0

I have a Nvidia GeForce GTX 660 graphics card and a dual monitor setup. Sometimes my screens go black and get a warning like:

Nvidia Kernel stops responding.

I tried a lot of things; specifically, I updated my OS, GPU driver, second GPU driver, motherboard BIOS and utilies. But I am still getting error sometimes, about once every two days. I downloaded a program called "EVGA PrecisionX 16" and use it with these settings:

enter image description here

As already mentioned I use two monitors which are identical, I mean they both have 1920x1080 resolution and are same model. I connect one of them first to my Nvidia GPU HDMI port and second one to Nvidia GPU DVI port. I usually use both of them, but while gaming I disable secondary (with help of a program called UltraMon) and use my primary monitor to play.

And while I am gaming I monitor values from this program and I observed max. 65°C in GPU temperature. I monitor my CPU temperature and fan speed and they appear to be normal. Am I wrong about my GPU voltage?

My hardware and software details:

  1. Motherboard: Gigabyte Z97M-D3H
  2. BIOS: 05/30/2014 14:12:20 ver:04.06.05
  3. Processor: Intel i5-4440 CPU @ 3.10 Ghz (4 CPUs) ~3.1 Ghz
  4. Memory: 8192MB (2 * 4096 MB Ram with Kingston brand)
  5. Operating System: Windows 7 Home Basic 64 Bit (6.1, Structure 7601) (with latest updates)
  6. First GPU: Intel HD Graphics Family 4600 (with latest driver and optimum settings)
  7. Secondary GPU: Zotac Nvidia GeForce GTX 660 2GB 192Bit GDDR5 (DX12) PCI-E 3.0 (with latest driver and optimum settings)
  8. DirectX: DirectX 11
  9. PSU: 600 Watt

If you want more information about my system I can edit post.

EDIT : I uninstalled whole W7 and installed W8.1 64-Bit. There is no driver kernel error but sometimes, espeacially while playing game with background music with Windows Media Player , I got some audio freezing. Except this, there is no problem. It may be not a hardware problem.

UPDATE : Today(Feb 4th) I got this error again in Windows 8.1

UPDATE : (March 2015) I upgrade my 8 GBs RAM to 16 GB and have never got that error. I think this is the real milestone for this problem.

17
  • How often does this happen? After your screens go black, what happens? Is this a custom built computer and did you buy any of the parts used?
    – hmit
    Feb 1, 2015 at 6:34
  • It is a built computer and a sales person helped me while I am building. We started from motherboard and psu and then gpu. So I think we should not have a problem with hardware. And it happens one or two times in a day nearly. Somedays it does not happen. But I am very unhappy about that situation. I paid a lot money on this computer and like everyone I do not want to enforced to using warranty and waiting coming. Feb 1, 2015 at 16:06
  • Have you reinstalled NVidia drivers?
    – hmit
    Feb 1, 2015 at 16:08
  • Yes I tried this method several times, but it should not connect to driver I think. Now I am using the latest stable version of GPU installed by GeForce Experience. Feb 1, 2015 at 16:41
  • It looks like you are not using the Intel HD Graphics - if that is the case then you should probably disable it in the BIOS. Section 2-5 of the manual indicates where to do that. Feb 2, 2015 at 19:22

5 Answers 5

0

It looks like you are not using the Intel HD Graphics - if that is the case then you should disable it in the BIOS. Section 2-5 of the manual indicates where to do that.

This will have small additional benefit of using a little bit less power (a few watts).

Note for other readers: you will need to consult the manual for your specific motherboard to find where to disable the on-chip graphics in the BIOS (or UEFI).

0

The Kernel crash happens for many reasons. I will list below some of them, and what you can try out.

1. Driver Error
Do you have the latest driver installed? If so, then try to install an older version and check if the problem persists. (Note that if you have a beta version of the driver, uninstall it and install a stable version. Also when a clean install is recommended. Keep in mind that if you do a clean install, all your profile settings will be deleted.)
2. Hardware Error
Kernel errors happen also when some capacitors (or other components sometimes) get blown on the card. Open your card, and check it out. But you have to consider the fact that you might need to apply new thermal paste to the card.
3. Third-party Software
Do you have MSI Afterburner beside EVGA PrecisionX 16? If so, please remove one of them. It is not recommended using two overclocking programs at the same time.
4. Other reasons
Some other reasons could be other programs you use that might affect the card's work, overheating, low spin of the fan, overclocking etc..
Some cards do not work well overclocked for some games. So you can try underclocking (downclocking) the card.

P.S.
Please try the above, and if none works, give us more updates so we can try to help you!

13
  • High voltage for gpu is not good for everytime you said hah? Feb 1, 2015 at 16:07
  • What do you mean?? Every gpu has a maximum overclocking limit (which means a voltage limit too).
    – Erlis D.
    Feb 1, 2015 at 18:33
  • How can I detect or calculate my GPU voltage limit? Does it depend on GPU model or motherboard or PSU? Feb 1, 2015 at 18:50
  • To check the current gpu voltage, use a tool like HWiNFO (hwinfo.com). When in welcome screen, select sensors only, and click Run. You will see in the list your currect GPU voltage also. It depends on the GPU model and PSU. But if you already didn't run into problems before, it doesn't make any difference. So, in this case you should concern about the GPU. Check out this review (where you can find also the max voltage allowed for your card) anandtech.com/show/6276/…
    – Erlis D.
    Feb 1, 2015 at 23:08
  • I tracked my GPU values for 5 hours and then this is result : Result Photo While the tracking was continuing, my screen got black 2 times and computer frozed. As you see, my GPU Clock and GPU memory clock's value is too low. Is this source of problem? Feb 2, 2015 at 12:59
0

I've seen similar behavior on several systems at my office. Generally it comes down to the two graphics processors interfering with each other. The only time I can think of wanting an integrated graphics processor working at the same time as a discrete card is in a laptop, where battery life is a consideration.

If you have a legitimate reason for having both operational, then disregard the following:

Go into your motherboard bios and disable the integrated intel graphics processor. It is generally in the "peripherals" or "On board" areas of a BIOS. After that is completed, I would recommend uninstalling the Intel graphics drivers. There are several graphics driver scrubbers available on the internet, and I would use one of them to make sure all parts of the old driver are removed.

1
  • I closed this feature two days and ago and install Windows 8.1 and since that day I have got no driver kernel error but as I mentioned in my edit, got only some audio freezes. I hope I will never have this silly error. Feb 4, 2015 at 5:03
0

Found this on another site - Worth a shot i reckon.

Use ''Run'' and type ''regedit'' HKEYLOCALMACHINE > SYSTEM > Currentcontrolset > control > graphicsdrivers. Right click ''graphicsdrivers'', click new >Dword (32-bit) value. Name the new value TdrDelay. Modify the value to 8.

What I did next was delete ALL of my graphic driver installed (the nvidia one). It will swap to the build-in one on your motherboard. Then use driver sweeper, select nvidia, and let it erase ALL of your nvidia driver files. I then re-installed the nvidia driver through the installation CD I got with the graphic card, and did NOT get any updates.

Now it all works flawlessly!

1
0

Struggled with a similar issue for years. I will often boot up with no image on display (display goes to sleep). Frequent freeze, anytime under load or without load. Sometimes the freeze result in auto restart, sometimes it could remain like frozen. Sound would continue to play. Things I did in order:

  1. Started replugging monitor when rebooting after freeze episode - this seemed to work most of the time. Though sometimes it would require several replugs and reboots to finally show on the monitor.
  2. Changed RAM Sticks within the first month of putting together the PC
  3. Started removing power plug as well and replugging. That somehow worked but was random.
  4. Out of the blue started having memory issues/ beeps and it prevented POST.Removed two of the newly installed RAM sticks at some point. That seemed to work, but random freezes persisted for years.

What finally worked after almost 8 years of unplugging and replugging when restarting: I changed casing after reading on a forum that metal touching motherboard can cause all sorts of random issues. Miraculously everything just worked, I could reinstate back my memory sticks.

I hope this can help someone else. Test your PC barebone (out of casing) before building or when having seemingly random issues.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .