Install Deepstream SDK on ARM64 SBSA without using docker container

Hi, I am running an EC2 instances server with ARM64 CPU and Nvidia GPU. And I want to install Deepstream SDK on this hardware. I have found on the internet but there is no support for it except using docker container nvcr.io/nvidia/deepstream:7.0-triton-arm-sbsa. I want to ask if it is possible to install the Deepstream SDK on this hardware. Thank you!

Below is my hardware

• Hardware Platform (Jetson / GPU): GPU NVIDIA T4G
• DeepStream Version: 6.26.3
• TensorRT Version: 8.4.1-1
• NVIDIA GPU Driver Version (valid for GPU only): 550.90.07

| NVIDIA-SMI 550.90.07              Driver Version: 550.90.07      CUDA Version: 12.4     |
|-----------------------------------------+------------------------+----------------------+
| GPU  Name                 Persistence-M | Bus-Id          Disp.A | Volatile Uncorr. ECC |
| Fan  Temp   Perf          Pwr:Usage/Cap |           Memory-Usage | GPU-Util  Compute M. |
|                                         |                        |               MIG M. |
|=========================================+========================+======================|
|   0  NVIDIA T4G                     On  |   00000000:00:1F.0 Off |                    0 |
| N/A   25C    P8              9W /   70W |       1MiB /  15360MiB |      0%      Default |
|                                         |                        |                  N/A |
+-----------------------------------------+------------------------+----------------------+
                                                                                         
+-----------------------------------------------------------------------------------------+
| Processes:                                                                              |
|  GPU   GI   CI        PID   Type   Process name                              GPU Memory |
|        ID   ID                                                               Usage      |
|=========================================================================================|
|  No running processes found                                                             |
+-----------------------------------------------------------------------------------------+

• CPU:

  CPU op-mode(s):         32-bit, 64-bit
  Byte Order:             Little Endian
CPU(s):                   8
  On-line CPU(s) list:    0-7
Vendor ID:                ARM
  Model name:             Neoverse-N1
    Model:                1
    Thread(s) per core:   1
    Core(s) per socket:   8
    Socket(s):            1
    Stepping:             r3p1
    BogoMIPS:             243.75
    Flags:                fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm lrcpc dcpop asimddp ssbs

We don’t have official guideline yet, but I think it’s possible. Please refer to the x86 installation guideline and then change the corresponding dependency to ARM64 SBSA. Especially CUDA and TensorRT.

https://docs.nvidia.com/metropolis/deepstream/dev-guide/text/DS_Installation.html#dgpu-setup-for-ubuntu

 sudo add-apt-repository "deb https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/sbsa/  /"

I still recommend you to use docker image, it is very convenient

Hi, Thank you for replying to me. I have tried to install arm64.deb file from this link DeepStream | NVIDIA NGC
but it required me to install the libnvvpi2. But I cannot install the library although I already added the apt repo

FYI: there is not Docker image for version 6.3, only version 7.1 has docker image for arm-sbsa

Which version of Ubuntu did you install?

Please follow the instructions to install Ubuntu 22.04 and DS-7.1. CUDA and TensorRT must also be installed with the specified version, otherwise they cannot be guaranteed to work properly

DS-6.3 has not been tested on ARM64 SBSA and vpi2 is no longer supported.

I checked the docker layer and found one layer that copies the built file, how can I download this file
image

This tarball is not currently available for download. You can extract it from the docker image /opt/nvidia/deepstream/deepstream. However, this is not recommended.

DeepStream does not support SBSA/dGPU on ARM with native debian/tarball installation yet. We will check this issue in subsequent versions. Thanks

You also can refer this documentation.

https://docs.nvidia.com/metropolis/deepstream/dev-guide/text/DS_Installation.html#dgpu-on-arm-setup-for-ubuntu-alpha

There is no update from you for a period, assuming this is not an issue anymore. Hence we are closing this topic. If need further support, please open a new one. Thanks

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