Recurring CTD during missions, unclear what else I can do to isolate a cause.

[silverdevil]

1 minute ago, Istari6 said:

Thanks for the additional ideas!  I've run a full 3DMark Steel Nova stress test on my entire system, passed 99.0%, no issues identified.  I've also done a complete MemTest86+ check on my RAM modules, which passed.  So i don't think it's the factory-set overclocking or RAM.  

On your point about hardware, I did some more testing last night, and I was able to run an entire mission without any CTDs in "flat screen" (2D) mode.  This is a mission that is particularly unstable for CTDs, almost always crashes in the first 5-10 minutes.  So I'm now thinking it has something to do with the chain of layers to my Varjo Aero headset.

I generated more CTDs last night, and plugged the Windows Event Log data into ChatGPT (o3-mini-high) for suggestions.  Remember, I'm not getting any error message from DCS and no crash log generated - it just crashes straight to Windows.  The DCS log just terminates, no "minidump" or other information.  So I'm left trying to use the Windows Event View logs to find out why it's crashing.

Interestingly, ChatGPT says the following:

What the Logs Tell Us

1. Faulting Module & Exception Code

   • Faulting Module: edCore.dll (a core part of DCS’s engine)
   • Exception Code: 0xc00000fd
     • This code typically means stack overflow, which generally indicates that a function (or a chain of functions) is calling itself repeatedly until the stack runs out of space.
   • Fault Offset: 0x0000000000182e37
     • This is the location within edCore.dll where the error occurs. Without symbol files (debug information) from Eagle Dynamics, it’s hard to pinpoint the exact function. However, the fact that it’s within a core engine DLL suggests the problem is internal to DCS when executing a particular code path.

2. VR-Only Occurrence

   • You mentioned that DCS runs fine in “flat screen” (2D) mode and only crashes in VR mode.
   • This strongly suggests that the crash (i.e., the stack overflow) is triggered by the VR code path inside DCS. In many cases, the VR rendering or related logic (possibly interfacing with the OpenXR runtime or Varjo-specific routines) may be executing a recursive or improperly bounded call sequence.

 

good work. so it seems the issue is VR. questions. is there a way to test the headset on another game?

here is where temp folder is.

%LOCALAPPDATA%\TEMP\DCS

just paste that into URL in file explorer

[MAXsenna]

@Istari6 Excellent! Happy you solved it. If I were you, I'd do that stress test again to see if it passes this time. Then you'll "know" it was the reseating of the memory sticks. I've seen other users here reporting the same. I also believe they were VR users.
Cheers!

Sent from my SM-A536B using Tapatalk

[Istari6]

On 2/8/2025 at 12:47 PM, MAXsenna said:

@Istari6 Excellent! Happy you solved it. If I were you, I'd do that stress test again to see if it passes this time. Then you'll "know" it was the reseating of the memory sticks. I've seen other users here reporting the same. I also believe they were VR users.
Cheers!

Just to be clear, I haven't solved it yet :(.  I realized that the formatting in my original message was unclear, I was cutting-and-pasting from another thread where they had fixed a similar issue.  I'm diving back into it tonight, hoping to finally pin this down.  I'm "jonesing" here without my DCS fix.

[Flappie]

@Istari6 I've just found this solved thread for another user with the same issue. You both have a 14900 CPU. Can you try what is suggested?

 

[BIGNEWY]

threads merged

[Istari6]

OK I'm back after several days of troubleshooting.  First, I'm still unable to play DCS in Varjo VR for more than 5-10 minutes with a CTD.  It's immensely frustrating.  2D mode seems to work OK.  I've flown several missions with friends in 2D mode without a CTD.  

The reason for the delay in following up is that it's taken a few days to execute the following steps:

• Reset the entire PC to Corsair's original factory settings.  All previous data wiped.  Confirmed a clean SSD, reinstalled everything from scratch including 320 GB of DCS, latest drivers, etc. --> still get CTD, even with this fresh system
• Rolled back NVIDIA driver from latest 572.16 to 566.14 (September 26th) -> CTD
• Rolled back Varjo Base from latest (4.6.1) to 4.4.0 -> CTD
• 3DMark Steel Nova Stress Test - 99.0% (earned Steam achievement "Rock Solid"), all CPU temperatures below 80C except for a single momentary spike to 99.5C across 20 minutes of stress testing.  
• RAM DDR 5 sticks - swapped, checked voltages in BIOS, confirmed 1.35V to both -> CTD

Having done a complete reinstall of my entire system from scratch, rolled back both NVIDIA drivers and Varjo Base/OpenXR to earlier configurations (where DCS worked), and done a 20 minute stress test with complete stability, where to go from here?

The only information I have are the Event Viewer logs from Windows.  There is nothing in the DCS.log files.  

What I get recurrently are errors like the following:
 

Quote

 

Event 1000, Application Error - 2/15/2025 - 5:32:22pm

 

Faulting application name: DCS.exe, version: 2.9.12.5336, time stamp: 0x678ab24b
Faulting module name: edCore.dll, version: 2.9.12.5336, time stamp: 0x6786a3c4
Exception code: 0xc00000fd
Fault offset: 0x0000000000182e37
Faulting process id: 0xBDC
Faulting application start time: 0x1DB8011185663D7
Faulting application path: C:\Program Files\Eagle Dynamics\DCS World\bin-mt\DCS.exe
Faulting module path: C:\Program Files\Eagle Dynamics\DCS World\bin-mt\edCore.dll
Report Id: a2ac1f36-94b3-4110-b81a-98e8d75d3bcf
Faulting package full name: 
Faulting package-relative application ID: 

 

---------------------------------

I do see Flappie's recommendation about a potential solution in the other thread.  Thanks for the pointer!  See comments below

On 2/11/2025 at 1:11 PM, Flappie said:

I've just found this solved thread for another user with the same issue. You both have a 14900 CPU. Can you try what is suggested?

@Flappie: I've reviewed this thread, and the only actionable step seems to be at the end where the author says: "

Quote

 

So after much grinding and gnashing of teeth, a friend of mine was able to look into this for me and provided me what I needed to ultimately fix the problem.

Enforce a 253 Watt power limit and disable multi-core enhancement. He did some searching around and said that some of the 14900 series is susceptible to a low I2 voltage whci can be tweaked at some other point. 

 

I am lost on what this means.  What is a low I2 voltage?

Note, regarding some of the other ideas in that thread, they sound reasonable, but I don't think they apply here.  I have a "stock" Corsair Vengeance i750 system, purchased 11/2023.  1000 W power supply, nothing OC that I'm aware of.  I've not done any changes to the system from the stock Corsair configuration.  Stress test shows CPU and GPU temperatures look good.  

How would I identify if I have something malfunctioning in an individual core during MT operations?  

 

Edited February 16 by Istari6

[SharpeXB]

@Istari6 Are you getting continual CTDs when using a 14th Gen Intel CPU? This is happening to me and it might be likely due to the CPU degrading prior to the microcode fix last year. I’m looking to just replace it. Are other games crashing too? Have you tried updating your BIOS? 

[Flappie]

8 hours ago, Istari6 said:

I am lost on what this means.  What is a low I2 voltage?

Here's the definition of PL1 and PL2 (sometimes they are simply called L1 and L2):

• PL1 = Power Limit for long-term steady power consumption. Intel recommends for PL1 to be set in consideration with limits of thermal cooling solution.
• PL2 = Power Limit for short-term maximum power consumption under heavy load. It means that PL2 shouldn't be set lower than PL1.

 

8 hours ago, Istari6 said:

I've reviewed this thread, and the only actionable step seems to be at the end where the author says:

"So after much grinding and gnashing of teeth, a friend of mine was able to look into this for me and provided me what I needed to ultimately fix the problem. Enforce a 253 Watt power limit and disable multi-core enhancement. He did some searching around and said that some of the 14900 series is susceptible to a low I2 voltage which can be tweaked at some other point. "

It appears that the Intel 14900K series has stability issues, in games only, because compatible motherboards use non-recommended default settings in their "Auto" mode. 

Following this finding, Intel has asked these motherboard manufacturers to use lower default settings by May 31, 2024. In this linked page, you can see that Intel now recommends a maximum of 253 W for both PL1 and PL2.

So it does not matter if your PSU is able to give 1000 W or more to your PC. As long as you have a motherboard with a 14900K in your system, your system may be affected - unless of course your motherboard has Intel recommended settings.

 

The next steps for you are to:

• Confirm that you have a MSI PRO Z790 motherboard (it's what the Corsair Vengeance i7500 system is supposed to be based on),
• Check your motherboard PL1 and PL2 settings in the BIOS (Overclocking > Advanced CPU Configuration). If you have an MSI Z790, they will be called:
  • Long Duration Power Limit
  • Short Duration Power Limit

If at least one of these two values is more that 253 W, it probably explains your system instability in games. And here's the procedure for the MSI Z790:

1. Go to your system BIOS.
2. Save your current settings into a profile so you can return to them later if you want.
3. Reset your BIOS to default settings.
4. Enable XMP.
5. Go to Overclocking > Advanced CPU Configuration.
6. Set Long Duration Power Limit(W) = 253
7. Set Short Duration Power Limit(W) = 253
8. Set CPU Current Limit(A) = 307
9. Save and exit.
10. Test DCS.

Of course, lowering your power limit settings means that performance will decrease a bit in games, but it's a choice you must make between "more frames per second" and stability.

Edited February 16 by Flappie

[SharpeXB]

I am definitely not an expert at any of this so I’ll refrain from giving out any actual advice and just relate my own experience with the a 14900KS. All the tech support I received was through my PC vendor.

My first 14900KS was purchased in April of 2024 and by July has started to exhibit repeated CTDs. I was able to solve this problem by updating the BIOS which likely included those fixes above. By the end of July just a few weeks later, I became aware of the widespread degradation issue with the 13 and 14 series chips. By mid-August a “microcode” fix was put out for my MB to halt the degradation problem. Afaik it did things like enforce the Intel defaults and disable the Multi-Core Enhancement feature of the MB. If going in and tinkering with the BIOS seems intimidating, updating the whole thing is rather easy. Some MB utilities can do this from Windows.  

Now the 14900KS is repeatedly crashing again and a BIOS update doesn’t solve this. From looking that that timeline above it seems pretty clear that the chip could have been damaged between April and July. I can only guess how these things behave when damaged varies a lot. It still shows very good (even better) performance on benchmarks and the games run very well until just quitting. My previous 13900K did exhibit noticeable performance drops but never crashed.

At this point I just plan to replace the 14900KS with another one as the microcode fix supposedly resolved the degradation issue. 

Edited February 16 by SharpeXB

[Istari6]

On 2/15/2025 at 8:40 PM, SharpeXB said:

@Istari6 Are you getting continual CTDs when using a 14th Gen Intel CPU? This is happening to me and it might be likely due to the CPU degrading prior to the microcode fix last year. I’m looking to just replace it. Are other games crashing too? Have you tried updating your BIOS? 

Yup, it's a 14th Gen Intel CPU: Intel Core i9-14900K.  No other games are having the same problem I'm having in DCS.  I was hopeful that updating the BIOS would solve the problem (perhaps due to the microcode fix), but I just updated the BIOS from the Corsair Support site, and boom.  DCS crashes again.  So updating the BIOS hasn't solved the problem.  

On 2/16/2025 at 3:10 AM, Flappie said:

Here's the definition of PL1 and PL2 (sometimes they are simply called L1 and L2):

• PL1 = Power Limit for long-term steady power consumption. Intel recommends for PL1 to be set in consideration with limits of thermal cooling solution.
• PL2 = Power Limit for short-term maximum power consumption under heavy load. It means that PL2 shouldn't be set lower than PL1.

OK thanks @Flappie.  Updating the BIOS hasn't worked, I'll next dive into the BIOS and see if I can adjust the settings around 253W for PL1/PL2 as you're suggesting.  More once I try that... 

On 2/16/2025 at 8:34 AM, SharpeXB said:

Now the 14900KS is repeatedly crashing again and a BIOS update doesn’t solve this. From looking that that timeline above it seems pretty clear that the chip could have been damaged between April and July. I can only guess how these things behave when damaged varies a lot. It still shows very good (even better) performance on benchmarks and the games run very well until just quitting. My previous 13900K did exhibit noticeable performance drops but never crashed.

@SharpeXB - thanks for the background on your issues with the 14xxxx chips.  I'm starting to wonder if my CPU was similarly damaged, given I had solid performance from this machine for almost 12 months, now it's a continuous problem that i can't seem to fix.  I've just updated the BIOS, hasn't helped.  Crashed again within 5-10 minute of mission start.  Do you know if there's any way to verify a damaged CPU?  I've run Prime95 and several other benchmarking tests that stress the CPU, and they haven't shown anything.  My Corsair machine has a 2-year warranty, wondering if I can prove the CPU is damaged if I can get a new chip under warranty.  Not that you'd know how to navigate Corsair's warranty, but I likely need to demonstrate that the chip is damaged to get a new one sent out.  

[SharpeXB]

On 2/19/2025 at 3:15 PM, Istari6 said:

Do you know if there's any way to verify a damaged CPU?  I've run Prime95 and several other benchmarking tests that stress the CPU, and they haven't shown anything.  My Corsair machine has a 2-year warranty, wondering if I can prove the CPU is damaged if I can get a new chip under warranty. 

I’m in the same boat, just going to buy a new 14900KS and get it replaced. From what I see they’re still reliable as long as they weren’t running without the BIOS fix. It’s worth a shot anyways IMO since I don’t want to replace my MB or CPU cooler at this stage. As far as I know the only way to verify a damaged chip is to take it out and test it, like send it to Intel or something. If you got it as part of the prebuilt machine you probably need to go through their warranty. I had actually bought this one myself so I can RMA it. But yeah DCS started crashing on me first, followed by other games. 

[Istari6]

So an update two months later...

I've been using a Quest 3 headset to workaround the CTD issue.  It's much inferior to the Varjo Aero, but at least I could continue to fly with my friends.  I was reluctant to buy a new CPU until I could prove it was malfunctioning, but none of the tests were showing anything.  The news of the Trump tariffs forced me "off the fence", and I bought a new i9 14900K before prices rose.  It was a bit of gamble, but I really miss flying DCS with high-resolution and a stable framerate.

The good news is that it worked.  I've now flown multiple hours in DCS without any CTD, including several missions that were reliable triggers in the past.  I think that SharpeXB and others on this thread were right, and my original i9 was damaged due to incorrect BIOS settings.  Thanks for pointing me in the right direction, it's great to be flying DCS "properly" again.  

[Flappie]

Thanks for your feedback.