NVIDIA Jetson

Nvidia Jetson is a series of embedded computing boards from Nvidia. Jetson devices carry a Tegra SoC from Nvidia that integrates an ARM architecture CPU and a GPU.

Jetson is a low-power system and is designed for accelerating machine learning applications

Useful resources

• Official NVIDIA Jetson page

• Jetson Software Documentation

• Jetpack SDK

• How to Install and Configure JetPack SDK

• SDK Manager - end-to-end development environment setup solution for NVIDIA’s Jetson, Holoscan and other SDKs

• JetsonHacks

• NVIDIA Isaac ROS - a collection of NVIDIA-accelerated, high performance, low latency ROS 2 packages for making autonomous robots which leverage the power of Jetson and other NVIDIA platforms.

Tools

• nvidia-smi - NVIDIA System Management Interface program. Provides monitoring and management capabilities for NVIDIA GPUs (not available on Jetson devices)

• jetson-stats (command jtop) - A package to monitor and control your NVIDIA Jetson [Xavier NX, Nano, AGX Xavier, TX1, TX2] embedded board. It is a collection of various tools to monitor the system, including CPU, GPU, and RAM usage, temperature, and power consumption. Repository

• prime-select - Command-line tool to switch between Intel and NVIDIA graphics cards on Ubuntu (on hybrid systems)

• nvidia-settings - GUI tool to configure the NVIDIA graphics driver

• glxinfo | grep "OpenGL renderer" - to check the OpenGL renderer (to make sure the NVIDIA GPU is being used) (install: sudo apt install mesa-utils)

  • glxinfo -B - to get more detailed information about 3D acceleration setup

Nvidia Jetson Xavier NX Setup

1. Install Nvidia SDK Manager on your host machine. NB! Xavier NX requires a host machine with Ubuntu 18.04 LTS or Ubuntu 20.04 LTS (not newer) Latest supported JetPack version is 5.1.4 Easiest way to use SDK Manager in a Docker container. It can be downloaded here NB! Use image Ubuntu 18.04 or 20.04

2. NB! When installing JetPack SDK using the SDK Manager Docker image install qemu-user-static and binfmt-support on the host, Then run following commands: sudo update-binfmts --enable and sudo killall adb

3. Load the docker image: docker load -i ./sdkmanager-[version].[build#]-[base_OS]_docker.tar.gz, Then tag the image: docker tag sdkmanager:[version].[build#] sdkmanager:latest

4. Run the Docker image: docker run -it --rm --privileged -v /dev/bus/usb:/dev/bus/usb/ -v /dev:/dev -v /media/$USER:/media/nvidia:slave sdkmanager --cli

5. When ready to flash, connect Jetson’s pins 9 and 10 to boot into force recovery mode. Connect Jetson to host via a USB cable. Turn on the Jetson.

6. After flash is completed, remove the jump wire and reboot Jetson.

Note

If SDK Manager can’t detect Jetson, even though it is showing up on both host and inside the container when running lsusb, try to do step 5 before running Docker.

Nvidia SDK Manager Docker docs