Is Anyone Else Experiencing Nvidia Driver Crashes? Is This Warframe Or The Gpu/driver?

[Aure7]

Using Nvidia GTX 650 Ti and now using beta drivers because the old drivers did not help and it doesn't make any difference anyway, it keeps crashing. Multi-threaded does not make a difference either.

With PhysX on it crashes even more often so i have to keep it off, what's the point in having nvidia card with PhysX when it keeps crashing with it?

 

Windows 7 64bit
Intel Core i5 3570 (stays at low temp, 55C-60C max I think)
GeForce GTX 650 Ti 1Gb (stays cool too, framerate is very clean and high at maxed settings and temp never goes above 50C)
8 Gb RAM
600 W power supply http://www.chieftec..../gps-600a8.html

 

playing 64bit mode, dx11.

 

Re installed drivers (and looks like updated them at the same time) played WITH PHYSIX on for like 3 days without any crashing, then it all returned again, with or without physx, nvidia driver crashes.

Edited September 27, 2013 by Aure7

[Saenol]

What's the exact error you are getting?

 

It's not likely that Warframe is responsible for an NVIDIA driver BSOD.

[Aure7]

What's the exact error you are getting?

 

It's not likely that Warframe is responsible for an NVIDIA driver BSOD.

 

That's a nvidia driver crash, not BSOD. How can I possibly make it more clear?

 

It stops working and then recovers quickly but the game can't deal with that and shuts down or leaves a "evolution engine" process running in the background.

[Schmetterlein]

same for me try to diable dx11 an 64 bit my drivers crash constantly only in this game i have gtx 570 oc

Edited September 22, 2013 by Schmetterlein

[Saenol]

That's a nvidia driver crash, not BSOD.

 

A BSOD is a form of crash and can certainly be precipitated by NVIDIA drivers.

 

From your original post I have no way of knowing that you aren't talking about this:

 

 

How can I possibly make it more clear?

 

You could post exactly what's in the event log, or otherwise detail the error you are getting.

 

"Display driver Nvidia Windows Kernel Mode driver, driverversionhere has stopped responding and has successfully recovered." Often isn't an application issue either.

 

Without more details, I would double check the stability of your CPU, memory, and GPU with some stress tests.

Edited September 22, 2013 by Saenol

[Aure7]

A BSOD is a form of crash and can certainly be precipitated by NVIDIA drivers.

 

From your original post I have no way of knowing that you aren't talking about this:

 

 

 

You could post exactly what's in the event log, or otherwise detail the error you are getting.

 

"Display driver Nvidia Windows Kernel Mode driver, driverversionhere has stopped responding and has successfully recovered." Often isn't an application issue either.

 

Without more details, I would double check the stability of your CPU, memory, and GPU with some stress tests.

 

I've heard bad things about nvidia and it's latest drivers while googling, so that's why I assume that it's nvidia drivers not working properly with warframe. (and some other games) It annoys me how easy it is for you to just push me away by saying "check the stability or whatever" but I can't argue even though the chance that my parts are broken is really low. If you're talking about overheating or overclocking, then neither of them is happening, my gpu and cpu are all chill and stock-clocked.

[FunkDoctaSpock]

Have you tried reverting to a previous driver version ?

[mattklanks]

Good luck, been looking for a fix since Update 8.

[Saenol]

I've heard bad things about nvidia and it's latest drivers while googling, so that's why I assume that it's nvidia drivers not working properly with warframe. (and some other games) It annoys me how easy it is for you to just push me away by saying "check the stability or whatever" but I can't argue even though the chance that my parts are broken is really low. If you're talking about overheating or overclocking, then neither of them is happening, my gpu and cpu are all chill and stock-clocked.

 

You stated that older drivers didn't help, which makes me think a driver issue is unlikely. Additionally, I am unable to reproduce this error on any of my NVIDIA graphics cards (I have a GTX 275, a GTX 480, and a 555M GT).

 

I'm not being dismissive, or pushing you away, by suggesting a stability problem. I work with a multitude computers on a daily basis, and in my 20 odd years of experience, I'd conservatively estimate that 1-in-5 bone stock systems has some sort of configuration or hardware issue that makes them not quite stable. Also, in my experience, a significant portion of GPU driver errors are related to system memory instability, or faulty PSUs.

 

Stability testing is simply basic troubleshooting, and hopefully a way to narrow down the problem.

 

Anyway, we are now on post #10, and you still haven't given the precise error or more detailed system configuration issue.

[Aure7]

You stated that older drivers didn't help, which makes me think a driver issue is unlikely. Additionally, I am unable to reproduce this error on any of my NVIDIA graphics cards (I have a GTX 275, a GTX 480, and a 555M GT).

 

I'm not being dismissive, or pushing you away, by suggesting a stability problem. I work with a multitude computers on a daily basis, and in my 20 odd years of experience, I'd conservatively estimate that 1-in-5 bone stock systems has some sort of configuration or hardware issue that makes them not quite stable. Also, in my experience, a significant portion of GPU driver errors are related to system memory instability, or faulty PSUs.

 

Stability testing is simply basic troubleshooting, and hopefully a way to narrow down the problem.

 

Anyway, we are now on post #10, and you still haven't given the precise error or more detailed system configuration issue.

What are you talking about, log files? I've already sent them few months ago to support and all they suggested was turning off physX until they fix it. Everything I managed to figure leads to a "driver crash" that I've told you several times already. Game freezes for a little, shuts down (or sometimes disappears with engine process running in the background) and I see a a pop up message that nvidia drivers have stopped working and recovered. I will try to stress test the components somehow soon, 1/5 ratio sounds very awful.

[Saenol]

What are you talking about, log files? I've already sent them few months ago to support and all they suggested was turning off physX until they fix it. Everything I managed to figure leads to a "driver crash" that I've told you several times already. Game freezes for a little, shuts down (or sometimes disappears with engine process running in the background) and I see a a pop up message that nvidia drivers have stopped working and recovered. I will try to stress test the components somehow soon, 1/5 ratio sounds very awful.

 

 

I'm not DE support, so I don't have your logs.

 

I'm pretty sure I have a good idea of the error you are getting, but since I can't reproduce it myself, the specific entry in Event Viewer "eventvwr.msc" in Run, then find the specific error under administrative alerts) might help.

 

Also, what CPU, memory, motherboard, PSU, and OS are you using? Any custom driver options? Newest OS updates? Have you verified the game's files/cache with the installer? Are you using DX11 and/or 64-bit for Warframe?

[Aure7]

I'm not DE support, so I don't have your logs.

 

I'm pretty sure I have a good idea of the error you are getting, but since I can't reproduce it myself, the specific entry in Event Viewer "eventvwr.msc" in Run, then find the specific error under administrative alerts) might help.

 

Also, what CPU, memory, motherboard, PSU, and OS are you using? Any custom driver options? Newest OS updates? Have you verified the game's files/cache with the installer? Are you using DX11 and/or 64-bit for Warframe?

Posted specs in OP. Please don't worry too much now, I will play around with various things and drivers for the last time in upcoming days, test things if possible and will come back if I encounter more crashes.

[Aure7]

Played for like 3 days with PhysX on when re installed and updated drives to 327.23 But then it all began again with or without PhysX.

How should I stress test my GPU? Do you recommend FurMark?

[Saenol]

FurMark is a good stress test (despite what some may say), but it doesn't have an artifact scanner so it can be difficult to tell when errors exist. It will likely reveal if your GPU has cooling issues or if your PSU is unable to deliver enough power to it. Pay close attention to temperatures while running it.

 

OCCT is potentially even more demanding, and has an artifact scanner that will reveal rendering errors you normally would not be able to see. Again, if you use it, keep a close eye on GPU temperatures.

 

I would recommend looping something like Unigine Valley Benchmark to better test the GPU's memory.

[Aure7]

FurMark is a good stress test (despite what some may say), but it doesn't have an artifact scanner so it can be difficult to tell when errors exist. It will likely reveal if your GPU has cooling issues or if your PSU is unable to deliver enough power to it. Pay close attention to temperatures while running it.

 

OCCT is potentially even more demanding, and has an artifact scanner that will reveal rendering errors you normally would not be able to see. Again, if you use it, keep a close eye on GPU temperatures.

 

I would recommend looping something like Unigine Valley Benchmark to better test the GPU's memory.

should I run Burn-in on FurMark? Also is it that important to test the memory? I really don't want to spend money on that benchmark only because I am unlucky with my hardware.

[Hameln]

Just turn off the 64 bits. It's crash me every single time.

[TheRipper7]

happens to me to i use the Nvidia GT640 M and whenenver i turn on Physx my game always crashes

[Aure7]

OK I guess it was my gpu that was unstable. Running FurMark test it crashed on the first few minutes.

I've read that on crysis 2 it was useful to increase your voltage just a tiny bit (or downclock) and it would get rid of the crashes, however I failed to test this on warframe and just ignored it thinking it wouldn't help. Now I've tried increasing voltage by 12mV and it ran 15 minutes of FurMark test with no problem, max temp being something like 65 or 67 Celsius.

So Saenol I am asking you now. Could it be that my gpu was overclocked with factory settings so it made it unstable? I am really starting to experience more and more crashes on other games and FurMark just proved it as well. I have not changed anything about my gpu after I bought it, just increased voltage sometimes for crysis 2 and now for FurMark.

NVIDIA Inspector says that
Current Clock, GPU clock and default clock are all 1032Mhz

Default Voltage is 1.100 V

 

[Saenol]

Could be the GPU or the PSU.

 

Anyway, when is the last time you cleaned out this system, including the PSU and the fins of the GPU heatsink?

 

If you GPU is not stable at the clock speeds it comes with, with the stock voltages, it's defective or broken and you should get warranty service on it.

 

Rule out the PSU as a possibility first, if you have any way to do so. A failing PSU can potentially damage or kill everything connected to it.

[Aure7]

Could be the GPU or the PSU.

 

Anyway, when is the last time you cleaned out this system, including the PSU and the fins of the GPU heatsink?

 

If you GPU is not stable at the clock speeds it comes with, with the stock voltages, it's defective or broken and you should get warranty service on it.

 

Rule out the PSU as a possibility first, if you have any way to do so. A failing PSU can potentially damage or kill everything connected to it.

well I just played loads of warframe with increased voltage, PhysX on and no crashes. So I guess it was an unstable gpu?

I've cleaned my system few days ago.

[Saenol]

well I just played loads of warframe with increased voltage, PhysX on and no crashes. So I guess it was an unstable gpu?

 

This is likely the case.

 

Really, there is no excuse for them sending you a part that is not stable at their specified stock clocks and voltages. I would RMA this GPU if at all possible.

[Aure7]

This is likely the case.

 

Really, there is no excuse for them sending you a part that is not stable at their specified stock clocks and voltages. I would RMA this GPU if at all possible.

Well thank you for all the help. Then for the final part, how much voltage is too much? I crashed once with +12.5mv so I tried +25mv and it seems stable and very cool, like 56 celcius max. I've heard that you should mostly worry only about your temps when playing around with voltage.

But I still think warframe is a little unstable. I've played loads of other games with stock voltage that are very demanding and draining my fps to 20-30, but the game runs fine.

Edited September 30, 2013 by Aure7

[c.BOLTSON]

Hey , I run an old GTX 460 here. Personaly I'm running OLD drivers because they perform better as absurd as that sounds. (driver 310. or something or other, from December 2012)

I recently updated my drivers to run benchmark tests in 3DMark. The test results, for the first time I might add, show me this latest driver did yield a very minor performance improvement. I thought great, the only problem was as soon as I got this new Nvidia driver I started getting multiple crashed while playing Warframe. Something I never had before.

I reverted back to my old 310. something driver (late 2012 driver) and now were good again.

However I hav experienced a performance drop since that last very small patch for Warframe that came out yesterday or whenever it was. (The update was like only kilobytes or megabytes, very small update)

Edited October 5, 2013 by c.BOLTSON

[Saenol]

Well thank you for all the help. Then for the final part, how much voltage is too much? I crashed once with +12.5mv so I tried +25mv and it seems stable and very cool, like 56 celcius max. I've heard that you should mostly worry only about your temps when playing around with voltage.

 

How much voltage is too much depends on a lot of other factors, but in general, 25mv extra is not going to be harmful.

 

Still, the point still stands that your card is supposed to be virtually perfectly stable in everything without any extra voltage at all, and if it's not, it's not living up to advertised specifications and is defective.

 

Hey , I run an old GTX 460 here. Personaly I'm running OLD drivers because they perform better as absurd as that sounds. (driver 310. or something or other, from December 2012)

 

Not absurd sounding.

 

Newer drivers are tweaked for newer architectures. A 460, which is a few generations old at this point, is no longer the focus of optimizations, and what is good for newer parts can occasionally cause a minor loss of performance with older ones.

[4shae]

from what i think is that

they try so hard to improve graphics and stuffs like that, pushing it way off their limit and stuff, and now, they neglected small and minor bugs, and im sure this is one of them which they neglected.....