GPU not detected in NVIDIA DRIVE AGX Orin DevKit after flashing Drive OS 6.0.10

[aastudil]

Please provide the following info (tick the boxes after creating this topic):
Software Version
DRIVE OS 6.0.10.0
DRIVE OS 6.0.8.1
DRIVE OS 6.0.6
DRIVE OS 6.0.5
DRIVE OS 6.0.4 (rev. 1)
DRIVE OS 6.0.4 SDK
other

Target Operating System
Linux
QNX
other

Hardware Platform
DRIVE AGX Orin Developer Kit (940-63710-0010-300)
DRIVE AGX Orin Developer Kit (940-63710-0010-200)
DRIVE AGX Orin Developer Kit (940-63710-0010-100)
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
2.1.0
other → docker ngc.nvidia

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 → native Ubuntu Linux 22.04 Host installed with DRIVE OS Docker Containers - NGC

Issue Description
Hello all,
After flashing my AGX Orin DevKit with Drive OS 6.0.10 using the provided Docker SDK (NVIDIA NGC and Set Up DRIVE OS Linux with NVIDIA GPU Cloud (NGC) | NVIDIA Docs), the GPU is not detected:

1. jtop shows “GPU: Not Available”.
2. nvidia-smi returns errors

Steps to Reproduce

1. Flashed the device using the Docker SDK image (Drive OS 6.0.10) from NGC.
2. Booted the system and verified the OS runs correctly.
3. Installed jtop (sudo pip install jetson-stats) and nvidia-smi (preinstalled).
4. Both tools fail to detect the GPU.

Expected Behavior

• nvidia-smi should display GPU stats (e.g., utilization, memory).
• jtop should show GPU usage (as on Jetson platforms).

Additional Context

• Goal: Run a YOLOv8 TensorRT model (Ultralytics guide), but GPU acceleration is unavailable.
• Questions:
  1. Is there a missing driver or configuration step for Drive OS 6.0.10?
  2. Could you provide a step-by-step guide to:
  • Fix GPU detection.
  • Deploy YOLO models with TensorRT on Drive AGX Orin?

Thank you for your support!
It goes further than the similar suggested topic: How to obtain GPU memory information on NVIDIA DRIVE AGX Orin DEV Kit devices with DRIVE OS 6.0.8? - #5 by SivaRamaKrishnaNV

[SivaRamaKrishnaNV]

Dear @aastudil ,
nvidia-smi and jtop are not for DRIVE. Could you run deviceQuery CUDA sample to get GPU info.

nvidia@tegra-ubuntu:/usr/local/cuda/samples/1_Utilities/deviceQuery$ ./deviceQuery
./deviceQuery Starting...

 CUDA Device Query (Runtime API) version (CUDART static linking)

Detected 1 CUDA Capable device(s)

Device 0: "Orin"
  CUDA Driver Version / Runtime Version          12.1 / 11.4
  CUDA Capability Major/Minor version number:    8.7
  Total amount of global memory:                 28954 MBytes (30360248320 bytes)
  (016) Multiprocessors, (128) CUDA Cores/MP:    2048 CUDA Cores
  GPU Max Clock rate:                            1275 MHz (1.27 GHz)
  Memory Clock rate:                             1275 Mhz
  Memory Bus Width:                              256-bit
  L2 Cache Size:                                 4194304 bytes
  Maximum Texture Dimension Size (x,y,z)         1D=(131072), 2D=(131072, 65536), 3D=(16384, 16384, 16384)
  Maximum Layered 1D Texture Size, (num) layers  1D=(32768), 2048 layers
  Maximum Layered 2D Texture Size, (num) layers  2D=(32768, 32768), 2048 layers
  Total amount of constant memory:               65536 bytes
  Total amount of shared memory per block:       49152 bytes
  Total shared memory per multiprocessor:        167936 bytes
  Total number of registers available per block: 65536
  Warp size:                                     32
  Maximum number of threads per multiprocessor:  1536
  Maximum number of threads per block:           1024
  Max dimension size of a thread block (x,y,z): (1024, 1024, 64)
  Max dimension size of a grid size    (x,y,z): (2147483647, 65535, 65535)
  Maximum memory pitch:                          2147483647 bytes
  Texture alignment:                             512 bytes
  Concurrent copy and kernel execution:          Yes with 2 copy engine(s)
  Run time limit on kernels:                     No
  Integrated GPU sharing Host Memory:            Yes
  Support host page-locked memory mapping:       Yes
  Alignment requirement for Surfaces:            Yes
  Device has ECC support:                        Disabled
  Device supports Unified Addressing (UVA):      Yes
  Device supports Managed Memory:                Yes
  Device supports Compute Preemption:            Yes
  Supports Cooperative Kernel Launch:            Yes
  Supports MultiDevice Co-op Kernel Launch:      Yes
  Device PCI Domain ID / Bus ID / location ID:   0 / 0 / 0
  Compute Mode:
     < Default (multiple host threads can use ::cudaSetDevice() with device simultaneously) >

deviceQuery, CUDA Driver = CUDART, CUDA Driver Version = 12.1, CUDA Runtime Version = 11.4, NumDevs = 1
Result = PASS

Deploy YOLO models with TensorRT on Drive AGX Orin?

If you have ONNX model, you can use trtexec tool to generate TRT model or TensorRT runtime APIs to generate model and perform inference. You need to implemenet preprocessing/post processing operations. Please check TensorRT samples at /usr/src/tensorrt/samples for reference.

Please file a new topic for issues related to YOLO v8

[aastudil]

Hello, @SivaRamaKrishnaNV thank you for your clarification on this topic.
Running the CUDA sample worked successfully:

Regarding the TensorRT topic, do you want me to open a new issue? Just for you to know, the tensorRT samples are not on the provided/flashed image.

[SivaRamaKrishnaNV]

Yes. Please open a new topic for each issue to avoid cluttering of topics
Please follow https://developer.nvidia.com/docs/drive/drive-os/6.0.10/public/drive-os-linux-sdk/external/drive-quickstart-guide/dita/common/topics/installation/build-samples/build-run-sample-apps-linux.html?hl=tensorrt#ariaid-title6 or you can copy the samples to target for on target compilation and test.
The samples are available on docker at /usr/src/tensorrt/.

[aastudil]

Okey!
No problem!
Thanks