Game Crashes Nvidia Display Driver (April 2015)

[[DE]Linkski]

In regards to anyone who has had their Nvidia display driver crash while playing Warframe --

Could I get a few people to record their GPU temperatures around the time this crash occurs? If you do not have a program to record temperatures with then you can use this one: http://openhardwaremonitor.org/

 

Please observe your temperature as you progress through the game and let me know what it reaches when you crash. Once you believe you have been able to reliably reproduce the crash then I would suggest changing your V-sync and/or Nvidia PhysX Effects settings. -- then record your temperature again.

 

Unfortunately we are limited in the number of in-house machines that we have been able to reproduce the issue on so any additional information from the community would be greatly appreciated. I understand how frustrating this can be. 

 

Thanks everyone. 

[HurricaneBones]

It's been 2 weeks i noticed the game was jerky like the game was removing frames per second even if the game told me i got 40 fps. (my wife is playing on it, so the problem may be far older)

Then I got 2 blue screen of the death last week while playing (my computers never did this on me)

i have an gamin' nvidia laptop.

 

I looked on my other gamin' nvidia laptop who had no problem (beside that warframe is always warming my computers for winter (which is coming))

 

2 main differences : (computers have same configs nvidia GTX660M)

- driver :

           - the one jerky got an older driver but approved by my computer manufacturer 331.65

           - the one working great had a newer one but not approved by my computer manufacturer 347.52

 

- launcher config :

          - direct X10(+X11) + full screen on the jerky one

          - direct X11 + windowed on the great one.

 

I changed the launcher to direct X11 + windowed and since then the game is now smooth and no crash to signal (yet)

Edited May 12, 2015 by HurricaneBones

[Me.KayOh]

Update: apparently dx10 and dx11 were suddenly turned off in my launcher settings (idk how. i didnt do it) and i was able to play fullscreen without crashing. occasionally the game would still freeze for a split second, seeming like its going to crash, but then continue normally...reproduce the crash?

 

ah man...i'll see what i can do.

[weezedog]

Had my NVIDIA driver crash again today. Seems to only happen in the Void for me. Temperature is only 60C at the time of the crash, definitely not overheating. I've uploaded my EE.log to a support ticket #352099 Now I noticed in the log right before the display driver crash starts, there are buffer underruns occuring.

 

 

7236.449 Snd [Error]: Buffer underrun streaming /Lotus/Music/DynamicMusic/Orokin/VoidOne/OrokinVoidOneCmbtPerc.wav (0 active streams)

7236.449 Snd [Error]: Buffer underrun streaming /Lotus/Music/DynamicMusic/Orokin/VoidOne/OrokinVoidOneCmbtPercB.wav (0 active streams)

7236.450 Snd [Error]: Buffer underrun streaming /Lotus/Music/DynamicMusic/Orokin/VoidOne/OrokinVoidOneCmbtPercC.wav (0 active streams)

7236.450 Snd [Error]: Buffer underrun streaming /Lotus/Music/DynamicMusic/Orokin/VoidOne/OrokinVoidOneCmbtStringsPerc.wav (0 active streams)

7236.450 Snd [Error]: Buffer underrun streaming /Lotus/Music/DynamicMusic/Orokin/VoidOne/OrokinVoidOneCmbtSynths.wav (0 active streams)

7236.450 Snd [Error]: Buffer underrun streaming /Lotus/Music/DynamicMusic/Orokin/VoidOne/OrokinVoidOneCmbtHorns.wav (0 active streams)

7236.450 Snd [Error]: Buffer underrun streaming /Lotus/Music/DynamicMusic/Orokin/VoidOne/OrokinVoidOneCmbtStringsPercB.wav (0 active streams)

7236.450 Snd [Error]: Buffer underrun streaming /Lotus/Music/DynamicMusic/Orokin/VoidOne/OrokinVoidOneCmbtStringsChoir.wav (0 active streams)

7236.450 Snd [Error]: Buffer underrun streaming /Lotus/Sounds/Ambience/Orokin/Interior/OroRingRotateLargeA.wav (0 active streams)

7236.450 Gfx [Warning]: D3DAssert: Dx11OcclusionQuery::Create() returned 0x887A0005 (DXGI_ERROR_DEVICE_REMOVED)

7236.450 Gfx [Warning]: D3DAssert: Dx11OcclusionQuery::Create() returned 0x887A0005 (DXGI_ERROR_DEVICE_REMOVED)

7236.450 Gfx [Warning]: D3DAssert: Dx11OcclusionQuery::Create() returned 0x887A0005 (DXGI_ERROR_DEVICE_REMOVED)

7236.450 Gfx [Warning]: D3DAssert: Dx11OcclusionQuery::Create() returned 0x887A0005 (DXGI_ERROR_DEVICE_REMOVED)

7236.451 Gfx [Warning]: D3DAssert: Dx11OcclusionQuery::Create() returned 0x887A0005 (DXGI_ERROR_DEVICE_REMOVED)

7236.451 Gfx [Warning]: D3DAssert: Dx11OcclusionQuery::Create() returned 0x887A0005 (DXGI_ERROR_DEVICE_REMOVED)

7236.451 Gfx [Warning]: D3DAssert: Dx11OcclusionQuery::Create() returned 0x887A0005 (DXGI_ERROR_DEVICE_REMOVED)

Edited May 12, 2015 by weezedog

[[DE]Linkski]

Can anyone experiencing this crash please try rolling back to the 345.20 drivers? If your card does not have a 345.20 driver (Titan X, etc.) then can you please roll back to the drivers closest to February 24, 2015? (so for Titan X that should be the 347.88 drivers)

 

Once you have done this then feel free to play the game as you normally would and then report back to this thread with your findings.

 

Thanks again.

Edited May 14, 2015 by DE_Dmitri

[weezedog]

NVIDIA driver crash again. This time I was playing Loki on a syndicate spy mission on a corpus tile set. I've uploaded the log (Loki_Corpus_Spy_EE.log) to support ticket 352099 for you to take a look at. The driver crash occurs at 12387.187 in the log.

I've got a whole bunch of NVIDIA driver error logs in the Windows Event Logs I could add too, if you would like.

Edited May 15, 2015 by weezedog

[[DE]Linkski]

NVIDIA driver crash again. This time I was playing Loki on a syndicate spy mission on a corpus tile set. I've uploaded the log (Loki_Corpus_Spy_EE.log) to support ticket 352099 for you to take a look at. The driver crash occurs at 12387.187 in the log.

I've got a whole bunch of NVIDIA driver error logs in the Windows Event Logs I could add too, if you would like.

 

From your log: "NVIDIA driver version is: 35012"

 

Can you try rolling back your drivers to 345.20 and see if that helps at all?

[weezedog]

I can certainly try to

[Sonicbullitt]

Can anyone experiencing this crash please try rolling back to the 345.20 drivers? If your card does not have a 345.20 driver (Titan X, etc.) then can you please roll back to the drivers closest to February 24, 2015? (so for Titan X that should be the 347.88 drivers)

 

Once you have done this then feel free to play the game as you normally would and then report back to this thread with your findings.

 

Thanks again.

 

Rolled back to those drivers, have dx11 , dx10 and 64 bit mode unchecked and so far warframe is running fine. If I experience more crashes i'll be be sure to let you know

[EWar]

Hi there.

 

I am suffering the same problem since I switched from an AMD Radeon HD6950 2GB to a Nvidia GTX 980.

 

GPU is not faulty.

After I got my first 980 which was a EVGA GTX 980 SuperClocked ACX 2.0 I experience those crashes and thought it's the card.

I RMA'd the card and got a brand new EVGA GTX 980 FTW ACX 2.0 for it, sporting the exact same crashes.

 

Display driver crashes range from none to 5+ a day.

 

Increasing the TDR Delay to 15 seconds only reduced the amount of display driver crashes, but did not solve it.

Disabling PhysX in Warframe seem to further reduce the crashes.

 

Now, temperatures and GPU load are on another page.

At Liset and post-round screen I have about 75°C (Temp Target is 80°C, fans are at ~50%)

and the card hits about 70-85% of the Power Target. Game runs at 300+ FPS.

 

In-game during missions the card is dipping below 60 fps on high complex scenes and median about 80-100 FPS.

BUT the card is below 70°C and only hits 50% or less of the Power Target all the time.

 

I have already set the power preference in the control panel to "Prefer Maximum Performance".

I feel like the game cannot tap into the full potential of the card somehow.

 

Here are the readings: http://imgur.com/a/kQuJU

 

No idea what is happening here.

Underclocking to 980 factory speeds did not help.

Is downgrading to 345.20 the only valid solution here?

 

System:

Windows 8.1 Pro x64

16 GB DDR3-1866 RAM

Intel Core i7-4770K

EVGA GTX 980 FTW ACX 2.0

[Sonicbullitt]

 

Hi there.

 

I am suffering the same problem since I switched from an AMD Radeon HD6950 2GB to a Nvidia GTX 980.

 

GPU is not faulty.

After I got my first 980 which was a EVGA GTX 980 SuperClocked ACX 2.0 I experience those crashes and thought it's the card.

I RMA'd the card and got a brand new EVGA GTX 980 FTW ACX 2.0 for it, sporting the exact same crashes.

 

Display driver crashes range from none to 5+ a day.

 

Increasing the TDR Delay to 15 seconds only reduced the amount of display driver crashes, but did not solve it.

Disabling PhysX in Warframe seem to further reduce the crashes.

 

Now, temperatures and GPU load are on another page.

At Liset and post-round screen I have about 75°C (Temp Target is 80°C, fans are at ~50%)

and the card hits about 70-85% of the Power Target. Game runs at 300+ FPS.

 

In-game during missions the card is dipping below 60 fps on high complex scenes and median about 80-100 FPS.

BUT the card is below 70°C and only hits 50% or less of the Power Target all the time.

 

I have already set the power preference in the control panel to "Prefer Maximum Performance".

I feel like the game cannot tap into the full potential of the card somehow.

 

Here are the readings: http://imgur.com/a/kQuJU

 

No idea what is happening here.

Underclocking to 980 factory speeds did not help.

Is downgrading to 345.20 the only valid solution here?

 

System:

Windows 8.1 Pro x64

16 GB DDR3-1866 RAM

Intel Core i7-4770K

EVGA GTX 980 FTW ACX 2.0

 

 I have a gtx 650 and after rolling back my drivers to version to 345.20 I have experienced no crashes. Warframe has been running for a full day without crashing so far, so yes rolling back to those drivers does seem to fix the issue.

[weezedog]

 I have a gtx 650 and after rolling back my drivers to version to 345.20 I have experienced no crashes. Warframe has been running for a full day without crashing so far, so yes rolling back to those drivers does seem to fix the issue.

 

I'd give it more time, my driver crashes were sporadic. I could go a week or more with no issues, then get 1 or more crashes in the same day.

[Damhaet]

I have a EVGA GTX 980 FTW ACX 2.0, i5 2500K, Win7 64 and my drivers also crash. The game freezes for a second, then I see my Dektop but I am still in the game. I hear sounds, can fire my weapon, do everything. I just cant see anything. Switching windows with ALT-TAB does not restore the Warframe-Window.

 

EE File shows this:

 

432.575 Game [info]: HeavyFemaleGrineerAvatar1130 client side kill! dd: 1087
440.090 Sys [info]: 1 melee attacks were aborted after receiving EndMeleeAction
448.782 Gfx [Error]: Dx11TextureMgr::Present() failed, hr = 0x887A0005
448.782 Gfx [Error]: Present returned DXGI_ERROR_DEVICE_REMOVED:
449.965 Gfx [Error]: LevelUpSuitRadialA failed to allocate vertex buffer data, disabling.
449.973 Sys [info]: Texture streaming NORMAL -> DISABLED [Heap: 779,327,824/797,507,584]
449.975 Sys [Warning]: LevelUpSuitRadialA: Reallocating particles, this should only happen after mode resets!
452.191 Gfx [Error]: SyndicateUISparksBurst2d failed to allocate vertex buffer data, disabling.
452.194 Sys [Warning]: SyndicateUISparksBurst2d: Reallocating particles, this should only happen after mode resets!

 

These two lines get spammed hundreds of times per second after this with different variable names:

 

452.281 Gfx [Error]: GrineerDeathExplosionFlowSparks failed to allocate vertex buffer data, disabling.
452.281 Sys [Warning]: GrineerDeathExplosionFlowSparks: Reallocating particles, this should only happen after mode resets!

 

Another example:

 

367.469 Game [info]: EliteRifleLancerAvatar135 client side kill! dd: 1975
371.588 Gfx [Error]: Dx11TextureMgr::Present() failed, hr = 0x887A0005
371.588 Gfx [Error]: Present returned DXGI_ERROR_DEVICE_REMOVED:
372.837 Gfx [Error]: RifleHitSpangLines failed to allocate vertex buffer data, disabling.
372.837 Gfx [Error]: RifleHitSpangShockwave failed to allocate vertex buffer data, disabling.
372.837 Gfx [Error]: RifleHitSpang failed to allocate vertex buffer data, disabling.
372.837 Gfx [Error]: RifleHitSpangLines failed to allocate vertex buffer data, disabling.
372.837 Gfx [Error]: RifleHitSpangBurstLines failed to allocate vertex buffer data, disabling.
372.838 Gfx [Error]: RifleHitSpang failed to allocate vertex buffer data, disabling.
372.838 Gfx [Error]: RifleHitSpangLines failed to allocate vertex buffer data, disabling.
372.838 Gfx [Error]: RifleHitSpangBurstLines failed to allocate vertex buffer data, disabling.
372.838 Gfx [Error]: RifleHitSpangShockwave failed to allocate vertex buffer data, disabling.
372.838 Gfx [Error]: RifleHitSpang failed to allocate vertex buffer data, disabling.
372.841 Gfx [Error]: SyndicateUISparksBurst2d failed to allocate vertex buffer data, disabling.
372.842 Gfx [Error]: SyndicatePowerFX failed to allocate vertex buffer data, disabling.
372.846 Sys [info]: Texture streaming NORMAL -> DISABLED [Heap: 721,810,464/729,874,432]
372.854 Sys [Warning]: RifleHitSpang: Reallocating particles, this should only happen after mode resets!
372.854 Sys [Warning]: RifleHitSpangLines: Reallocating particles, this should only happen after mode resets!
372.854 Sys [Warning]: RifleHitSpangShockwave: Reallocating particles, this should only happen after mode resets!
372.854 Sys [Warning]: RifleHitSpang: Reallocating particles, this should only happen after mode resets!
372.854 Sys [Warning]: RifleHitSpangLines: Reallocating particles, this should only happen after mode resets!
372.854 Sys [Warning]: RifleHitSpangBurstLines: Reallocating particles, this should only happen after mode resets!
372.854 Sys [Warning]: RifleHitSpang: Reallocating particles, this should only happen after mode resets!
372.854 Sys [Warning]: RifleHitSpangLines: Reallocating particles, this should only happen after mode resets!
372.854 Gfx [Error]: RifleHitSpangLines failed to allocate vertex buffer data, disabling.

 

It happens nearly every mission, after a few minutes (When I play the Event-Sabotage I reach the Terminal to flood the base and crash. This happened multiple times). I tried the 345.20 and the 350.12 drivers, reinstallled both at least two times.

Edited May 17, 2015 by Damhaet

[Sonicbullitt]

I'd give it more time, my driver crashes were sporadic. I could go a week or more with no issues, then get 1 or more crashes in the same day.

 It has been a few days now and the game is still running fine with not a single crash since I rolled back.

Edited May 18, 2015 by Sonicbullitt

[Zruvan]

Overclocking the card seems to be the real cause, try bringing the clocks to reference design specs or even lower. It did fix my crashes.

[weezedog]

Got an NVIDIA driver crash with the downgraded driver version 345.20. Uploaded the ee.log to support ticket 352099, file name, Driver_345.20_EE.log.

I noticed the driver crash is slightly different with driver version 345.20. When it crashes to the desktop, a second later the game crashes triggering the warframe crash reporter, rather than the game engine continuing to run in the background indefinitely like it did with the newest driver.

 

The nvidia driver logs the following to the windows event logs

 

The following information was included with the event: 

 

\Device\UVMLiteProcess4

Graphics Exception: ESR 0x404490=0x80000002

 

The following information was included with the event: 

 

\Device\UVMLiteProcess4

Graphics Exception: EXTRA_MACRO_DATA

 

Edited May 19, 2015 by weezedog

[Me.KayOh]

i've been crashing recently as well. I was playing void t1 survival last night and crashed 3 times in-game. i'd be playing, and then it would just freeze for one second. the windows would then switch and i'd be put back onto my desktop and warframe would seem as though its minimized. but then it would end and a window would come up saying the evolution engine has crashed. it happened just now again so that makes it 4 times.

 

out of the four times, only once did it say my Nvidia display driver has stopped working and has recovered from an error and that was last night... i also got a code when this happened which was WAR-460499.

 

and of course, it ate the key i used and the rewards gained were lost as a result of this. For me, warframe has never crashed so consistently.

 

my current specs are:

 

ASUS Maximus VI Extreme Motherboard ROG Edition

Intel Core i5 4670k 3.4 GHz Quad Core

Kingstone HyperX DDR3 16.0GB

ASUS Strix nVidia GTX 970 (GPU Clock - 1253 MHz, Memory Clock - 7010 MHz)

its me again. sorry i have not tried to replicate this problem, as i was too busy playing the event. however after running in windowed mode for some time, i switched back to full screen and for many days, i had no problems at all with everything enabled.

 

then nvidia drivers updated yesterday for me.

you all released a hotfix today and for a good few hours everything ran smoothly until i was in the void and i began losing frames quickly. then eventually the game minimized and i could not open it back but it was still running as i could hear the sound.

 

i wanted to get back in the mission immediately so i exited the client and restarted and rejoined the mission. i was successful in doing so, then the game kicked me from the mission. 

 

i got reinvited and managed to finish the mission. it was a t3 MD. then i started another one with my own key but someone else as the host. and in the mission i began to lose frames again.. really badly...then the game said the connection to the host was lost and it was returning me back to the menus

 

since it was my key it kicked everyone from the game

 

so i restarted the client in windowed mode and was able to do the mission to completion successfully without losing frames....

 

i did not catch the video card temperature when this occurred though. tmrw if it happens again i'll try to get it.

[BladedSoul]

I'm having this same issue. A couple of minutes ago, in a Void mission (it happens A LOT in Void missions in my case, by the way), my game froze for a second, then my monitor screen turned black and I was back to the desktop. I could still play and hear all sounds playing in the game, but I just could not switch back to the game screen. No Alt+Tab and not even clicking on the Warframe icon on the taskbar.

 

My computer specs: AMD Phenom X4 955 CPU, EVGA GTX 970, 8GB RAM, Windows 7 64 bits, running Warframe on DirectX 11 and my NVidia drivers are up to date at 350.12.

I basically have the same issue as you except my screen stays black and I don't get booted to desktop.

A few days ago, probably one or two patches ago I could play through very well. A few crashes here or there but not as frequent as I have them now. I don't even get the crash reporter like I did before

Specs: Lenovo y410p, Intel i7-4700MQ @ 2.4GHz, 8GB ram, Windows 8 64bit, running DX11 on nvidia 750m updated to latest drivers (also tried previous drivers but also crashing)

[[DE]Xander]

I basically have the same issue as you except my screen stays black and I don't get booted to desktop.

A few days ago, probably one or two patches ago I could play through very well. A few crashes here or there but not as frequent as I have them now. I don't even get the crash reporter like I did before

Specs: Lenovo y410p, Intel i7-4700MQ @ 2.4GHz, 8GB ram, Windows 8 64bit, running DX11 on nvidia 750m updated to latest drivers (also tried previous drivers but also crashing)

It sounds in your case that the display driver is crashing but failing to recover. 

 

We released shader optimizations this morning that might remove some of these display driver crashes due to memory. I hope to hear back from any/all of you if there's a change.

[Fate_Epsylon]

I basically have the same issue as you except my screen stays black and I don't get booted to desktop.

 

 That's not a driver issue. That's a common windows 8 issue. There are a few videos on youtube that show how to fix this problem. I had this for a while as well. (About a year ago)

[pauli133]

It sounds in your case that the display driver is crashing but failing to recover. 

 

We released shader optimizations this morning that might remove some of these display driver crashes due to memory. I hope to hear back from any/all of you if there's a change.

I turned on dx11 this morning on my GTX 970 rig (latest driver), to see if the patch helped - hard locked on wave 18 of T1D. So, in one case, no.

[Hooride]

Been having same issue with driver crashing and doesnt recover constantly with gtx 960.

 

 

 

 

Win 7 x64

16Gb 1866 ram

Intel i5 4590

Msi Gtx 960

Edited May 21, 2015 by Hooride

[kildyu]

I've been having these crashes as well recently. Just today I had a bluescreen with the DPC_WATCHDOG_VIOLATION, unfortunately my page file was not on the same hard drive as my OS, so the dump wasn't saved. When a driver crash occurs there is no extra information in the event viewer - just "Display driver nvlddmkm stopped responding and has successfully recovered."

 

EVGA 980 ACX 2.0 Driver versions 350.12(Bluescreen) and 352.86(No bluescreen yet, updated today)

i7 3770k

32Gb 1866 ram

Win 8.1 64bit

 

Here are some things from my EE.log

 

6141.191 Gfx [Warning]: D3DAssert: Dx11OcclusionQuery::Create() returned 0x887A0005 (The GPU device instance has been suspended. Use GetDeviceRemovedReason to determine the appropriate action.)

 

6142.076 Gfx [Error]: Dx11TextureMgr::Present() failed, hr = 0x887A0005

6142.076 Gfx [Error]: Present returned DXGI_ERROR_DEVICE_REMOVED:

 

6142.166 Sys [info]: Texture streaming NORMAL -> DISABLED [Heap: 648,825,728/866,844,672]

6142.176 Gfx [Error]: GrineerShieldHitSparks failed to allocate vertex buffer data, disabling.

6142.187 Sys [Warning]: GrineerShieldHitSparks: Reallocating particles, this should only happen after mode resets!

6142.359 Gfx [Error]: MeleeSlamSparks failed to allocate vertex buffer data, disabling.

 

6142.786 Sys [Warning]: LastStandAttachFlow: Reallocating particles, this should only happen after mode resets!

6142.786 Gfx [Error]: LastStandAttachFlow failed to allocate vertex buffer data, disabling.

 

6143.598 Gfx [Error]: GrineerDeathLeadUpExhaust failed to allocate vertex buffer data, disabling.

6143.599 Sys [Warning]: GrineerDeathLeadUpExhaust: Reallocating particles, this should only happen after mode resets!

 

6143.922 Gfx [Error]: KneeSlideSparks failed to allocate vertex buffer data, disabling.

6143.922 Sys [Warning]: KneeSlideSparks: Reallocating particles, this should only happen after mode resets!

 

6146.466 Sys [Warning]: KneeSlideSmoke: Reallocating particles, this should only happen after mode resets!

6146.466 Gfx [Error]: KneeSlideSmoke failed to allocate vertex buffer data, disabling.

 

There's about 10,000 of the last 4 lines over 9 seconds in the log (I'm guessing the first number in the line is seconds, they go from 6142-6151).

 

I will try some of the things listed in this thread and see if any work.

Edited May 22, 2015 by kildyu

[Urlan]

That actually is pretty good info, we would just need to know what caused the device, the Graphics card, to go into a stopped state. It would be interesting if you still got a similar crash while turning off Direct x 10 and 11 as that particular error is a Direct x 11 specific error message; the messages afterward could mean that either the textures were not finished in memory or the device was still partially unavailable for rendering the texture memory. I am going to go with the card was not recovered yet, but it would be best to test whether you get the same issue using only direct x 9 mode in the launcher.

 

I would think that the core problem is the graphics driver trying to run Direct x 11 calls and failing but I wouldn't rule out hardware issues until we test the card using other versions of Direct x from the launcher. Users back in January had a similar issue with Arma 3 using Intel 4000 integrated cards on their laptops. By the errors there and here, it may be the same scenario with a different card. In the Arma intel HD card issue, going back to a much earlier driver corrected their issue until Windows 8.1 tried to update the driver to a newer version. Perhaps rolling back a bit might help.

[Schweinekaramell]

We released shader optimizations this morning that might remove some of these display driver crashes due to memory. I hope to hear back from any/all of you if there's a change.

Last driver crash was yesterday 03:26 PM EDT.

 

GTX 960, Win7x64