Page MenuHomeFeedback Tracker

GPU hangs / DX12 Error, crash since update to 1.2.1.169
Assigned, NormalPublic

Description

Game Version number:
1.2.1.169

Modded?:

  • CORE Pack1 GUID = "5EB139459EBF5C16"
  • CORE Pack2 GUID = "61B8FA7B3BF8656B"
  • MISSION: "Just in Time" GUID = "617B2ED33DE29156"

Issue Description:

Since the last update, GPU crashes have unfortunately been occurring very frequently. This doesn't just affect PC players, the Xbox version also crashes repeatedly. Unfortunately, the log file doesn't help us to narrow down the error.

00:43:10.621 WORLD        : UpdateEntities
00:43:10.621  WORLD        : Frame
00:43:10.621   NETWORK      : PerformActions
00:43:10.621    DEFAULT   (E): InputActionManager @"ENTITY:4611686018427402957" ('SCR_ChimeraCharacter') at <4840.715820 57.680000 5439.483887> @"{29DBADACDD8AE8E3}Prefabs/Characters/Player_US_06.et" CharacterControllerComponent: Action: Trying to perform action CharacterAction_FinishClimbCommand which wasn't queued
00:43:25.693 WORLD        : UpdateEntities
00:43:25.693  WORLD        : Frame
00:43:25.693   NETWORK      : PerformActions
00:43:25.693    DEFAULT   (E): InputActionManager @"ENTITY:4611686018427390721" ('SCR_ChimeraCharacter') at <4972.434082 2.219999 3851.345947> @"{8BA50170236438CF}Prefabs/Characters/Player_US_01.et" CharacterControllerComponent: Action: Trying to perform action CharacterAction_FinishClimbCommand which wasn't queued
00:43:46.809 RENDER    (E): Render error (-2005270523 = 0x887A0005): The video card has been physically removed from the system, or a driver upgrade for the video card has occurred. The application should destroy and recreate the device.
00:43:46.809 RENDER    (E):   func: res
00:43:46.809 RENDER    (E):   C:\jenkins\workspace\continuous_branches_stable_1.2.1\ARGamecode\Enfusion\Enfusion\src\render\dx12\enf_texture_dx12.cpp(636): enf::Texture::AllocateDX12Texture
00:43:46.809 RENDER    (E): Device removed, reason: DXGI_ERROR_DEVICE_REMOVED
00:43:46.809 ENGINE    (F): Crashed
00:43:47.027 ENGINE    (E): Application crashed! Generated memory dump: C:/Users/bs99/Documents/My Games/ArmaReforger/logs/logs_2025-01-04_22-36-27\ArmaReforgerSteam_2025-01-05_00-43-46.mdmp
05.01 2025 00:43:46
Unhandled exception

Program: D:\Steam\steamapps\common\Arma Reforger\ArmaReforgerSteam.exe
Reason: Access violation. Illegal write by 0x7ff6b70c1a93 at 0x0

Crash dump:

Details

Severity
Major
Resolution
Open
Reproducibility
Always
Operating System
Windows 11 x64
Category
General
Steps To Reproduce

Random GPU crashes, see above.

Event Timeline

R34P3R created this task.Jan 5 2025, 9:46 AM
Geez changed the task status from New to Awaiting internal Testing.Jan 6 2025, 1:51 PM
Alienasa added a subscriber: Alienasa.EditedJan 26 2025, 10:03 PM

Been having the exact same problem. Here several more Error logs and memory dumps.

Edit: I use no mods and only play on the main branch.

I find this one especially interesting.

16:39:42.410 RENDER (E): Invalid copy queue fence value (complete = -1, expected 1496982), GPU crash is likely
16:39:42.410 RENDER (E): Render error (-2005270523 = 0x887A0005): The video card has been physically removed from the system, or a driver upgrade for the video card has occurred. The application should destroy and recreate the device.
16:39:42.410 RENDER (E): func: m_pD3DDevice->GetDeviceRemovedReason()
16:39:42.410 RENDER (E): D:\jenkins\workspace\continuous_branches_stable_1.2.1\ARGamecode\Enfusion\Enfusion\src\render\enf_rendmainimpl.cpp(3185): enf::RendererImpl::EndFrame
16:39:42.410 RENDER (E): Device removed, reason: DXGI_ERROR_DEVICE_REMOVED
16:39:42.410 ENGINE (F): Crashed
16:39:42.607 ENGINE (E): Application crashed! Generated memory dump: C:/Users/Bas-PC/Documents/My Games/ArmaReforger/logs/logs_2025-01-24_15-11-31\ArmaReforgerSteam_2025-01-24_16-39-42.mdmp

Jetty added a subscriber: Jetty.EditedJan 26 2025, 11:55 PM

What the fuk... My ticket about this has not been set to "awaiting internal testing" And I posted mine a while back.
I have the same issue. It has been reported several times in steam as well as Discord and other ticket here.

**I do not have any mods tho. And it's the same issue on experimental.

My ticket also includes some troubleshooting and tests that has been done.
If almighty @Geez would have cared.

Jetty added a subscriber: Geez.Jan 26 2025, 11:59 PM

I did a full GPU driver uninstall with DDU and reinstalled the latest drivers. I had one session without crashes yesterday but today another crash after about 3-4 hours of play.

The latest logs and dumps here.

I'm just going to keep uploading crash logs and dump files.

Any change from the previous post is I'm now on Nvidia drivers 572.16. Different drivers don't seem to change anything. Also I disconnected my second monitor in case that caused any issues.

Is any of you using multiple control devices? That might also be a thing. I use mouse and keyboard for infantry and ground vehicles but I use an Xbox controller for flying helicopter. Maybe changing back and forth between them creates anomalous conditions.

Just making another post to inform that a month later I'm still having the same issues.

R34P3R added a comment.EditedFeb 25 2025, 3:31 AM

Same with EXP. 1.3.0.84 - Crash: a9a9a051-bf45-468a-893e-1477ca60d8e9

R34P3R added a comment.EditedFeb 25 2025, 11:51 PM

Experimental Dump:

23:44:25.366 RENDER    (E): Invalid copy queue fence value (complete = -1, expected 1768966), GPU crash is likely
23:44:25.366 RENDER    (E): Render error (-2005270523 = 0x887A0005): The video card has been physically removed from the system, or a driver upgrade for the video card has occurred. The application should destroy and recreate the device.
23:44:25.366 RENDER    (E):   func: m_pD3DDevice->GetDeviceRemovedReason()
23:44:25.366 RENDER    (E):   D:\jenkins\workspace\continuous_branches_stable_1.3.0\ARGamecode\Enfusion\Enfusion\src\render\enf_rendmainimpl.cpp(3138): enf::RendererImpl::EndFrame
23:44:25.366 RENDER    (E): Device removed, reason: DXGI_ERROR_DEVICE_REMOVED
23:44:30.368 RENDER    (E): Parsing DRED output:
23:44:30.368 RENDER    (E):   failed to get page fault allocation data...
23:44:30.368 RENDER    (E):   DRED failed, have you enabled this functionality by D3DConfig?
23:44:30.368 RENDER    (E):     1) add application to debug: D3DConfig apps --add path_to_executable
23:44:30.368 RENDER    (E):     2) allow DRED: d3dconfig.exe dred --force-on-all
23:44:30.368 ENGINE    (F): Crashed
23:44:30.454 RENDER       : ----------------------------------------------------------------
23:44:30.454 RENDER       : ----------------------------------------------------------------
23:44:30.454 RENDER       : Rend: Fps:1000000.00 (0.00ms) | Present 0.00ms | Wait4GPU 0.00ms (graphical 0.00 + copy 0.00) | defrag 0.00ms
23:44:30.454 RENDER       : Rend: DP:0000 DPI:0000 DPR:0000 DPID:0000 Tri:00000 Vert:000000 CSRun:00|AllV:00x(00000) AllI:00x(00000) BegV:00x(00000) BegI:00x(00000)|TRes:0356
23:44:30.454 RENDER       : ----------------------------------------------------------------
23:44:30.641 ENGINE    (E): Application crashed! Generated memory dump: C:/Users/bs99/Documents/My Games/ArmaReforger/logs/logs_2025-02-25_21-26-45\ArmaReforgerSteam_Experimental_2025-02-25_23-44-30.mdmp

And another Crash on Exp. 1.3.0.84 from yesterday

23:33:53.480 RENDER    (E): Invalid copy queue fence value (complete = -1, expected 1568478), GPU crash is likely
23:33:53.480 RENDER    (E): Render error (-2005270523 = 0x887A0005): The video card has been physically removed from the system, or a driver upgrade for the video card has occurred. The application should destroy and recreate the device.
23:33:53.480 RENDER    (E):   func: m_pD3DDevice->GetDeviceRemovedReason()
23:33:53.480 RENDER    (E):   D:\jenkins\workspace\continuous_branches_stable_1.3.0\ARGamecode\Enfusion\Enfusion\src\render\enf_rendmainimpl.cpp(3138): enf::RendererImpl::EndFrame
23:33:53.480 RENDER    (E): Device removed, reason: DXGI_ERROR_DEVICE_REMOVED
23:33:58.487 RENDER    (E): Parsing DRED output:
23:33:58.487 RENDER    (E):   failed to get page fault allocation data...
23:33:58.487 RENDER    (E):   DRED failed, have you enabled this functionality by D3DConfig?
23:33:58.487 RENDER    (E):     1) add application to debug: D3DConfig apps --add path_to_executable
23:33:58.487 RENDER    (E):     2) allow DRED: d3dconfig.exe dred --force-on-all
23:33:58.487 ENGINE    (F): Crashed
23:33:58.567 RENDER       : ----------------------------------------------------------------
23:33:58.567 RENDER       : ----------------------------------------------------------------
23:33:58.567 RENDER       : Rend: Fps:1000000.00 (0.00ms) | Present 0.00ms | Wait4GPU 0.00ms (graphical 0.00 + copy 0.00) | defrag 0.00ms
23:33:58.567 RENDER       : Rend: DP:0000 DPI:0000 DPR:0000 DPID:0000 Tri:00000 Vert:000000 CSRun:00|AllV:00x(00000) AllI:00x(00000) BegV:00x(00000) BegI:00x(00000)|TRes:0217
23:33:58.567 RENDER       : ----------------------------------------------------------------
23:33:58.753 ENGINE    (E): Application crashed! Generated memory dump: C:/Users/bs99/Documents/My Games/ArmaReforger/logs/logs_2025-03-01_21-14-33\ArmaReforgerSteam_Experimental_2025-03-01_23-33-58.mdmp

I do have the Same issue in Arma Reforger Tools while using the Terrain Editor.
My crash Files are attached. AHving this issue on multiple occasions during the last weeks.

Geez changed the task status from Awaiting internal Testing to Assigned.
Geez added a subscriber: edible_lemon.
Exp. 1.3.0.102 - Crash:  210a16bd-8e97-4dc1-bd55-7391b3108135

Alienasa added a comment.EditedMon, Mar 17, 7:49 PM

Another one! On the main branch.

Following this guide for Ark: Survival Evolved (uses unreal engine) seems to have fixed my issue, been playing for 2 days with no crashes as of yet.
https://www.reddit.com/r/ARK/comments/17plsn4/possible_fix_for_dxgi_error_device_removed_with/

R34P3R added a comment.EditedThu, Mar 20, 11:27 AM

@edible_lemon do you know what this reg entrys do?

i already have such entrys, but with a different Value.

EDIT:

https://learn.microsoft.com/de-de/windows-hardware/drivers/display/tdr-registry-keys

@R34P3R Not too familiar with regedit but looking at the documentation:
"TdrDdiDelay
Indicates how many seconds the operating system allows threads to exit the driver. After a certain time, the operating system checks the computer with the code VIDEO_TDR_FAILURE (0x116)."

Would guess its something to do with how long it waits for threads to exit the driver, perhaps there is some bug with unreal engine not allowing enough time by default? - note your screenshot shows "TdrDelay" and not "TdrDdiDelay".

Update: The reg changes does not fix the problem. Had the same Crash yesterday in 1.3.130 and i already changed the values in the registry.

Geez added a comment.Mon, Mar 31, 11:20 AM

Hello everyone.
Can you please try to run the experimental release and provide us a crash dump from that one?
Thank you

@Geez Hi Geez, did not have a Dump from Experimental, because we switched all to the current public build. But here is a Dump from the current 1.3.130 public, maybe this also helps.

00:32:24.981 RENDER    (E): Invalid copy queue fence value (complete = -1, expected 2045415), GPU crash is likely
00:32:24.981 RENDER    (E): Render error (-2005270523 = 0x887A0005): The video card has been physically removed from the system, or a driver upgrade for the video card has occurred. The application should destroy and recreate the device.
00:32:24.981 RENDER    (E):   func: m_pD3DDevice->GetDeviceRemovedReason()
00:32:24.981 RENDER    (E):   C:\jenkins\cpu\workspace\continuous_branches_stable_1.3.0\ARGamecode\Enfusion\Enfusion\src\render\enf_rendmainimpl.cpp(3138): enf::RendererImpl::EndFrame
00:32:24.981 RENDER    (E): Device removed, reason: DXGI_ERROR_DEVICE_REMOVED
00:32:24.982 ENGINE    (F): Crashed
00:32:25.244 ENGINE    (E): Application crashed! Generated memory dump: C:/Users/bs99/Documents/My Games/ArmaReforger/logs/logs_2025-03-28_22-01-07\ArmaReforgerSteam_2025-03-29_00-32-25.mdmp

Geez added a comment.Mon, Mar 31, 11:33 AM

We will require experimental specificaly, thank you

I'm not running experimental sadly and sometimes it takes like 3 hours of play to get the crash so it's unlikely I will invest the time to do that. Here however is another log from after the 1.3.0.144 update.