2 year long bug, Nvidia Display drivers crashed. - Tried everything, I'm now clueless

[renegadexan]

I been having this issue ever since I bought my computer and so far everything I tried seems to only make the matter worse.
I've read many thread regarding the issue and nothing worked.

Here is a list of what I tried.

Memtest86+ for a whole day, no error.
Tried 1 stick of 4Gig ram, I got 4(16Gig)
Tried countless different drivers versions over the course of 2 years.
RMA'ed: GPU,PSU,Motherboard,SSD
Only pieces not RMA'ed are cpu and ram.
Updated bios to latest version.
Tried Windows 7 and 8 64 bits, no difference.
Run default no overclock settings
Removed HDD

Tweaks I tried recommended for threads:

-Set fan speed to 60% on my GPU
-Uninstalled the Nvidia 3D vision drivers
-Disabled firefox's "use hardware acceleration"
-Adjust image settings with in Nvidia from Quality to Performance.
-Set frequency to 61Hz.
-Tried Microsoft registry fix
-Tried settings display to performance rather then quality.

Yup noting worked, so yeah I am not sure what else to try at this point, hopefully someone with 30years of IT experience can figure it out.

Seriously considering buying a new board just cause why not? Don't see what could fix it at this point.

Computer Specs:

Motherboard: Asus P8P67 LE LGA1155
CPU: Intel Core i5 2500k
Ram: Corsair Vengeance 16G 4x4 GB
GPU: EVGA GTX 580
SSD: OWC Mercury Extreme pro 6G 120GB
HDD: WD1002FAEX 1GB
PSU: Thermaltake Toughpower XT TPX-775M 775W

 

[radrok]

Could mean nothing but it's worth a try.

Have you tried going into Nvidia (going to guess translation cause I have nvidia CP in Italian) control panel under 3D settings -> Global Settings -> "something related to power delivery" set it to prefer maximum performance instead of adaptive.

That solved some issues for me, namely crashing nvidia driver while using Firefox.

 

[maksic87]

Well mate i gues you got no options left, but to borrow a cpu that fits your mobo and give it a try. Of course there is always the possibility of getting a gpu with the same flaw from rma. I got a gpu like that for 3 times, after that i got my money back from the supplier.

Good luck

 

[mastrdrver]

What exactly is the problem again? Kind of hard to help without knowing as nVidia display driver crashing in not exactly an explanation of the problem your experiencing.

 

[95Viper]

What you using to cool that setup?
What case you got? How is the air flow?
What are your ambient temps and what are the temps of the GPU and, maybe, the CPU?
What does the "event viewer" in Windows Admin tools show when this occurs?
Does it crash when certain games are run or in all apps?
Have you tried turning off your A/V & Firewall and testing?
Are you sure that the card is seating in the slot securely and not hitting anything that would block it?
Tried a different slot? (Just for testing)

 

[MxPhenom 216]

I would just do an RMA.

EDIT: Nevermind missed the RMA of the GPU and various other parts.

Motherboard won't change anything I don't think.

 

[de.das.dude]

RMA the i5. i ihave seen a dude have to rma his 2500k 2-3times. non overclocked.

intel may perform well, but it dont last!

 

[MxPhenom 216]

RMA the i5. i ihave seen a dude have to rma his 2500k 2-3times. non overclocked.

intel may perform well, but it dont last!

I have never had an Intel chip "not last" for me, and I have only ran Intel chips for the 6-7 years I have done computer shit, and I upgrade A LOT.

I have also had a Intel Pentium 4 that still runs and I have had it since I was like 7 years old. I am 20 now.

Not sure what you mean by "it don't last!"

 

[Wile E]

I had a similar issue with my 580 on my current 980X rig. On a whim, I bumped my PCIe voltage from 1.5 to the next step up on my mobo, which was either 1.52 or 1.54, and the problem disappeared.

 

[MxPhenom 216]

I had a similar issue with my 580 on my current 980X rig. On a whim, I bumped my PCIe voltage from 1.5 to the next step up on my mobo, which was either 1.52 or 1.54, and the problem disappeared.

Good advice. Try this OP ^^^

 

[de.das.dude]

I have never had an Intel chip "not last" for me, and I have only ran Intel chips for the 6-7 years I have done computer shit, and I upgrade A LOT.

I have also had a Intel Pentium 4 that still runs and I have had it since I was like 7 years old. I am 20 now.

Not sure what you mean by "it don't last!"

it was good back in the day. even i have seen some really old machines running now. reliability seems to be an issue in recent years.
maybe they just cant take the burden of overclocking. XD

 

[renegadexan]

ya my default voltage is 9,62. I haven't seen the bug so far it comes and go, for some reason last night it happened a lot and now, nothing, once it does I'll try to play with the voltages.

My temp are good dont seem to go much over 70 for GPU, I'm using a CNPS9900 MAX for cpu fan.

As far as my power, I moved to 3 different apartment over the 2 year period and the bug happened at each place.

My windows error log pretty much says the same thing all the time, it doesn't seem to be related directly to the crash because I had no crash today and I got an error message.

Activation context generation failed for "C:\Program Files (x86)\NVIDIA Corporation\nTune\MFC80.DLL". Dependent Assembly Microsoft.VC80.MFCLOC,processorArchitecture="amd64",publicKeyToken="1fc8b3b9a1e18e3b",type="win32",version="8.0.50608.0" could not be found. Please use sxstrace.exe for detailed diagnosis.

 

[de.das.dude]

ya my default voltage is 9,62. I haven't seen the bug so far it comes and go, for some reason last night it happened a lot and now, nothing, once it does I'll try to play with the voltages.

My temp are good dont seem to go much over 70 for GPU, I'm using a CNPS9900 MAX for cpu fan.

As far as my power, I moved to 3 different apartment over the 2 year period and the bug happened at each place.

My windows error log are in French but it pretty much says the same thing all the time, it doesn't seem to be related directly to the crash because I had no crash today and I got an error message.

Nom du journal*:Application
Source*: SideBySide
Date*: 2013-08-11 09:14:39
ID de l’événement*:33
Catégorie de la tâche*:Aucun
Niveau*: Erreur
La création du contexte d’activation a échoué pour «*C:\Program Files (x86)\NVIDIA Corporation\nTune\MFC80.DLL*». Assembly dépendant Microsoft.VC80.MFCLOC,processorArchitecture="amd64",publicKeyToken="1fc8b3b9a1e18e3b",type="win32",version="8.0.50608.0" introuvable. Utilisez sxstrace.exe pour un diagnostic détaillé.

what :O

 

[renegadexan]

I found the English version of the error, I edited my post.

 

[unclewebb]

What sort of stress testing have you done on your system? I like MemTest86+ but that's just the first thing you need to run before installing Windows so you know you have some basic stability. After that I would be running Prime95, the Small FFT test as well as the Balanced test and I would also run LinX or IBT to give your system a much more thorough workout.

Prime 95
http://www.mersenne.org/freesoft/

Stasio has a comprehensive list of updated testing programs here:

http://forums.tweaktown.com/gigabyt...-info-benchmarking-stability-tools-print.html

Grab one of the newer versions of LinX or IntelBurnTest.

When you build a new computer with a collection of random parts you need to do some thorough testing.

The other thing is do you know what low power C States your CPU is using? Some users have found the CPU package C States can cause random BSOD problems when lightly loaded. Use RealTemp T|I Edition to check what C States you have set your CPU up to use.

RealTemp T|I Edition
http://www.overclock.net/t/1330144/realtemp-t-i-edition

I would turn off in the bios C3/C6 and the package C States until you find out what is going on.

What core voltage does CPU-Z show? Are you running your CPU at its default speed? Some motherboards will automatically overclock your CPU without your knowledge. I worked on one friend's computer recently and noticed the shop that built it added a big overclock to his computer without him asking for it. His new computer was for his office so an overclock was completely unnecessary. Most shops don't have the time to properly test an overclock so should leave that stuff to the end users. Post some screen shots of CPU-Z and RealTemp while doing some of the above stress tests.

Your problem might be caused by your Nvidia GPU or some bad drivers but you have to eliminate a lot of possible issues first before you start pointing the finger at your graphics card.

 

[kn00tcn]

i vouch for interlburnintest

is the mobo set to auto volts for cpu+ram? does it have a 'loadline calibration' or 'eliminate vdroop' setting?

 

[renegadexan]

Well here is what cpuz and Realtemp shows if it can help, only WoW running, not doing a benchmark.

My bios is using auto turbo and all the auto settings pretty much, I have it set to normal as far as the overclock option in asus bio, so it isn't set to overclock like it could.

I wouldnt pay too much attention on the gpu-z sensors as I didn't enable the option to run full while watching 2D.

Edit: Yea just crashed again, I had voltage at 1V but not while in 2D settings, ima try to put it a 1.05 for everything and see if it works.

 

[Wile E]

Try upping the PCIe voltage one notch in your motherboard bios.

 

[renegadexan]

The default voltage is I think 0964 or something, I put it to 1050 and so far it hasn't crashed, I heard it might even need a boost up to 1100 to be stable. I don't understand why the default settings would be so unstable and why upping the voltage would work, but so far it's ok.

Too soon to tell if it will fix it for good though.

I feel the P67 motherboard might really be a very crappy product.

 

[Wile E]

Not your GPU/video card voltage. I mean the PCIe voltage setting in your motherboard bios.

This is what the setting looks like in my BIOS. Default is 1.5v. Mine is set one notch higher.

 

[renegadexan]

I dont believe my bios allow to change it, it seem to be rather limited.

 

[Wile E]

Well crap. I was hoping it allowed you. It fixed mine perfectly.

 

[WaroDaBeast]

Well... You'd expect to find PCI-E voltage adjustment on a board of that caliber. Guess Asus made some rather strange choices there.

Anyhow... I've got a couple of ideas. You could first test your PCI-E frequency — that is, if you have access to that settings in your BIOS. Try to lock it at 100 or 101 and see if that solves your problem.

Next up is checking if the issue persists under another. Grab a Linux ISO (Ubuntu or its derivatives will do), make a live disc or a live USB with the ISO file. Boot on your CD/DVD drive or USB drive and check whether or not you still have crashes under Linux.

 

[renegadexan]

Only setting I have is PCH voltage.

 

[WaroDaBeast]

Only setting I have is PCH voltage.

The motherboard you got is an Asus P8P67, right? You're telling me you can't adjust frequencies on that motherboard?