CTD continues with 1.0.1

[mtbparker]

Hey Folks,

The 1.0.1 update has a lot of great improvements.  I just want to report in, however, and let you know that I still get CTDs very regularly.

The error message in the log reads: 

Quote

0:24:30.994 E/GFX/VK: Encountered Vulkan error VK_ERROR_DEVICE_LOST. Expression: result. C:/jenkins/design-triggered/source_code/app/X-Plane-f/../../core/gfx/Vulkan/gfx_vk_query.cpp:98

Thanks again for the effort.  This is a beautiful plane.
 

I'm running XPlane b17.  EVGA 1080Ti graphics card on an Intel-5930K and 16GB RAM

Tom

Edited August 12, 2020 by mtbparker

[Cameron]

That appears to be a Laminar error. You should report that to them!

[mtbparker]

I'm not so sure about that.  Yes, the error is a very generic error indicating that the graphics card is crashing and not recovering.  (See link below for a good explanation.) 

The problem is that it only occurs with this aircraft.  There is a corresponding secondary symptom that also immediately precedes the crash.  The audio will "buzz".  In my experience as a SW developer whose worked on D2A conversion of packetized audio, this happens when the stream of audio packets stops flowing to the A2D converter.  Since 1.0.1, I hear that audio buzzing now multiple times in a single flight.  Approximately half the time, that buzzing is immediately followed by the CTD with the error mentioned above.  The other half of the time it seems to recover.  Never do I have a CTD without that error.  And, again, it only happens with this aircraft.

https://forums.x-plane.org/index.php?/forums/topic/209672-ctd-vulkan-device-loss-error/&tab=comments#comment-1896351
 

Yes, I'm open to the possibility that it's outside the scope of this aircraft.  If I hear the buzz and/or see the CTD anywhere else, I'll be sure to update here so that I don't waste precious time chasing the wrong problem.

Edited August 12, 2020 by mtbparker

[Cameron]

It may only occur with this aircraft simply for the fact we reach into the areas of the sim some others don't, therefore exposing a bug in Laminar's code.

Having read Sid's reply, he's really not pointing fingers at third party devs there. Jenkins is their compile platform, and only they will have access to what's going on there. Sounds are also 100% controlled by Laminar code via FMOD integration.

Sid flat out gives a link to the bug reporter at the end of his post, and seems a lot of people have resolved the error by keeping the factory overclock off on the GPU.

[mtbparker]

Okay. Thanks for the help.