Running ./sample_hello_world cudaErrorSystemDriverMismatch: system has unsupported display driver / cuda driver combination

[christopher.tuch]

Please provide the following info (tick the boxes after creating this topic):
Software Version
DRIVE OS 6.0.4 SDK
DRIVE OS 6.0.4 (rev. 1)
other

Target Operating System
Linux
QNX
other

Hardware Platform
DRIVE AGX Orin Developer Kit (940-63710-0010-D00)
DRIVE AGX Orin Developer Kit (940-63710-0010-C00)
DRIVE AGX Orin Developer Kit (not sure its number)
other

SDK Manager Version
1.8.3.10426
other (1.8.4-10431)

Host Machine Version
native Ubuntu Linux 20.04 Host installed with SDK Manager
native Ubuntu Linux 20.04 Host installed with DRIVE OS Docker Containers
native Ubuntu Linux 18.04 Host installed with DRIVE OS Docker Containers
other

I just installed DriveOS 6.0.4 rev.1 via SDK Manager and I tried running the ./sample_hello_world on my Linux host machine. I am getting the following output:

+++++++++++++++:/usr/local/driveworks/bin$ ./sample_hello_world 
*************************************************
Welcome to Driveworks SDK
[13-10-2022 16:26:06] Platform: Detected Generic x86 Platform
[13-10-2022 16:26:06] TimeSource: monotonic epoch time offset is 1665671077039348
[13-10-2022 16:26:06] Adding variable DW_Base:DW_Version
[13-10-2022 16:26:06] Addded variable DW_Base:DW_Version
[13-10-2022 16:26:07] Driveworks exception thrown: Platform: cannot acquire CUDA context. Error cudaErrorSystemDriverMismatch: system has unsupported display driver / cuda driver combination

Cannot init SDK

Any ideas?
Edit: to give more input
Before the installation I was checking my driver version with nvidia-smi, and I was using the nvidia-driver-515.
After the installation via sdkmanager-1.8.4-10431, i could not check my driver-version via nvidia-smi and in the additional drivers GUI menu, it said i would use a manually installed driver.

[VickNV]

Hi @christopher.tuch
Please try if rebooting your host system helps with this.
Also, don’t you run the program successfully in the other topic?

[christopher.tuch]

Yes i did run it successful, but i had an cuda unknown error, so i decided to try and reinstall it with SDK Manager all again. I just booted and it didn’t change, i guess it is something with the nvidia driver, as I described in the EDIT section in the first post. How can I check which driver is in use, when nvidia-smi is not working?

[christopher.tuch]

Ok, i got it working: The issue was due to an incomplete setup of my nvidia driver, as my system has secure boot on.
I had to set a key in the MOK Manager which I had to set after a driver installation. Now the sample_hello_world runs with the same output as in the linked thread ./sample_hello_world on host machine fails: Could not find ./recources.resources.pak
I still can’t run certain other samples due to Error CudaErrorUnknown: unknow Error. Should i make a new topic? As I can’t find a working fix on the internet.

[VickNV]

Thanks for sharing. also, saw your created another topic (Running precompiled Sample: "sample_camera_replay" results in Error: Error cudaErrorUnknown: unknown error).