solved DCS ST and MT crashing.

[Akula108]

Hello all,

I have recently stumbled upon an issue where my DCS in both Single threaded and Multithreaded DCS are crashing within half an hour of playing single player or multiplayer. Listed below are the attempted fixes I have tried, which all have failed, followed by system specs. 

Attempted fixes.
1. Set page file to 32GB (even tried 64GB)
2. DDU to uninstall and reinstall latest version of graphics driver.
3. Updated to the latest version of Windows 11.
4. Launched game with debug mode on in Nvidia Control Panel.
5. Lowered in game graphics settings.
6. Uninstalled all user mods under C drive.
7. Deleting fxo and metashaders2 folders and rebuilding.
8. Memtest via windows and memtest86 (No errors). 
9. Renamed C drive DCS to rebuild files.

System Specs:
Asus Rog Z790 Extreme
i9 13900k (not overclocked)
RTX 4090 MSI Suprim Liquid X
128GB DDR5 on XMP at 5200mhz
5x 2TB Samsung 990 PRO
Windows 11 Pro
Samsung Odyssey neo G9

I have ruled out hardware issues as I am able to play equally intensive (if not more intensive) titles such as Ray-traced Cyberpunk 2077, Star Citizen, BF2042, and many other titles for hours without crashes. Temperatures, voltage, and stability are also nominal during benchmarking software. At the moment DCS is the only application that has issues on my PC. I have attached the logs and crash logs in this post. I genuinely do not know how to proceed. 

Many Thanks. 

 

dcs.log-20230501-022153.zip dcs.log

Edited May 2, 2023 by Akula108

[BIGNEWY]

Hi, 

is your 4090 overclocked? or has a factory overclock? 

 

Spoiler

2023-05-01 02:21:53.309 WARNING DX11BACKEND (10832): Can't map structured buffer.
2023-05-01 02:21:53.309 ERROR   DX11BACKEND (10832): Failed assert `0` at Projects\render\dx11backend_win8\Source\DX11StructuredBuffer.cpp:100
2023-05-01 02:21:53.311 ERROR   DX11BACKEND (10832): DX device removed. Reason: 0x887A0006
2023-05-01 02:21:53.312 WARNING WORLD (Main): ModelTimeQuantizer: ANTIFREEZE ENABLED
2023-05-01 02:21:53.312 ERROR   DX11BACKEND (16452): Failed assert `false && "failed to create texture"` at Projects\render\dx11backend_win8\Source\DX11TextureManager.cpp:711
2023-05-01 02:21:53.312 INFO    EDCORE (16452): try to write dump information
2023-05-01 02:21:53.314 INFO    EDCORE (16452): # -------------- 20230501-022153 --------------
2023-05-01 02:21:53.316 INFO    EDCORE (16452): DCS/2.8.4.39313 (x86_64; MT; Windows NT 10.0.22621)
2023-05-01 02:21:53.317 INFO    EDCORE (16452): D:\Eagle Dynamics\DCS World OpenBeta\bin-mt\GraphicsCore.dll
2023-05-01 02:21:53.318 INFO    EDCORE (16452): # C0000005 ACCESS_VIOLATION at 00007ffe6231e168 00:00000000
2023-05-01 02:21:53.319 INFO    EDCORE (16452): SymInit: Symbol-SearchPath: 'D:\Eagle Dynamics\DCS World OpenBeta\bin-mt;', symOptions: 532, UserName: 'David'
2023-05-01 02:21:53.321 INFO    EDCORE (16452): OS-Version: 10.0.22621 () 0x100-0x1
2023-05-01 02:21:53.322 INFO    EDCORE (16452): 0x000000000003e168 (GraphicsCore): render::RenderGraph::~RenderGraph + 0x2E48
2023-05-01 02:21:53.322 INFO    EDCORE (16452): 0x0000000000041fd3 (GraphicsCore): render::RenderGraph::computeOptimalBarriers + 0x403
2023-05-01 02:21:53.323 INFO    EDCORE (16452): 0x0000000000041ace (GraphicsCore): render::RenderGraph::bakeGraph + 0x18FE
2023-05-01 02:21:53.323 INFO    EDCORE (16452): 0x000000000000e20f (SceneRenderer): DCSSceneRenderer::bakeRenderGraph + 0x3F
2023-05-01 02:21:53.323 INFO    EDCORE (16452): 0x0000000000107f27 (Visualizer): smCamera_Implement::getClipRegion + 0x17837
2023-05-01 02:21:53.323 INFO    EDCORE (16452): 0x00000000000483fb (edCore): ed::this_thread::yield + 0x1531B
2023-05-01 02:21:53.323 INFO    EDCORE (16452): 0x00000000000476d2 (edCore): ed::this_thread::yield + 0x145F2
2023-05-01 02:21:53.323 INFO    EDCORE (16452): 0x0000000000048f73 (edCore): ed::this_thread::yield + 0x15E93
2023-05-01 02:21:53.323 INFO    EDCORE (16452): 0x00000000000326b1 (edCore): ed::thread::_get_current_thread_id + 0x71
2023-05-01 02:21:53.323 INFO    EDCORE (16452): 0x0000000000029363 (ucrtbase): _recalloc + 0xA3
2023-05-01 02:21:53.323 INFO    EDCORE (16452): 0x00000000000126ad (KERNEL32): BaseThreadInitThunk + 0x1D
2023-05-01 02:21:53.548 INFO    EDCORE (16452): Minidump created.

 

[Akula108]

My 4090 has a factory overclock from MSI. However, launching the game in debug mode in the Nvidia control panel via Help ---> Debug Mode should turn off the factory overclock correct? I just want to clarify because I have already tried playing the game with Debug Mode enabled and the crashes persist. 

[BIGNEWY]

6 minutes ago, Akula108 said:

My 4090 has a factory overclock from MSI. However, launching the game in debug mode in the Nvidia control panel via Help ---> Debug Mode should turn off the factory overclock correct? I just want to clarify because I have already tried playing the game with Debug Mode enabled and the crashes persist. 

Yes using debug mode would be the best way to check, if you have done that already then we can rule that out also. 

Could you post your nvidia settings here also 

thank you

[Akula108]

Unsure if there was a more efficient way to do this, but I have posted all submenus where I have made changes to the Nvidia control panel.

I want to also note that I did monitor my GPU clock speeds during debug mode on and off (listed below) incase any of it helps.

With debug mode off, my 4090 clocked 2830 mhz.
With debug mode on, my 4090 clocked 2760 mhz.
 

Edited May 2, 2023 by Akula108

[BIGNEWY]

That is fine thank you 

Can you disable vsync in DCS and retest, let me know if you still get the crash

[Akula108]

1 minute ago, BIGNEWY said:

That is fine thank you 

Can you disable vsync in DCS and retest, let me know if you still get the crash

Apologies, I should have mentioned this in the original post, the game crashes with Vsync on and off, with full-screen enabled.
Additionally, I have just crashed two more times, I have attached the two newest crash files below. I have noticed in the files there are four lines that are present in both,

0x00000000001059b5 (d3d11): CreateDirect3D11SurfaceFromDXGISurface + 0x10535
0x000000000002db71 (dx11backend): createRenderer + 0x15C51
0x000000000007edbb (edterrainGraphics41): edtg41::TerrainRenderable::dumpRenderItem + 0x69A1B
0x00000000000d6a8c (edterrainGraphics41): createInstancerRenderable + 0x38FEC

 

dcs.20230502-061149.crash dcs.20230502-062829.crash

[Maverick87Shaka]

With 128Gb of ram you probably don't need to worry about page file even for long duration mission.
From log file we can see:

2023-05-01 02:21:53.309 WARNING DX11BACKEND (10832): Can't map structured buffer.
2023-05-01 02:21:53.309 ERROR   DX11BACKEND (10832): Failed assert `0` at Projects\render\dx11backend_win8\Source\DX11StructuredBuffer.cpp:100
2023-05-01 02:21:53.311 ERROR   DX11BACKEND (10832): DX device removed. Reason: 0x887A0006
2023-05-01 02:21:53.312 WARNING WORLD (Main): ModelTimeQuantizer: ANTIFREEZE ENABLED
2023-05-01 02:21:53.312 ERROR   DX11BACKEND (16452): Failed assert `false && "failed to create texture"` at Projects\render\dx11backend_win8\Source\DX11TextureManager.cpp:711

That is typically caused by a driver crash, and normally the GPU is the first cause

 

3 minutes ago, Akula108 said:

With debug mode off, my 4090 clocked 2830 mhz.
With debug mode on, my 4090 clocked 2760 mhz.

As far I can see, in debug mode the GPU frequency is lowered, but not even match the NVIDIA standard frequency: 2235/2520Mhz
If you have some tools like afterburner, try to lower frequency to match the RTX 4090 baseline, and see what happens

[Akula108]

4 minutes ago, Maverick87Shaka said:

With 128Gb of ram you probably don't need to worry about page file even for long duration mission.
From log file we can see:

2023-05-01 02:21:53.309 WARNING DX11BACKEND (10832): Can't map structured buffer.
2023-05-01 02:21:53.309 ERROR   DX11BACKEND (10832): Failed assert `0` at Projects\render\dx11backend_win8\Source\DX11StructuredBuffer.cpp:100
2023-05-01 02:21:53.311 ERROR   DX11BACKEND (10832): DX device removed. Reason: 0x887A0006
2023-05-01 02:21:53.312 WARNING WORLD (Main): ModelTimeQuantizer: ANTIFREEZE ENABLED
2023-05-01 02:21:53.312 ERROR   DX11BACKEND (16452): Failed assert `false && "failed to create texture"` at Projects\render\dx11backend_win8\Source\DX11TextureManager.cpp:711

That is typically caused by a driver crash, and normally the GPU is the first cause

 

As far I can see, in debug mode the GPU frequency is lowered, but not even match the NVIDIA standard frequency: 2235/2520Mhz
If you have some tools like afterburner, try to lower frequency to match the RTX 4090 baseline, and see what happens

I will try this via afterburner but getting a result may take some time as I am not well versed in undervolting or underclocking GPUs.

[Maverick87Shaka]

8 minutes ago, Akula108 said:

I will try this via afterburner but getting a result may take some time as I am not well versed in undervolting or underclocking GPUs.

I've mentioned Afterburner just because you have an MSI GPU and maybe you have already installed since it's quite popular, but it's not my favorite one.
If you don't have anything at the moment, my personal choice will be NVIDIA Inspector, more simple and lightweight program.

[Akula108]

2 minutes ago, Maverick87Shaka said:

I've mentioned Afterburner just because you have an MSI GPU and maybe you have already installed since it's quite popular, but it's not my favorite one.
If you don't have anything at the moment, my personal choice will be NVIDIA Inspector, more simple and lightweight program.

Thank you very much for the suggestion, I am a little more versed around MSI Afterburner.
For now, I have turned down the power limit from 100 to 80 and will launch the game to look for a result.
I will update this thread ASAP when I have a result. 

[Akula108]

59 minutes ago, Maverick87Shaka said:

With 128Gb of ram you probably don't need to worry about page file even for long duration mission.
From log file we can see:

2023-05-01 02:21:53.309 WARNING DX11BACKEND (10832): Can't map structured buffer.
2023-05-01 02:21:53.309 ERROR   DX11BACKEND (10832): Failed assert `0` at Projects\render\dx11backend_win8\Source\DX11StructuredBuffer.cpp:100
2023-05-01 02:21:53.311 ERROR   DX11BACKEND (10832): DX device removed. Reason: 0x887A0006
2023-05-01 02:21:53.312 WARNING WORLD (Main): ModelTimeQuantizer: ANTIFREEZE ENABLED
2023-05-01 02:21:53.312 ERROR   DX11BACKEND (16452): Failed assert `false && "failed to create texture"` at Projects\render\dx11backend_win8\Source\DX11TextureManager.cpp:711

That is typically caused by a driver crash, and normally the GPU is the first cause

 

As far I can see, in debug mode the GPU frequency is lowered, but not even match the NVIDIA standard frequency: 2235/2520Mhz
If you have some tools like afterburner, try to lower frequency to match the RTX 4090 baseline, and see what happens

I reduced power limit to 75% which did result in lower clock speeds (wasn't able to make out the numbers) but unfortunately I have crashed again in MT Multiplayer after almost exactly 30 minutes of gameplay, latest crash file is attached below. 

dcs.20230502-083821.crash

[Flappie]

Have you tried a DCS repair in slow mode yet?

https://www.digitalcombatsimulator.com/en/support/faq/repair/#2463807

[Akula108]

Just now, Flappie said:

Have you tried a DCS repair in slow mode yet?

https://www.digitalcombatsimulator.com/en/support/faq/repair/#2463807

Thank you for the reply, is the slow mode the second option that scans the entire game? If so, I have tried that and the repair tool did not find any corrupt files that needed to be replaced, but the crashes persisted afterwards. I don't remember what the option was called off the top of my head so apologies for that. 

[Flappie]

Yes, that's the mode I was talking about.

Have you tried disabling MSI Afterburner before playing? For a few users, it causes DCS to crash.

[Akula108]

4 minutes ago, Flappie said:

Yes, that's the mode I was talking about.

Have you tried disabling MSI Afterburner before playing? For a few users, it causes DCS to crash.

I have just started re-using MSI afterburner to reduce the clock speeds of the card as the game was crashing without MSI afterburner running. I have actually just crashed out of another session on MT, this session lasted about 15 minutes. I had my 4090 running at 70% power limit and a clock speed of 2420mhz, the output of the crash dump is identical to last one I submitted in a previous comment. Are you able to decipher anything in the log? (new log attached below) I am very confused now why DCS in particular, is the only game on my system that behaves in this matter. 

dcs.20230502-090335.crash

[Akula108]

Another update, another crash in MT with the GPU at 2310mhz in game. I am beginning to think it is not a clock speed issue. Crash dump file output is the exact same as the previous two, have attached it below. I'm sure people are tired of me mentioning it but I am able to play more intensive titles such as Cyberpunk 2077 with ray-tracing with clock speed going above 2860mhz for hours with no issues. Please advise on how to proceed, below are my MSI afterburner settings. null

dcs.20230502-092849.crash

Edited May 2, 2023 by Akula108

[Flappie]

The message seen in your crash log, "RenderGraph::computeOptimalBarriers", is relatively new in DCS. I'm not entirely sure if it replaced the "good old" DXGI_ERROR_DEVICE_REMOVED message or not, but it may be a possibility.

Sorry, I don't have any more idea at this stage.

You should open a support ticket: log into DCS website, click Support, then Create a new ticket.

[Akula108]

11 hours ago, Flappie said:

The message seen in your crash log, "RenderGraph::computeOptimalBarriers", is relatively new in DCS. I'm not entirely sure if it replaced the "good old" DXGI_ERROR_DEVICE_REMOVED message or not, but it may be a possibility.

Sorry, I don't have any more idea at this stage.

You should open a support ticket: log into DCS website, click Support, then Create a new ticket.

Unfortunately I have submitted a ticket but it seemed like support was unable to get anywhere either so created this post afterwards. However, I was able to find another thread mentioning "RenderGraph::computeOptimalBarriers", supposedly it was fixed in a patch but it seems like users are still experiencing the issue, however the thread leads to a dead end as there are no suggested fixes.

The Crash log's output in that thread are basically identical to the output of my own crash log. I am still crashing on MT and ST after testing this morning. 

[Akula108]

Update to this thread, I may have what was causing the persistent crashes. First off, my windows power plan was set to "Balanced" though I do remember picking the "High Performance" option, a windows update may have reverted this change. Secondly, I uninstalled all additional terrains only leaving Caucasus and Marianas, One of these two changes has allowed me to play multiplayer on MT build for a 35 minute session (before being pulled away by work)without crashing. I will continue to monitor later this evening.

Edited May 2, 2023 by Akula108

[Sr.]

1 hour ago, Akula108 said:

Update to this thread, I may have what was causing the persistent crashes. First off, my windows power plan was set to "Balanced" though I do remember picking the "High Performance" option, a windows update may have reverted this change. Secondly, I uninstalled all additional terrains only leaving Caucasus and Marianas, One of these two changes has allowed me to play multiplayer on MT build for a 35 minute session (before being pulled away by work)without crashing. I will continue to monitor later this evening.

 

Now that you mention it... my similar problem seems to have started after the last Normandy 2.0 patch.

 

 

[Akula108]

Update: After a one hour session, the crashes have returned. However I have noticed that once the first crash happens, any time I relaunch and play the game, I can only stay in single player or multiplayer for around 10 minutes before another crash appears. Latest crash log attached, output is basically the same. "ComputeOptimalBarriers" still present. 

# -------------- 20230503-014114 --------------
DCS/2.8.4.39313 (x86_64; MT; Windows NT 10.0.22621)
D:\Eagle Dynamics\DCS World OpenBeta\bin-mt\GraphicsCore.dll
# C0000005 ACCESS_VIOLATION at 00007ffa173de168 00:00000000
SymInit: Symbol-SearchPath: 'D:\Eagle Dynamics\DCS World OpenBeta\bin-mt;', symOptions: 532, UserName: 'David'
OS-Version: 10.0.22621 () 0x100-0x1
0x000000000003e168 (GraphicsCore): render::RenderGraph::~RenderGraph + 0x2E48
0x0000000000041fd3 (GraphicsCore): render::RenderGraph::computeOptimalBarriers + 0x403
0x0000000000041ace (GraphicsCore): render::RenderGraph::bakeGraph + 0x18FE
0x000000000000e20f (SceneRenderer): DCSSceneRenderer::bakeRenderGraph + 0x3F
0x0000000000107f27 (Visualizer): smCamera_Implement::getClipRegion + 0x17837
0x00000000000483fb (edCore): ed::this_thread::yield + 0x1531B
0x00000000000476d2 (edCore): ed::this_thread::yield + 0x145F2
0x0000000000048f73 (edCore): ed::this_thread::yield + 0x15E93
0x00000000000326b1 (edCore): ed::thread::_get_current_thread_id + 0x71
0x0000000000029363 (ucrtbase): _recalloc + 0xA3
0x00000000000126ad (KERNEL32): BaseThreadInitThunk + 0x1D
 

dcs.20230503-014114.crash

[Flappie]

7 hours ago, Akula108 said:

Update: After a one hour session, the crashes have returned. However I have noticed that once the first crash happens, any time I relaunch and play the game, I can only stay in single player or multiplayer for around 10 minutes before another crash appears.

What do your PC temperature sensors say?

[Akula108]

20 minutes ago, Flappie said:

What do your PC temperature sensors say?

CPU and GPU both remain under 55C. 

[Akula108]

I have just experienced another crash in MP using the MT build, this time I crashed within minutes of joining the 4YA PVE servers. 
Crash log output was different this time. Text pasted below and log attached. 

# -------------- 20230503-093428 --------------
DCS/2.8.4.39313 (x86_64; MT; Windows NT 10.0.22621)
C:\Windows\SYSTEM32\d3d11.dll
# C0000005 ACCESS_VIOLATION at 00007fff304859b5 00:00000000
SymInit: Symbol-SearchPath: 'D:\Eagle Dynamics\DCS World OpenBeta\bin-mt;', symOptions: 532, UserName: 'David'
OS-Version: 10.0.22621 () 0x100-0x1
0x00000000001059b5 (d3d11): CreateDirect3D11SurfaceFromDXGISurface + 0x10535
0x0000000000042270 (dx11backend): createRenderer + 0x2A360
0x0000000000005422 (dxgui_edge_render): (function-name not available) + 0x0
0x00000000000054b5 (dxgui_edge_render): (function-name not available) + 0x0
0x0000000000019a68 (dxgui): gui::getWidgetCallbackType + 0x4D8
0x000000000001d09e (dxgui): gui::GUI::doRedraw + 0x2E
0x000000000011de90 (Visualizer): smSceneManager::regLua + 0x46D0
0x0000000000107121 (Visualizer): smCamera_Implement::getClipRegion + 0x16A31
0x00000000000f6996 (Visualizer): smCamera_Implement::getClipRegion + 0x62A6
0x000000000010a072 (Visualizer): smSceneManager::DestroySceneManager + 0x1542
0x00000000000326b1 (edCore): ed::thread::_get_current_thread_id + 0x71
0x0000000000029363 (ucrtbase): _recalloc + 0xA3
0x00000000000126ad (KERNEL32): BaseThreadInitThunk + 0x1D

dcs.20230503-093428.crash