2 minutes ago, bobcat999 said:

It will be great if they have fixed it as I love the NXi, but I uninstalled it due to CTD's as well. 

I completed an hour long flight through the Austrian Alps then - it was great!  Then after a touch and go, I had to pause the sim for a few minutes (escape key), and when I came back it had crashed to desktop. Strange how it would crash when paused, but there you go.  Getting better though!

Yeah it is a really good mod but caused a CTD fest on my end. And actually, the bug is not really the fault of WT but rather the engine itself. This bug could lead to instability in a lot of areas of the sim, not just with this mod, so now that WT is aware of it and understands what is causing it, hopefully Asobo can address this in the engine itself. 

4 minutes ago, RobJC said:

Yeah it is a really good mod but caused a CTD fest on my end. And actually, the bug is not really the fault of WT but rather the engine itself. This bug could lead to instability in a lot of areas of the sim, not just with this mod, so now that WT is aware of it and understands what is causing it, hopefully Asobo can address this in the engine itself. 

Yes, I am sure that Asobo will be glad of any assistance they can get on this at the moment.  We all know it isn't easy, and it must be just as frustrating for them.

More likely than not, the problem is on your end, as I don't see massive CTD complaints on WT's NXi Discord channel. 

3hr flight this morning and still no CTDs since removing this mod 👍

Edited August 10, 20214 yr by hanhamreds

Ok just caught up on the thread and can see that WT are on this👍

Found another source of CTDs. I took off from Denver to Dallas and 10 minutes in it crashed. I do have Flightbeam’s KDEN installed, so i removed it and tested again. No CTDs. Reproduced with other planes. This add-on has been rock solid prior to SU5. The only other 3rd party airport i have is Vancouver so i will check that next. 

One observation. This crash was the freeze then CTD variety. 

Edited August 11, 20214 yr by RobJC

Just to give a thorough update on the NXi and CTDs,

Instruments that are coded in Javascript can only cause a crash due to an underlying bug in the simulator that is triggered when the instrument calls the API that has the bug. Instruments in Javascript cannot, however cause a crash on their own, no matter how bad the error is in the Javascript itself. This always makes us scratch our heads a bit when we encounter people with repeatable crashes, because it means only one of two things:

- There's a bug in the simulator triggered when we make a call into the sim, which is not good

- There's a totally unrelated bug that has nothing at all to do with the instrument at all, but people (understandably) believe it is the source of their issue without sufficient isolation testing

Oftentimes the situation is much more the latter than the former (and it's awfully hard for a user to tell with so many moving parts), but in this case we've located a race condition in the sim during map initialization in the sim that is timing dependent. This is why it affects only some users. Additionally, this is also why it is affecting the NXi and not other instruments with maps, as the NXi loads faster than the stock instruments, making this timing tighter and making the race condition in the sim map system show its face.

We've identified a workaround we can use on the instrument end that seems to lessen or eliminate the issue. We won't quite know for sure how helpful it will be until we get it out in the wild in the next release, but it seems to eliminate the CTD for the users that have tested it.

These issues are especially sticky for us as developers because they both feel like they should be theoretically impossible as well as being terrifically difficult to debug, as we aren't able to reproduce it repeatably (or oftentimes at all).

That being said, we always do investigate these when a pattern this strong emerges from the noise.

-Matt

14 minutes ago, MattNischan said:

Just to give a thorough update on the NXi and CTDs,

Instruments that are coded in Javascript can only cause a crash due to an underlying bug in the simulator that is triggered when the instrument calls the API that has the bug. Instruments in Javascript cannot, however cause a crash on their own, no matter how bad the error is in the Javascript itself. This always makes us scratch our heads a bit when we encounter people with repeatable crashes, because it means only one of two things:

- There's a bug in the simulator triggered when we make a call into the sim, which is not good

- There's a totally unrelated bug that has nothing at all to do with the instrument at all, but people (understandably) believe it is the source of their issue without sufficient isolation testing

Oftentimes the situation is much more the latter than the former (and it's awfully hard for a user to tell with so many moving parts), but in this case we've located a race condition in the sim during map initialization in the sim that is timing dependent. This is why it affects only some users. Additionally, this is also why it is affecting the NXi and not other instruments with maps, as the NXi loads faster than the stock instruments, making this timing tighter and making the race condition in the sim map system show its face.

We've identified a workaround we can use on the instrument end that seems to lessen or eliminate the issue. We won't quite know for sure how helpful it will be until we get it out in the wild in the next release, but it seems to eliminate the CTD for the users that have tested it.

These issues are especially sticky for us as developers because they both feel like they should be theoretically impossible as well as being terrifically difficult to debug, as we aren't able to reproduce it repeatably (or oftentimes at all).

That being said, we always do investigate these when a pattern this strong emerges from the noise.

-Matt

Thanks for the info Matt. This was a strange one but glad you guys sorted it out. 
 

 

1 hour ago, MattNischan said:

we've located a race condition in the sim during map initialization in the sim that is timing dependent. This is why it affects only some users. Additionally, this is also why it is affecting the NXi and not other instruments with maps, as the NXi loads faster than the stock instruments, making this timing tighter and making the race condition in the sim map system show its face.

This confirms my guess about concurrency issue. Glad you've found the source.

I have a fast system I would say... Ryzen 5950X, FS on NVME... things load faster

On 8/9/2021 at 6:24 PM, hanhamreds said:

Just wanted to come back with an update.

since removing the G1000 nxi ive had about a dozen flights today, ranging from 1hr 34mins down to a couple of minutes - I’ve had no CTDs!

You removed it from the Content Manager, or there's a different procedure for the uninstall?

3 hours ago, edpatino said:

You removed it from the Content Manager, or there's a different procedure for the uninstall?

Yep content manager. This is an old post. I’m now running the update with no issues.

9 hours ago, hanhamreds said:

Yep content manager. This is an old post. I’m now running the update with no issues.

Ok, thanks. I know it’s an old post but it look to me that there are still issues after the most recent update 

I may be late, forgive me...
did you manage to solve it? me getting inexplicable CTD's
MSFS just closes, without errors.. I'm even afraid of messing with the G1000NXi all the time, and flying around with that feeling is really unpleasant.