Never ending CTDs which seem to be related to Nvidia...?

[WebMaximus]

Thank you using the Bug section, using templates provided will greatly help the team reproducing the issue and ease the process of fixing it.

Are you using Developer Mode or made changes in it?

No

Brief description of the issue:

Random CTDs when loading a flight

Provide Screenshot(s)/video(s) of the issue encountered:

Detailed steps to reproduce the issue encountered:

Unfortunately not reproduceable since it’s random

PC specs and/or peripheral set up if relevant:

Intel Core-i7 8086K, Nvidia 3090

Build Version # when you first started experiencing this issue:

Had this since way back, can’t recall when I first saw it and have it now in the latest version

---

For anyone who wants to contribute on this issue, Click on the button below to use this template:

Do you have the same issue if you follow the OP’s steps to reproduce it?

Provide extra information to complete the original description of the issue:

If relevant, provide additional screenshots/video:

I realize how hard it is to help someone having random CTDs but what I really wanted to ask about in here and/or highlight is the possible relation to Nvidia. Reason is how I always see the same message in the Windows Event Viewer System log go hand-in-hand with the MSFS CTD message in the Windows Event Viewer Application log. Both these messages are always seen together at the very same second as seen below.

This is what’s seen in the application log:

This is what’s seen in the system log the very same second, always the same thing:

I should also mention that I’ve seen this exact same issue over and long time and many different Nvidia drivers. So it’s not related to a specific Nvidia driver version. Also, I’ve tried using DDU to fully delete any traces of old drivers before installing a new Nvidia driver version.

Last thing I want to mention is how I was quite hopeful when I found the below tip. Unfortunately, following the advice given didn’t fix the problem for me.

https://substance3d.adobe.com/documentation/spdoc/gpu-drivers-crash-with-long-computations-tdr-crash-128745489.html

I hope I one day will be able to figure out why this is happening and fix it. Because it does take away a big part of the fun with MSFS when you constantly need to worry about it crashing to the desktop.

[LongHaulSky]

Moved to #self-service:ctd-help which is more appropriate

[XR219]

Have you tried uninstalling the Nvidia Phsyx driver and restarting the PC?

[SteveOhhNo]

next option would be , Limit MSFS FPS in the Nvidia driver. default is whatever as long as it’s a high number , just limit it to a number you concider ‘normal’

[WeirdNeville]

Is ‘hardware accelerated gpu scheduling’ off?

This is one suspected cause of crashes and is suggested by zendesk as a potential fix.

You could also try reinstalling an older version of the drivers, that you knew to be stable, as a potential fix.

Double check that it is always that fault logged: my crashes very occasionally show a d3d.dll or nvidia driver crash, but that’s not actually the root cause -the sim crashes and causes a driver crash, not the other way around.

Good luck and happy hunting!

[OldpondGL]

My thinking is that the GPU stops responding because it’s waiting for the sim to supply it some texture, and the sim fails to read that memory location because it didn’t come across the network. A good test would be to turn OFF online functionality for a week to see if the CTD’s stop.

[WebMaximus]

As mentioned, I’ve had this issue for a long time and with various versions of the Nvidia drivers and after using DDU for a guaranteed clean install.

Haven’t tried not installing the PhysX part of the driver though. Are you suggesting that one being the culprit?

[tamsini7744]

I strongly believe that CTDs are an unwanted feature of flight sims going back to the old MSFS, FSX, P3D and today’s MSFS (cannot speak for any other flight sim).

I also believe that, try as you might, no one can pinpoint exactly why one gets a CTD or where the problem stems from.

You can fly one route with a specific aircraft and get a CTD 30 minutes in, then restart your system and sim, try the exact same flight and succeed. Why? Impossible to tell.

Example, I have had CTDs in the last two weeks by:

• clicking a “secret spot” on the Milviz 310
• clicking HDG mode on the Flysimware 414
• clicking on the APU Bleed on the PMDG 737

Is it the developer’s coding? Highly doubt that they are the cause and also highly doubt they could possibly understand why it would happen.

Yes, you can lower your chances of getting CTDs by adopting suggestions made by the more knowledgeable folks on forums as well as developers but for some of us, they will never disappear.

Until the OS can properly log the exact cause of a particular CTD, I’m pretty sure they won’t completely disappear. But it will not stop me from enjoying this bizarre pasttime of ours

[XR219]

I haven’t used the latest Nvidia drivers since November when almost everyone had to roll it back to stop the sim crashing (Nvidia bug).

I had been having these intermittent CTDs over the last few month though.

First thing I did was to edit the power plan in the OS. Not just selecting high performance, but edited it so that the USB ports couldn’t go into power saving states.
This had an immediately noticeable affect, and I thought I had fixed it.

But it returned, although much less frequently.

Next was OneDrive. Pausing it before i started taxiing for the duration of the flight seemed to sort it. I had weeks with no problems. So much so that I thought I’d give the Nvidia drivers a try on Thursday. I didn’t install Geforce Experience.

First flight, 10 minutes in CTD.

Uninstalled the Physx driver, restarted, and so far (in combination with the other two points above) no CTD. I’m still convinced something will happen again with the Nvidia driver, but if it does, I’ll go back to using the one installed through Windows Update which seems infinitely more stable.

[StatesideSiren4]

If you’re using GeForce Experience, try rolling back to an older version of that. You can find the list of old versions on Nvidia’s website and if you Google a particular version number, there are plenty of download sites to snag it from. I had one particular version of GFE that just killed MSFS (I use it to record flights). Once I rolled it back, all good. Then one day they updated GFE to a new version, I got that, and never had any more problems.

[XR219]

A good way to rule out Nvidia is to do a complete uninstall of all of the Nvidia components. Then restart the OS and Windows will automatically install just the driver in stable form from the Windows Update library, without all the Nvidia bloat.

[N6722C]

maybe the Nvidea drive stopped responding, because it was MSFS that crashed ? (Not the other way around)

Your crash example clearly shows it was cause by a forbidden attempt to access memory location Zero, ie a Null Pointer Error.

The only real question, is if that a Null Pointer error is in the Asobo Code, or is it in any of the library (dlls) that MSFS is using. be it MSFS dlls, or Nvidea Driver dlls.

The fact that this is NOT appearing in other applications that rely on Nvidea Drivers, would logically tend to point to it being in MSFS itself, and not the Nvidea code.

---

I have looked at multiple Crash dumps, mostly null pointer type crashed, but without a .pdb file, its difficult top pinpoint exactly what & where then fault is coming from.

(I am use to having that .pdb file when I debug my own code !!)

But looking at the crash dump disassembly info from a Crash Dump analyzer, one can clearly see case when a Null Pointer is not being tested for, or, it is, but the pointer is in fact -1, and that is not being tested for, and hence the resulting CTD.

GUT feeling – may be cause by some of the extra security code, merged into the release version, to prevent tracing and debugging – IF SO, how ironic is that !!!

It is this difficult for a user, looking at their own crash dump of Release Compiled code, to tell if the issue is in Asobo’s written Code, or in one of the Libraries that MSFS is using, (ie Coherent)

Maybe it is possible, but it is beyond my knowledge or incentive, to figure it out, without the .pdb file .

Not that it should be for me to figure this stuff out or need to understand it, as the USER - that task if for the Developer who received my money for the product, expecting it to be functional and “Fit for purpose”.

[V8SHARK]

I am having the same experience as you with CTD and that error message in Event Viewer.

I have no solution.

I rebuilt my system yesterday and it seems to be OK, so far.

[WebMaximus]

Many thanks for all the great input here guys!

These are the things I’ll try for starters and I’ll try one thing at a time. To make sure I know which of these things (hopefully) fixed the issue in the end. Always easy when having an issue to try all sorts of things at once. In a desperate move to get rid of the issue. This time however and based on for how long I’ve had this issue, I really want to understand what the actual culprit is.

1. Disable Hardware Accelerated GPU Scheduling
2. Get rid of PhysX
3. Use the Nvidia driver that comes with Windows Update

I hope one of these will address the problem. If not, I would be willing to invest time and effort into getting a better understanding of what exactly is happening. By using some form of CTD analyzer tool or whatever is required to see what is going on behind the scenes when I have these CTDs. As mentioned above, I like to think it shouldn’t be that hard to track this one down. Since it always shows the exact same thing in the Windows Event Viewer logs.

As someone suggested, you can of course get CTDs for lots of different reasons and in different situations. In my case though, I want to repeat I only have these CTDs by the time I start loading a flight after having picked my aircraft and my origin airport. I never get any CTDs (fingers crossed) these days in-flight.

Happy to read how one of you got rid of the issue. At least for the time being…by rebuilding the complete system. This “solution” however is one I’m quite allergic to. I know how I’ve been in contact with Asobo support in the past when I wasn’t able to receive any updates for MSFS. Where things eventually turned into that direction. Being asked to rebuild my complete system.

Of course there might be situations when such a measure is the only way in the end. However, having to do that kind of thing to get a single application running without crashing to the desktop, be it MSFS or any other application, that to me simply isn’t acceptable. To me, that one always feel like support staff choosing the easy way out. Or at least hoping it will be. Not for the guy having the issue but for them.

Let’s see what happens. Will report back and thanks again for the input and suggestions so far!

[WebMaximus]

So…back after having disabled Hardware Accelerated GPU Scheduling and fingers crossed…not a single crash so far.

Need to do more testing but it does look promising and I hope it will stay like this. If so, I’ll come back here and mark this issue as resolved.

Thanks @WeirdNeville for the suggestion!

[Seven7Tango]

I’ve had this exact same issue with an AMD card: RX 5700XT with all manner of different drivers since August 2020. I could go months without a single CTD event, and then go through periods when I would get a half-dozen CTDs in a single day.

About a month ago I built a new rig using an RTX 3080 10GB card. I have not had a single CTD event since. I hope the good luck with the new card continues.

FWIW: I have always used Nvidia GPUs but decided - based on reviews and comparisons - that I’d give “Team Red” a chance to impress me. The conclusion I have drawn is that the Radeon card runs hotter, even though it draws less power. (That’s a head scratcher to me.) Performance is good, but the 3080 is better… as you might suspect. I’ll stick with Nvidia cards for future builds unless there is overwhelming proof that a Radeon card is superior.

My $0.02… good luck with chasing the root cause of the CTD events down.

[WebMaximus]

Thanks but with some luck, it looks like the problem might be solved (see my post above yours).

Good to hear things are working good for you with the new 3080

[Salem978]

I’ve since forever given anyone but team blue a chance to the point I bought a Kyrocera video card once … once. It had never worked out in my favor.

[PacificSet90456]

I’ve been having this nvlddmkm error since day one with all the NVIDIA drivers and MSFS updates. This error occurs on my system with non-MSFS graphic intensive applications as well. This error has been discussed in the NVIDIA forums FOR YEARS with many different NVIDIA cards, PC configurations, and Windows versions/updates/fixes without any resolution.

Basically, there is a Windows timer that makes sure the GPU responds within a certain time. If there isn’t a GPU response by timer expiration, then Windows assumes the GPU or the GPU driver has failed and Windows attempts recovery. Windows kills any applications that are using the GPU to attempt recovery. This is when the CTD happens (Windows kills MSFS) and the nvlddmkm error is logged. Rolling back drivers, Windows updates, changing GPU cards, or using DDU does not permanently fix this problem.

There might be certain NVIDIA settings that cause this error to happen randomly but so far there isn’t any way to force this error to happen.

[WebMaximus]

Did you already try to disable ‘Hardware-accelerated GPU scheduling’ as previously suggested in this thread?

As mentioned, in my case it does seem to have made things better. Still haven’t had a CTD since I disabled that one and I’m still keeping my fingers crossed it will carry on that way…