is rampant. After the error occurs, even video streaming can be affected by it. Replacing the video card can resolve your issue for a period of time but the damaged processor and components will eventually damage the video card. The only known way to prevent or slow the occurrence of this error is by not playing games or by playing on a small monitor with very low resolution.

The only current fix for this issue is to either attain the confidential software fix from Intel or purchase/build a AMD processor based rig. If you purchase/build an AMD, you cannot use any components that have been plugged into your Intel machines. If you do, the AMD will have permanent issues. I have been in the system building and semiconductor industry for over 20 years. I have seen many companies cover up there flaws in the same manner (examples: 1st generation DDR memory and AMD’s K6 series processors). You may find people that have made driver changes, software changes, and/or components changes. By doing this you may fix the issue for a period of time but it can reoccur any time. The way people use, setup, and/or play games (types of games and DX10) will determine how fast or delayed the error will occur with an Intel processor. Until Intel is forced to fix this problem, you will continue to see forum entries on this error for future generations of Intel processors. In addition, the Intel processor flaw is not just limited to NVidia video cards but also AMD video cards. If you search the web you will see many AMD video cards with the same black screen and driver stop error but with a different Windows error (type of error notification is OS, hardware, and driver dependent). Intel must fix this architectural problem and revert back to the quality company they based their name and reputation on.

Good luck!!!!

Is there a source for that report?

If a known issue is out and around enough for a 3rd party to come up with a software fix I'm sure we would have heard of it, more likely AMD fans spreading Anti Intel messages.

 

[Tomken15]

Re: [SOLVED] Nvidia driver crash when computer on for long periods

Took a bit of finding again but it's on page 3 of Display Driver stopped responding and has recovered. - NVIDIA Forums - Page 3 and it is by The Hun but the article also criticises AMD for the same problem so I don't think rivalry would have been the reason for The Hun to have posted it.

This article may help as a guide to changing the voltage from stock as these OPs had found Lost...BF3 crashing (gtx570 driver restarts) - [H]ard|Forum

 

[Wrench97]

Re: [SOLVED] Nvidia driver crash when computer on for long periods

I wouldn't put much stock in it, the part about the cpu damaging the video card and then damaging any hardware used with the card afterward is a bit far fetched.
And CPU damaging the power supply is not possible.

 

[Tomken15]

Re: [SOLVED] Nvidia driver crash when computer on for long periods

I didn't really know what to make of it as it was above my head, but thought I should include it as there weren't any rebuttals that I could see and my priority was to try and find a fix for muboa.

From the last link I gave, it would seem that the low stock voltage for muboa's card mirrors that of others and would seem to be the crux of the problem.

 

[Wrench97]

Re: [SOLVED] Nvidia driver crash when computer on for long periods

Post #49 was about as good a rebuttal as you'll see on the nvidia forums

Could be, but keep in mind they dropped the voltage to control the heat issues of previous cards. Raising the voltage is sometimes a temporary repair as the card further degrades the problem comes back.

 

[Tomken15]

Re: [SOLVED] Nvidia driver crash when computer on for long periods

Post #49 was about as good a rebuttal as you'll see on the nvidia forums

Could be, but keep in mind they dropped the voltage to control the heat issues of previous cards. Raising the voltage is sometimes a temporary repair as the card further degrades the problem comes back.

I didn't get as far as Post #49 as the suggested voltage increase fix was a repitition of that I'd seen on the other sites I'd visited and would seem to confirm that as the one to go for.

I think I just had an open mind about The Hun's post and didn't take it too seriously (possibly a conspiracy theory) as that was the only site that mentioned anything like that.

While I don't know how old the GTX570 is, but perhaps the low stock voltage isn't able to cope with the advanced games that have been released.

References are made to a BIOS upgrade in Posts #15 through to #17 in my last link, which may negate the need to manually increase the voltage, but muboa will need to check that out.

 

[Tomken15]

Re: [SOLVED] Nvidia driver crash when computer on for long periods

@Wrench97

Had a read through Post #49 but will leave it to those who know what they're talking about to argue the toss.

@muboa

If you enter the GPU's details into EVGA Download Center you will find the drivers that have been mentioned in the Nividia forum link, some of which worked and those which didn't for some.

Whichever option you choose (voltage or driver), I hope this finally sorts the problem for you.

 

[muboa]

Took a while to respond since I wanted to give my attempted solutions time to see if they worked. They did not. I increased voltage via MSI afterburner, and that seemed to solve the problem until today. I upgraded to latest drivers again (there was another update since I last tried this) and driver crashes continued.

I also got a blue screen today, didn't have a chance to get what it said. I let the PC restart automatically and let it sit for a bit, and driver crashes continued without a blue screen, so I assume it doesn't mean much.

From reading the forum you linked, Tomken, it seems to continue to be a widespread problem, which somehow comforts me a bit. The fact that, like I said, my PC only needs to be off for a few seconds to stop the crashes for several days, is also fortunate. I don't suppose anyone around here knows a solution if no one on the Nvidia forums does, so I guess I'll keep an eye out for more solutions there.

I did see a link which recommends setting your power plan in Windows to High Performance, which I thought I had before, but apparently not. I don't know if any relevant individual settings are any different now than what I had before, but I'm going to give that a chance for a few days anyway.

Thanks for taking the time to help.

 

[muboa]

Cool blue screens are happening constantly now, even with increased voltage and my power settings at "performance." They are also happening even with stock voltage and other settings.

Actually, for all I know, the driver crashes could be a separate issue from my blue screens, though I doubt it.

I'm going to try looking around for a solution. In the meantime, attached is the .dmp from the most recent bluescreen (which I have no idea how to interpret - I'll be reading up on that in the meantime). I don't know if this is really in the right forum at this point, but whatever.

All the bluescreens I've had on past computers have been due to RAM. I don't know if that's the most common hardware issue behind BSODs, but I'm hoping that raising the voltage on my GPU didn't somehow damage my RAM. I don't know if that even makes sense.