Sign in with
Sign up | Sign in
Your question

NVidia driver stop responding... looking for the 'why'

Last response: in Graphics & Displays
Share
March 10, 2012 11:07:48 PM

Ok I'll try to make ir short. First, these are my specs:

Intel i5-750
Motherboard Asus P7P55D Deluxe
8 GB RAM DDR3 1600 Mhz
3 HDs (320 GB - 1 TB - 500 GB)
DVDRW Optical Drive
Power Suppley Thermaltake TR2 RX-850 W0319RU (v2.3)
MSi N580GTX
Windows 7 64 Ultimate

Well, the problem I have is that randomly I could receive a warning that my driver has stopped working and has recovered.

It happened when playing Battlefield 3, League of Legends, and Batman Arkham City. Sometimes it woulnd't happen at all.

Games would generally will be forced to minimize when this happens, and I can't return to them, so I need to close the application (except League of Legends that would cause my screen to go black for a second and then recover by itself).

Before I had this new brand MSi videocard, I bought a EVGA 580 GTX, and I had to return it because it was preciselly returning that random error message when playing. I didn't note the problem with BF3 that time, but I would suffer it when playing League of Legends.

After I returned the EVGA thinking it was flawed somehow, I installed the new MSi 580 GTX and here comes the oddity... the problem persisted when playing League of Legends! So I thought it was an application related issue.

Time later, I bought a new LG LED Monitor and I started to play the previous games in HD. And was then when the problem presented itself this time playing Batman Arkham City and BF3 also.

I can't believe I just bought two faulty videocards from two different providers. It makes no sense. I can rule out RAM from the equation because I recently changed 2x2GB 1333 Mhz modules for the current 2x4GB 1666 Mhz.

I didn't perform any overclocking in any element of my PC. I use MSi Afterburner and when the problem appears, I would note that every value in the graph will go down and at the same time I will suffer a 100% GPU Usage spike.

I was thinking that maybe the PSU is somehow not feeding well my videocard, but just I don't know, I'm guessing.

My PSU is a couple of years old but I never had any weird problem besides this one. I do know that my PSU has 4 lines of 12v, 18A each. I noticed that recent versions of the same PSU come with an unified 12v line.

So I suppose my question is... could my PSU be the reason behind the cited problems?
a c 91 Î Nvidia
March 10, 2012 11:29:39 PM

It very well could be the power supply. The Thermaltake TR2 is a smoking bag of crap on a good day so it really would not surprise me. If you can try the computer with a different power supply you can confirm if that is the problem.

Have you tried uninstalling the driver and reinstalling it?
m
0
l
March 11, 2012 3:28:37 AM

anort3 said:
It very well could be the power supply. The Thermaltake TR2 is a smoking bag of crap on a good day so it really would not surprise me. If you can try the computer with a different power supply you can confirm if that is the problem.

Have you tried uninstalling the driver and reinstalling it?


Yes I tried the whole uninstall/reinstall thingy weeks ago with no avail.

Today I made a test using OCCT of my videocard for about 10 minutes, no errors and 70° max temperature reached.

However, when I try to test my PSU with OCCT it would finish quickly with an error because my CPU reaches 85° so the test stops for security reasons.

I have no way to test with another PSU except buying a new one, but since I'm not sure my PSU is the problem, I'm hesitating to make such purchase. I need to be sure.

What shoould I do next?
m
0
l
Related resources
Anonymous
a b Î Nvidia
March 11, 2012 3:33:52 AM

what driver are you using?

i just rolled back to 270 because 295 was giving me the same headaches in crysis.
m
0
l
March 11, 2012 12:46:03 PM

Anonymous said:
what driver are you using?

i just rolled back to 270 because 295 was giving me the same headaches in crysis.


Hi there.

I'm using the 295 right now, and before the 285.

The problem don't seem to be related to the drivers.

Yesterday I made a little test, I played for about two minutes BF3 and checked in background with OCCT, my CPU was over 90°... is that even normal?
m
0
l
a c 91 Î Nvidia
March 11, 2012 4:37:49 PM

No those temps are too high. You also mentioned your CPU reaching 85C and that should never happen even with stock cooling as long as you are not overclocking on it.

What case do you have? Sounds like you have very poor to no airflow. Use some monitoring software and note your max temps while stressing the computer. Then try taking the side off and do the same thing and report the results.

Have you cleaned the computer with compressed air lately?
m
0
l
March 11, 2012 5:10:52 PM

anort3 said:
No those temps are too high. You also mentioned your CPU reaching 85C and that should never happen even with stock cooling as long as you are not overclocking on it.

What case do you have? Sounds like you have very poor to no airflow. Use some monitoring software and note your max temps while stressing the computer. Then try taking the side off and do the same thing and report the results.

Have you cleaned the computer with compressed air lately?


Yes I've doing some research and these values are too high!

Today tested BF3 by about 2 minutes and temps of my cores according to OCCT reached:

CPUTIN 98°C
Core #0 88°C
Core #1 97°C
Core #2 82°C
Core #3 99°C

But, in the meantime, my GPU just reached 60°C

When iddle, the cores will stay between 35° and 40°.

My case is a Thermaltake, it's pretty big and with sufficient air flow.

I'm buying a new cooler for my i5 this week, I don't honestly know if this is the cause of my problem, but definetly must be addressed.

I'll try with the side off and check back here.
m
0
l
March 11, 2012 5:15:59 PM

Well tried with the side off, same results.

Changing my stock cooler ASAP. Reporting later this week with results.
m
0
l
a c 91 Î Nvidia
March 11, 2012 6:00:49 PM

Sounds like the cooler you have now is not seated correctly. Unless it is completely clogged with dust and dirt. I hate the stock Intel cooler, it's easy to break and can be hard to tell if it's mounted correctly.

If you are going with aftermarket cooling then the Hyper 212+ or Hyper 212 Evo are good low cost options. They will also fit with most RAM even RAM with high heat spreaders.
m
0
l
March 16, 2012 1:10:28 PM

Well here I am reporting the news.

I just installed a new cooler for my CPU, Zalman CNPS5 Performa.

Doing a load test I noticed how much the temperature went down. It's amazing. Now I never reach the 50°C under full load.

Also, I noticed something in the connection to my motherboard. The EATXPWR connection was only partially connected. I mean, it was connected to the PSU with just 4 pins instead the 8 I'm able to. Weird, I mean, it was working that way, but maybe not enough power was being pumped into the system?

Well, I'll do some tests the next days and report any news.
m
0
l
March 17, 2012 1:42:08 AM

Well I played by 15 minutes BF3 and the problem appeared again, and this time even I was able to see some green artifacs.

Same pattern, sudden 100% GPU spike and all the other values going down.

I'm out of options right now.

I can't believe both the EVGA and the MSi are flawed or damaged. Makes no sense.

I don't know what else to do.
m
0
l
a c 91 Î Nvidia
March 17, 2012 2:42:18 AM

Still could be that crappy Thermaltake power supply. The TR2 is an especially bad unit.
m
0
l
March 17, 2012 11:14:14 AM

anort3 said:
Still could be that crappy Thermaltake power supply. The TR2 is an especially bad unit.


I know but I can't be sure. The only way is to buy a new PSU and if the problem persists I'll be even more pissed.
m
0
l
March 17, 2012 4:08:07 PM

anort3 said:
Try clearing CMOS either with the motherboard jumper or by removing the battery for 10 minutes.

Review of the 750w version of your power supply.

http://www.hardwaresecrets.com/article/Thermaltake-TR2-...


Wow that review destroys the PSU lol.

Before the two 580 videocards, I was using a 550 GTi and I never had any kind of problem.

Today I made two video stress tests using OCCT for over 15+ minutes, not a single error detected.

Later I made the PSU test that will stress both CPU and GPU for over 10 minutes, no error whatsoever.

Cleaned the registry using CCleaner. Modified the TdrDelay parameter on registry without avail.
m
0
l
a c 91 Î Nvidia
March 17, 2012 4:13:30 PM

The GTX580 stresses your power supply a whole lot more than a GTX550 Ti. As you can see from the review stressing that power supply is a bad thing that can damage components and cause instability and crashes.
m
0
l
March 18, 2012 12:01:53 AM

anort3 said:
The GTX580 stresses your power supply a whole lot more than a GTX550 Ti. As you can see from the review stressing that power supply is a bad thing that can damage components and cause instability and crashes.


Yeah I got that, but I can't understand how the tests will not generate any error and these are stressing the system a lot more than games like BF3. In fact BF3 will crash with a suspicious sudden 100% GPU use spike, but before it, the average use was about 75%-80% and the temps around 60°. In the other hand, the GPU test will put the GPU use almost to 100% all the time and the temps will get around 80°.

I'll continue testing, I downclocked a bit some parameters with MSI Afterburner.

Reporting as fast I got news.
m
0
l
March 18, 2012 12:57:08 AM

Well after 4 or 5 matches of BF3 the problem appeared again.

According to the Event Manager, this is what always happens:

In some random point, under the Application tab:

'A request to disable the Desktop Window Manager was made by process (4)' being the origin the Desktop Windows Manager, event ID 9010.

At the same time, also Desktop Windows Manager under event ID 9013 informs that it can't be restarted because an application make it so.

4 sencods later, under the System tab on Event Manager appears the classic 'nvlddmkm' error under event ID 4101 cited as a 'warning'.

It's always the same pattern.

I understand that the cause of such error could be virtually ANYTHING.

Should I think I bought 2 flawed 580 videocards? Sigh...

m
0
l
a c 91 Î Nvidia
March 18, 2012 1:05:30 AM

You KNOW you have a flawed power supply that can be dangerous to your components. A bad power supply can also cause general system instability. So what's logical? Two bad graphics cards in a row? Or a bad PSU causing all your problems? ;) 

The power supply is the single most important part of a computer. Since you were willing to fork out $500 or better for a GPU don't you think it's a good idea to protect that investment with a good quality power supply?
m
0
l
March 18, 2012 1:22:42 AM

anort3 said:
You KNOW you have a flawed power supply that can be dangerous to your components. A bad power supply can also cause general system instability. So what's logical? Two bad graphics cards in a row? Or a bad PSU causing all your problems? ;) 

The power supply is the single most important part of a computer. Since you were willing to fork out $500 or better for a GPU don't you think it's a good idea to protect that investment with a good quality power supply?


Yeah I agree with you, but I'm gonna be REALLY pissed when I buy a new PSU and still the problem persists. I almost can feel I will not resolve this particular issue that way. But well, I guess I should make a try.

BTW, what PSU should I pick? I was reading that Corsair make nice ones. What about the wattage according to my specs? Corsair Tx850m 850w Enthusiast maybe? Do I really need 850w?
m
0
l
a c 91 Î Nvidia
March 18, 2012 1:40:14 AM

You only need 850 watts if you plan on adding a second GTX580. For a single GTX580 a quality 550-600w PSU will be fine.

Seasonic makes the best power supplies. They actually make many of the power supplies for Corsair, XFX, PC Power and Cooling and quite a few other good brands.

Everything you ever wanted to know about power supply makers.

http://www.realhardtechx.com/index_archivos/Page541.htm

A few brands always stand out as not having bad models. Corsair, Seasonic, PC Power and Cooling, XFX, Enermax, Silverstone, OCZ and Antec are almost always high quality.

This is on great sale and is modular.

http://www.newegg.com/Product/Product.aspx?Item=N82E168...
m
0
l
March 18, 2012 1:50:14 AM

anort3 said:
You only need 850 watts if you plan on adding a second GTX580. For a single GTX580 a quality 550-600w PSU will be fine.

Seasonic makes the best power supplies. They actually make many of the power supplies for Corsair, XFX, PC Power and Cooling and quite a few other good brands.

Everything you ever wanted to know about power supply makers.

http://www.realhardtechx.com/index_archivos/Page541.htm

A few brands always stand out as not having bad models. Corsair, Seasonic, PC Power and Cooling, XFX, Enermax, Silverstone, OCZ and Antec are almost always high quality.

This is on great sale and is modular.

http://www.newegg.com/Product/Product.aspx?Item=N82E168...


Thanks for your guide. As you say, a 550-600w PSU should be fine, so how is possible that my Thermaltake can't handle it if it were the case? I mean, I know it's a faulty PSU according to the reviews, but damn, it's over 200+ W of what I really need.

Also, I'm trying to find any other case where somebody solved a similar problem replacing the PSU. I can't find none! Everytime it's almost the same case, a faulty videocard. It's full of similar topics and almost everytime it's the GPU.

Now I'm depressed lol.
m
0
l
a c 91 Î Nvidia
March 18, 2012 1:58:04 AM

The problem with the TR2 is ripple and noise. It does not put out "clean" power. Hell it could even be the power supply damaging the video cards. HardwareSecrets and jonnyguru are the 2 best power supply review sites on the net. If they call a unit dangerous I believe them. I would replace it in my system even if I was not having any problems.
m
0
l
March 18, 2012 2:13:59 AM

Thanks again. Some points to clarify my case.

In the case of the EVGA I noticed the problem the very same day I installed it when playing League of Legends. That's why I returned it.

The same happened with the MSi, very same day I got it, but since Skyrim and BF3 were doing it well I tought it was software related. The problems with BF3 started like 1 month and half later when I bought the LG LED Monitor and started to play in 1360*768 resolution. Before that I had a crappy monitor so I used to play BF3 in 1024x768.

The funny part is that NOW League of Legends works fine, except in a couple of times in like 1 month when the same error happened, but a lot, lot less than before, and when it used to fail more (daily) was when I was playing it in 1024x768, and now I'm playing it in 1080 dpi. How does that make any sense?
m
0
l
a b Î Nvidia
March 18, 2012 3:23:54 AM

This could have no relevance to your issue at all, but I believe it does.

When I first built my computer, it was pretty awesome. 965BE, 6950, 8GB DDR2, etc. Now it played everything flawlessly for quite some time, until one day, I got the exact same problem you seem to be having. When doing something graphics-intensive, it would run for a few minutes fine, and the screen would go black, and come back, saying "AMD driver has stopped responding and has recovered." I played LoL like this for a good month until I got tired of it happening. My friend gave me his "broken" Nvidia 8800GT, which I fixed, and I thought it solved the problem, but low and behold it still happened. I figured the card was actually broken, and took out my fully working 8800GTX, (what the 6950 replaced) and the problem still persisted. I even swapped power supplies and most of the parts in the computer.

Raging about my computer one day while staring at the insides of it, I figured "might as well" and took 2 RAM sticks out of my computer, so I was running on 4GB, and low and behold THE PROBLEM WAS GONE. It never happened again after that. Since then I have used that 965 in another build, but I thought this might be helpful to you.
m
0
l
a c 91 Î Nvidia
March 18, 2012 3:26:59 AM

You were BARELY even using the GPU at 1024 x 768 and even at 1366 x 768 you could max every game with a GTX 460 / HD 6850. You are very CPU bound at either resolution. You have to have a 1920 x 1080 monitor to be full HD. A GTX 580 is pretty much even overkill for a single 1920 x 1080 monitor.

What we know.

CPU temps were high but now are fixed.

Not driver related.

Not RAM related.

2 cards with identical issues.

That pretty much leaves the power supply. Especially since it's known to be flawed.
m
0
l
a c 91 Î Nvidia
March 18, 2012 3:30:24 AM

Deemo13 said:
This could have no relevance to your issue at all, but I believe it does.

When I first built my computer, it was pretty awesome. 965BE, 6950, 8GB DDR2, etc. Now it played everything flawlessly for quite some time, until one day, I got the exact same problem you seem to be having. When doing something graphics-intensive, it would run for a few minutes fine, and the screen would go black, and come back, saying "AMD driver has stopped responding and has recovered." I played LoL like this for a good month until I got tired of it happening. My friend gave me his "broken" Nvidia 8800GT, which I fixed, and I thought it solved the problem, but low and behold it still happened. I figured the card was actually broken, and took out my fully working 8800GTX, (what the 6950 replaced) and the problem still persisted. I even swapped power supplies and most of the parts in the computer.

Raging about my computer one day while staring at the insides of it, I figured "might as well" and took 2 RAM sticks out of my computer, so I was running on 4GB, and low and behold THE PROBLEM WAS GONE. It never happened again after that. Since then I have used that 965 in another build, but I thought this might be helpful to you.


^ Not a badf idea.


Even though you have used 2 different sets of RAM and the problem has remained you might as well run memtest overnight to be sure. It's something that can be ruled out for sure.

http://www.memtest.org/
m
0
l
March 18, 2012 10:52:12 AM

Deemo13 said:
This could have no relevance to your issue at all, but I believe it does.

When I first built my computer, it was pretty awesome. 965BE, 6950, 8GB DDR2, etc. Now it played everything flawlessly for quite some time, until one day, I got the exact same problem you seem to be having. When doing something graphics-intensive, it would run for a few minutes fine, and the screen would go black, and come back, saying "AMD driver has stopped responding and has recovered." I played LoL like this for a good month until I got tired of it happening. My friend gave me his "broken" Nvidia 8800GT, which I fixed, and I thought it solved the problem, but low and behold it still happened. I figured the card was actually broken, and took out my fully working 8800GTX, (what the 6950 replaced) and the problem still persisted. I even swapped power supplies and most of the parts in the computer.

Raging about my computer one day while staring at the insides of it, I figured "might as well" and took 2 RAM sticks out of my computer, so I was running on 4GB, and low and behold THE PROBLEM WAS GONE. It never happened again after that. Since then I have used that 965 in another build, but I thought this might be helpful to you.


Thanks for your answer, I read something similar that happened to another user regarding the same problem. A couple of weeks ago I ran Memtest and no error was detected. Besides, the problems started as fast I try to use my system with 580 videocards. Like I said, I had a 550 GTi like for 3 months before getting the EVGA 580 and I never had any kind of problem. And something else, I changed my previous 2x2GB 1333 Mhz. DDR3 memories 1 month ago and got 2x4GB 1666 Mhz new ones, so I suppose I can pretty rule out this being memory related.
m
0
l
March 18, 2012 10:59:46 AM

anort3 said:
You were BARELY even using the GPU at 1024 x 768 and even at 1366 x 768 you could max every game with a GTX 460 / HD 6850. You are very CPU bound at either resolution. You have to have a 1920 x 1080 monitor to be full HD. A GTX 580 is pretty much even overkill for a single 1920 x 1080 monitor.

What we know.

CPU temps were high but now are fixed.

Not driver related.

Not RAM related.

2 cards with identical issues.

That pretty much leaves the power supply. Especially since it's known to be flawed.


Indeed, or it could be some sort of 'problem' with my motherboard too. Or even software related.

I'm going to do this:

1) Try with a new PSU. If persists...

2) Install a new HD with a clean install of Windows 7, updated it, only install Origin and BF3 and test again. If persists...

3) Move the recently acquired PSU and the videocard and move it to a different PC, clean install of Windows 7, Origin and BF3 and test again.

I suppose this works this way:

If I solve it in step 1, the problem was the PSU.

If solved in step 2, the problem is software related.

If solved in step 3, some problem with my whole system, maybe motherboard, RAM configuration, or whatever.

If not solved in anyone of the above, the videocard is damaged. Which gives me a new nightmare since I don't know if I can RMA it now.

It seems I have work to do... sigh.
m
0
l
March 22, 2012 5:59:52 PM

Tried phase 1 today, problem persists. New PSU didn't solve the problem. Got a Corsair 750.

So I installed a fresh copy of Windows 7, updated, installed Origin, BF3 and MSI Afterburner, nothing else.

Testing starts again...
m
0
l
March 24, 2012 12:26:10 AM

Well today after playing less than 2 hours artifacts on screen and game minimized again, with the same driver error.

Called my vendor, I have 2 more months of warranty. I'm RMAing this ASAP.
m
0
l
a c 91 Î Nvidia
March 24, 2012 3:03:45 AM

I have to think if you have 2 bad cards in a row that your old PSU may have damaged them. The odds are like winning the lottery to get 2 bad ones in a row.
m
0
l
March 26, 2012 12:49:13 AM

anort3 said:
I have to think if you have 2 bad cards in a row that your old PSU may have damaged them. The odds are like winning the lottery to get 2 bad ones in a row.


I'm starting to think the same. I hope my vendor can corroborate the failure and give me a new card.
m
0
l
March 29, 2012 10:29:12 PM

Well after 3 days of 'RMA' the tech guy told me that he couldn't reproduce the errors I described. He will continue testing tough, but I'm starting to believe he will come up saying that everything seems to be fine. He claimed to test using games, stress programs and becnh programs. I asked him if he tried playing BF3, he told me he didn't have that particular game for test right now.

No way I'm letting him return this videocard. This is becoming ugly...
m
0
l
a c 91 Î Nvidia
March 30, 2012 12:12:01 AM

Man. I hate these "Ghost in the Machine" problems. At this point I have no clue.

A bad PCI-Express slot does not make sense. If it was the board or your CPU was damaged from getting so hot then you would have errors in more than gaming.

A hard drive having random errors maybe? Have you used a program like Speedfan or Passmark's DiskCheckup to check for SMART errors?

http://www.passmark.com/products/diskcheckup.htm

http://www.almico.com/speedfan.php
m
0
l
March 30, 2012 1:12:31 AM

anort3 said:
Man. I hate these "Ghost in the Machine" problems. At this point I have no clue.

A bad PCI-Express slot does not make sense. If it was the board or your CPU was damaged from getting so hot then you would have errors in more than gaming.

A hard drive having random errors maybe? Have you used a program like Speedfan or Passmark's DiskCheckup to check for SMART errors?

http://www.passmark.com/products/diskcheckup.htm

http://www.almico.com/speedfan.php


Just impossible, because when I made a clean install I used a completely different HD, in fact, I also disconnected the other two I was using in first place.

Today I updated my motherboard BIOS but since I haven't the card I can't test again. My motherboard is an Asus P7P55D LE, I suppose if it were a PCI-e problem I would have issues with the previous 550 GTi aswell, thing that never happened.

In fact, I tried BOTH PCI-E connectors with the conflicting videocard and the problem remains.

I noticed that my RAM modules weren't properly configured, these were running at 1333 Mhz. I just changed this enabling a feature in my mainboard and now they're running at 1600 MHz. as supposed. Concurrently the CPU was sightly overclocked because of this.

But considering I had completely different modules around 1 month ago I can pretty rule out that.

I was checking the MSI forum and the mod there would usually suggest to rise the voltage a bit. I remember my videocard was set at 1.025v with charge, he will usually suggest to increase it a bit, like to 1.038v or even 1.050v. I must admit I never tried that, just downclocked some values like memory clock from 2040 to 2000 and core clock from 800 to 772 if I recall correctly. I'm not sure about the values but it was something along these lines.

I find pretty odd that as fast I got the EVGA 580 GTX every time I tried to play League of Legends I'll receive at least 1 error.

Changed the videocard for the MSi 580 GTX and the problem continued! How can it be? You suggested that maybe the former PSU damaged both cards... but SO FAST? And the same pattern? I find it extremelly odd.

It's like somehow my system just can't hande these videocards, but why? The only parts I didn't replaced were the motherboard and the CPU. I'm lost but mostly mad, because I'll end having a strong discussion if they try to send me back this card.
m
0
l
March 30, 2012 4:05:22 PM

Guess what, they confirmed and reproduced the error! The change of the card was authorized! Geez finally!
m
0
l
March 30, 2012 4:46:41 PM

I'm so gettin' an ATI 7950 now lol... I don't wanna have anything to do with any 580 ever again.
m
0
l
a c 91 Î Nvidia
March 30, 2012 5:28:36 PM

Well that's good news. Hopefully I was correct in that the Thermaltake PSU was damaging your cards and now that you have a Corsair you will have no more problems.

Now you just need a 3 x 1920 x 1080 monitor setup so you actually need a graphics card that high end. Or at least a single 30" 2560 x 1600 monitor. Either card is just complete overkill even for a single true HD 1920 x 1080 monitor. ;) 
m
0
l
March 30, 2012 7:18:03 PM

anort3 said:
Well that's good news. Hopefully I was correct in that the Thermaltake PSU was damaging your cards and now that you have a Corsair you will have no more problems.

Now you just need a 3 x 1920 x 1080 monitor setup so you actually need a graphics card that high end. Or at least a single 30" 2560 x 1600 monitor. Either card is just complete overkill even for a single true HD 1920 x 1080 monitor. ;) 


lol true... I'm gonna get an ATI 7970 after all, even better! I hope I have no problems at all... I never had an ATI before.

About the causes we'll never know I guess, it could be a faulty GPU since the very beginning, maybe the PSU damaged it both, I just don't know.

Maybe this thread will help other users aswell when dealing with this mysterious and sadly common 'driver' error.

BTW, the RMA guy told me they tested the card with 3DMark 2011 and Heaven Benchmark 2.5, both programs will reproduce the error within the 20 minutes of testing.
m
0
l
!