Keeps crashing in Mariannas. I keep sending report. But nothing happens

[DmitriKozlowsky]

So when DCS crashes. Is it the module that crashes, or is it DCS? How do we know?

[Flappie]

Hi. Please attach the most recent ZIP file you can find in your "Saved Games/DCS.../Logs" folder.

[DmitriKozlowsky]

OK. For nth time.

dcs.log-20211008-104817.zip

Another one.

dcs.log-20211008-063729.zip

DCS: C:\WINDOWS\SYSTEM32\gpapi.dll, 133760 bytes, created 13:8:2021-13.13.54, written 13:8:2021-13.13.54, accessed 8:10:2021-10.48.19
2021-10-08 10:48:19.602 INFO    DCS: C:\WINDOWS\SYSTEM32\WINNSI.DLL, 35640 bytes, created 16:10:2020-5.7.32, written 16:10:2020-5.7.32, accessed 8:10:2021-10.48.19

 

What do these files do?

[MAXsenna]

5 hours ago, DmitriKozlowsky said:

So when DCS crashes. Is it the module that crashes, or is it DCS? How do we know?

 

You post the log here like this. I haven't seen you do it nine times though.

 

Or you can open and examine it yourself. The DCS.log file that is. Or this crash file.

You can even monitor the DCS.log file live with the correct tools, and watch when it crashes.

 

Anyway. Seems that your GPU crashed. Have you overclocked it? For some strange reason DCS doesn't like that.

Only game that crashes when I overclock.

Cheers!

 

 

 

 

[Hiob]

That should be common sense, to first undo any overclock (that includes "stock" overclock and XMP) of GPU, CPU and RAM. (step by step)

Quoting Jay (Jayztwocents): "An Overclock can only be regarded as stable, when ALL you applications and games run without issue."

 

When it still crashes, well at least you ruled out one possible cause, if not, you found the culprit

[DmitriKozlowsky]

No overclock. Not on CPU not on GPU. In fact except for Nvidia reference drivers, last being 472.12, my machine config has not changed at all. Windows Pro has updated itself. Razer Pro mouse drivers updated automatically. Multiple times. Never had a problem.

For past two years, on same machine, with same hardware. No problem. Then all of a sudden. Just started going to black.

My GPU is RTX 2060Super Founder's Edition is from Nvidia, not any 3rd party.

 

Is crash due to heat? Is the GPU slowly dying after two years.

[Hiob]

Probably not dying, but the thermal pads on the VRAM may degrade (due to dry out or bleed out of silicone oil), that could cause the GPU to become more sensitive to spikes in load.

I'm not saying, that this is the most probable cause, but it is a possible one.

What temps is your card while running dcs?

 

A way to test that would be to reduce the load on the card by lowering the clock speed with Afterburner. Have you enabled an xmp-profile for your memory? Just try to disable it.

 

Edited October 8, 2021 by Hiob

[DmitriKozlowsky]

98-99% 3D load with 76-77 C GPU temp. During flight. According to Task Manager.

[Hiob]

Have you tried to run it with an open case.

 

The whole purpose of this is to figure out if you have borderline stability issues with your GPU, that happen to manifest in DCS.

You could also try out an older version of the GeForce-Drivers. Oh and btw, can you right click on the nvidia task icon and check which applications run on your card?

 

Problem-hunting is a nerve wrecking and time consuming matter...

Edited October 8, 2021 by Hiob

[DmitriKozlowsky]

Well the issue is incipient. It just began perhaps a week ago, with older Nvidia drivers. For two years DCS was pretty rock stable.

[Flappie]

Please read my answer about the same issue here:

 

 

[DmitriKozlowsky]

I dont have any custom scripted mods. When machine is idle , GPU is temperature steady at 45 C.  A GPU renderer (Renderman 24.1 XPU with GPU only) IPR render GPU holds 46-49 C.

Running DCS and flying AV-8B in Marrianas, with no other 3D apps, GPU usage is pegged at 98-99% while GPU temp is steady at 76 C. I get random black screen crash even flying Instant Action/Nevada/Cold & Dark. Missions that never had any problem. Now, for about a week I am experiencing more and more frequent black screen crashes. Never at same point. Sometimes at end, sometimes at start, sometimes during takeoff, sometimes during INS alignment. As far as can tell A-10CII, Gazelle, and now AV-8B are effected. There is no difference in reliability when going back to earlier Nvidia drivers. There is NO overclocking of CPU, or GPU, or of memory. There is no warning. Crash freeze is instantaneous and goes to full black screen, while I can hear engine in the back. After 30 seconds, a DCS crash dialog appears. Seems about time that Mosquito came out, but I don't have that module, is when it started.

[Hiob]

Ok, that STRONGLY hints at some defect or becoming defect hardware.

Could be GPU, power supply, mainboard….

As a first aid measure, I would disassemble the whole pc, clean it, smell all components for burnt parts, assemble it with an open case…

Messy business. But random crashes all over the place, without any kind of overclocking? My bet is on faulty hardware (either somewhere in the power delivery, or due to heat).

[DmitriKozlowsky]

I opened PC, its all clean. No dust blocking air flow. Ran it with open box. which lowers temps by 1 or 2 degrees C, but had no effect. DCS has become unstable. No other software has problem with the machine. Not Maya, not Houdini, not Nuke, not Prepare 3D, not GPU renderers (Renderman XPU , Arnold GPU, and Redshift). Chrome runs fine. HP diagnostics report no issue. HeavyLoad ran fine. I suspect ED has borked up DCS since latest 2.7.X with Mosquito.

[Hiob]

Sorry, pal. Nothing else up my sleeve.

Have you done a clean re-install of DCS?

 

Don‘t get me wrong. I really don‘t want to suggest that DCS can‘t be faulty. The all-over-the-place-errors are just not very common.

Usually you experience somewhat  repeatable errors with a certain module/map/situation.

 

Hardware-Errors can be very well limited to certain software though. 

 

Edited October 9, 2021 by Hiob

[DmitriKozlowsky]

I had HP Performance Advisor running during one session of DCS runninf an Instant Action AV-8B Cold & Dark mission. After 10 minutes. Dark screen error. Here is report from AdvisorError.txt

 

HP Performance Advisor (version 2.1.4431.0)

An unexpected problem occurred at 7:26 PM on 10/9/2021.

    "UCEERR_RENDERTHREADFAILURE (Exception from HRESULT: 0x88980406)"

Diagnostic information will be written to the AdvisorError.txt file on the Windows desktop. If this problem persists, please send this file to ap-assist@hp.com.

Manufacturer: Hewlett-Packard
       Model: HP Z440 Workstation
        BIOS: BIOS - M60 v02.47
          OS: Windows 10 (64-bit) - Microsoft Windows 10 Pro for Workstations
    Graphics: NVIDIA GeForce RTX 2060 SUPER

An unhandled System.Runtime.InteropServices.COMException exception was caught at:
   at System.Windows.Media.Composition.DUCE.Channel.SyncFlush()
   at System.Windows.Interop.HwndTarget.UpdateWindowSettings(Boolean enableRenderTarget, Nullable`1 channelSet)
   at System.Windows.Interop.HwndTarget.UpdateWindowPos(IntPtr lParam)
   at System.Windows.Interop.HwndTarget.HandleMessage(WindowMessage msg, IntPtr wparam, IntPtr lparam)
   at System.Windows.Interop.HwndSource.HwndTargetFilterMessage(IntPtr hwnd, Int32 msg, IntPtr wParam, IntPtr lParam, Boolean& handled)
   at MS.Win32.HwndWrapper.WndProc(IntPtr hwnd, Int32 msg, IntPtr wParam, IntPtr lParam, Boolean& handled)
   at MS.Win32.HwndSubclass.DispatcherCallbackOperation(Object o)
   at System.Windows.Threading.ExceptionWrapper.InternalRealCall(Delegate callback, Object args, Int32 numArgs)
   at System.Windows.Threading.ExceptionWrapper.TryCatchWhen(Object source, Delegate callback, Object args, Int32 numArgs, Delegate catchHandler)

 

[DmitriKozlowsky]

SO I noticed that prior to freeze. DCS in game visuals display these random yellow'ish flares which l;ast for couple of frames, then I get these missing slivers of black, followed shortly by black DCVS screen and freeze. Temp of GPU is 76 and holding steady during incident.

[silverdevil]

i had similar problems long ago with another PC. I replaced what i thought was a faulty card. it turned out to be a mismatched DLL in the video drivers. maybe you already tried this. I suggest trying to clean install the nvidia drivers. maybe even installing an old card to bump the drivers and trying current card. my issue was only happening in DCS out of several other games i had.

[DmitriKozlowsky]

I re-installed Nvidia drivers, reinstalled DCS, OS updated. The problem persists. As I went through my modules, I find that all are effected. So problem is DCS core related. I get no spikes in heat, and no spikes in 3D GPU load and temp. Once flying, DCS ramps GPU to 98-99%, GPU memory to full 8gb, and temp is steady at 76-77 C .DCS runs fine for a random undetermined amount of time. Sometimes 1 minute, sometimes over 10, sometimes for hour or more, then crashes with black screen freeze. Started about time of last update , when Mosquito , got included.

[DmitriKozlowsky]

Seems that 78 deg C is the cutoff. At which point My 2060Super driver shuts down. However max operating temperature of RTX 2060Super is 89 C. Card does not even start ramp down and fan speed UP until 80 C.

[maximov]

After DCS crash, open the Event Log of Windows. View events -> Logs of the Windows -> System. Save the file (*.evtx) and attach here.

There is some likelihood that the video card gives an error due to the power supply of the computer. Over time, the power supply gives much less energy than the manufacturer declared.

[DmitriKozlowsky]

GPU fail has , sadly, been confirmed. Hardly unique.  Seems fail mode and symptoms of my RTX 2060SUPER FE are  quite common  also effecting 2070 and 2070Super. From 6 months to a year on average, without crypto mining . Mine 2060S lasted 2 years. Failpoint appears, in my case , to be VRAM on GPU. There is no fix. I have to RMA GPU to Nvidia. Kind of a weird process.  Nvidia delegated RMA on FE cards, in North America , to PNY.