Recently, my computer is having some errors with nvidia kernel mode driver. Damn thing keep crashing. Anyone know what is with it? I have the latest patch and every thing. I even rollback on older driver and it sill crashes.

I'm running liquid cooling for Nivida 580 x2 SLI, average temp is 50 degrees.

Advice please.

 

Screen blacking out and coming back. Later there will be a small error bubble appearing task bar stating that it had stop working and just got recovered.

 

I've gotten the same thing in the past, it used to happen alot more if my laptop had been running for a long time so a quick reboot always helped.

Also updating the driver helped alot and i've only had the error once since updating.

 

Nope, reboot and update still happening..

 

Try cleaning your computer's interiors, this might be caused by your card overheating.

 

What kernal update did you get?

I kept getting the problem with version 295, currently using 301 only had the problem once from what i recall

 

There was a computer guy here once, ages ago... Cervus, I think his name was. He said if you used vodka on the inside of your PC (while its running), you could fix shit like this.

I did it, and it works great.

 

I used an air compressor to clean the interior of my comp weekly, and I'm running on liquid cooling. Average temp is 50 degrees.

@Shade: I rolled backed all the way to 275 and still having the same problem...

 

Hey man, just searched around the internet a bit more and apparently this errors been reported since version 190, and that once you start getting the error no matter how far you roll back the driver you'll keep getting the error.

People on different forums were complaining that Nvida knew about the problem but haven't fixed it.

Heres a posts which I hope may help

Someone replied to that saying, overclocking does cause problems with solid state components.

I thought that one might help as I have a MSI laptop with a button which overclocks the graphics card a small amount and that seemed to help as I haven't had it crash and recover in quite some time.

Hope that helped somewhat

 

Hi.

Lurked for the last two years or so, figured I'd make my own account finally just to give you some advice.

Go download EVGA Precision X, you'll probably have to register for the site but you don't have to pay anything. They have a version without the requirement to sign up, but the Precision X version is much nicer.

Download. Install.

Looking online the GPU Clock for the 580's base model is 772. If you have the base models try dropping the GPU Clock down to 750. Usually these kinds of problems only come in pre-OC'd cards because, as far as most companies are concerned, so long as the card runs at the OC'd amount, without failing immediately, they're a "perfectly good product." After a bit of stress is applied they start to conk out though.

Hopefully you have one of the overclocked models, at which point you might want to drop the clock down so much as 50. I highly doubt you'll notice any impact ingame.

If that's not your problem? Come back and we'll get a bit more technical.

Also, vodka wasn't a bad idea!

If this fixes the problem and you do have the base models, I'd suggest running one card at a time and seeing if it's a problem with one or both of them. Then return the faulty card(s) at least a month or two before your warranty ends. At least that way they've usually gone out of production, depending on manufacturer, which allows you to trade them in at the exact same dollar value you paid.

Free upgrades!