CUDA is not installed on Jetson Orin

Hi,

I followed all the steps as mentioned in Getting Started with Jetson AGX Orin Developer Kit | NVIDIA Developer

But still when I ran jetson_release -v, I am getting the following message that cuda is not installed. Can you please help me?

idia@ubuntu:~/Desktop$ jetson_release -v

  • NVIDIA Jetson UNKNOWN
    • Jetpack UNKNOWN [L4T 34.1.1]
    • NV Power Mode: MAXN - Type: 0
    • jetson_stats.service: active
  • Board info:
    • Type: UNKNOWN
    • SOC Family: tegra23x - ID:
    • Module: UNKNOWN - Board: P3737-000
    • Code Name: concord
    • CUDA GPU architecture (ARCH_BIN): NONE
    • Serial Number: 1421622124328
  • Libraries:
    • CUDA: NOT_INSTALLED
    • cuDNN: 8.3.2.49
    • TensorRT: NOT_INSTALLED
    • Visionworks: NOT_INSTALLED
    • OpenCV: 4.5.4 compiled CUDA: NO
    • VPI: ii libnvvpi2 2.0.14 arm64 NVIDIA Vision Programming Interface library
    • Vulkan: 1.3.203
  • jetson-stats:
    • Version 3.1.4
    • Works on Python 3.8.10

Thank you,

Hi user158496,

Please try below two methods to install CUDA:

  1. Using apt command to install:
$ sudo apt update
$ sudo apt install cuda-toolkit-11-4
  1. Install deb file by manually:
Copy "cuda-repo-l4t-11-4-local_11.4.14-1_arm64.deb" file to target Orin
$ sudo dpkg -i cuda-repo-l4t-11-4-local_11.4.14-1_arm64.deb
$ sudo apt-key add /var/cuda-repo-11-4-local-xxxxx.pub
$ sudo apt-get update
$ sudo apt-get -y install cuda-toolkit-11-4

Hi,

I tried both approaches, but getting the same status for cuda:

nvidia@ubuntu:~/Desktop$ jetson_release -v

  • NVIDIA Jetson UNKNOWN
    • Jetpack UNKNOWN [L4T 34.1.1]
    • NV Power Mode: MAXN - Type: 0
    • jetson_stats.service: active
  • Board info:
    • Type: UNKNOWN
    • SOC Family: tegra23x - ID:
    • Module: UNKNOWN - Board: P3737-000
    • Code Name: concord
    • CUDA GPU architecture (ARCH_BIN): NONE
    • Serial Number: 1421622124328
  • Libraries:
    • CUDA: NOT_INSTALLED
    • cuDNN: 8.3.2.49
    • TensorRT: 8.4.0.11
    • Visionworks: NOT_INSTALLED
    • OpenCV: 4.5.4 compiled CUDA: NO
    • VPI: ii libnvvpi2 2.0.14 arm64 NVIDIA Vision Programming Interface library
    • Vulkan: 1.3.203
  • jetson-stats:
    • Version 3.1.4
    • Works on Python 3.8.10

in the sdkmanager, all the packages installed successfully.
I tried to flash the jetson orin with jetpack version 5.0 but still the same status for cuda.
This is the output of nvcc --version:

nvidia@ubuntu:~/Desktop$ nvcc --version
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2021 NVIDIA Corporation
Built on Thu_Nov_11_23:44:05_PST_2021
Cuda compilation tools, release 11.4, V11.4.166
Build cuda_11.4.r11.4/compiler.30645359_0

nvidia@ubuntu:~/Desktop$ python3
Python 3.8.10 (default, Jun 22 2022, 20:18:18)
[GCC 9.4.0] on linux
Type “help”, “copyright”, “credits” or “license” for more information.

import torch
torch.cuda.is_available()
False

This is the output of sudo apt install cuda-toolkit-11-4

Can you please help me with this error?

Many Thanks,

I think we need to clarify this…

Are you able to use CUDA or not? I mean currently it looks like the issue is “jetson_release -v” cannot tell you whether the CUDA is installed or not.

But nvcc is already installed and gave you ther version number. The issue looks like just “jetson_release” does not work well but not “cuda cannot be installed”.

Also, just in case you don’t know, this “jetson_release” is not an official tool released by us, so we are not sure whether it has problem in detecting or not.

I believe that jetson_stats has not been released by @rbonghi for JetPack 5.0 yet, so that may explain why jetson_release tool isn’t reporting correctly. If CUDA Toolkit can be found under /usr/local/cuda, then it should already be installed.

Hi,

No, I am not able to use Cuda, the applications that need cuda, it will give me the error of Cuda not installed as well.

Is there anyway to make Cuda visible to system, I enabled CUDA_VISIBLE_DEVICES=0, 1 but still cannot detect cuda.

Can you please help me to make cuda visible to system?

Thank you,

Hi,

The /usr/local/cuda is already there

and this is the last lines in ~/.bashrc:

export PATH=/usr/local/cuda-11.4/bin:$PATH
export LD_LIBRARY_PATH=/usr/local/cuda-11.4/lib64:$LD_LIBRARY_PATH

Thank you,

Are you able to run deviceQuery okay?

$ cd /usr/local/cuda/samples/1_Utilities/deviceQuery
$ sudo make
$ ./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          11.4 / 11.4
  CUDA Capability Major/Minor version number:    8.7
  Total amount of global memory:                 30623 MBytes (32110190592 bytes)
  (008) Multiprocessors, (128) CUDA Cores/MP:    1024 CUDA Cores
  GPU Max Clock rate:                            1300 MHz (1.30 GHz)
  Memory Clock rate:                             1300 Mhz
  Memory Bus Width:                              128-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 = 11.4, CUDA Runtime Version = 11.4, NumDevs = 1
Result = PASS

If that works, then your CUDA driver should be fine.

2 Likes

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.