Skip to content


android  notes  hal  aidl  hidl  kernel  arduino  esp32  book  management  c/c++  arm  freertos  stm32  debug  sysview  jetson  linux  vuejs  nodejs  electron  chromium  dmm  hardware  raspberry-pi  camera  i2c  ffmpeg  performance  v4l2  picamera  stream  hls  dash  python  h264  mjpeg  ros  jinja  javascript  mkdocs  markdown  encryption  asm  rtt  swv  virtual-machine  docker  hyper-v  serial  yocto  gnss  lidar 


Nvidia Jetson »

NVIDIA Jetson

NVIDIA® Jetson™ is the world's leading platform for AI at the edge. It's well known to deliver the performance and power-efficiency that customers need to build software-defined intelligent on advanced embedded systems.

#jetson  

Last update: 2022-05-07

Table of Content

Jetson platform#

The Jetson platform includes small, power-efficient developer kits and production modules that offer high-performance acceleration of the NVIDIA CUDA-X™ software stack.

They have 2 main product lines:

Jetson development kits
Each developer kit includes a non-production specification Jetson module attached to a reference carrier board with standard hardware interfaces for flexible development and rapid prototyping. The development module usually has a micro-SD Card. These kits are not intended for production use.
Jetson modules

Jetson modules come with production specification which usually has built-in eMMC storage, more advanced parts, and longer operating lifetime. Each Jetson module ships with no software pre-installed; need to be attached to a carrier board designed or procured for your end product, and flash it with the software image you’ve developed.

Jetson modules have many signals in common, but the connector pin-out and electromechanical footprints do vary:

  • Jetson Nano, Jetson TX2 NX, and Jetson Xavier NX modules are pin compatible.
  • These Jetson TX2 series modules are pin compatible, including Jetson TX2, Jetson TX2 4GB, and Jetson TX2i
  • Jetson AGX Xavier series modules (Jetson AGX Xavier and Jetson AGX Xavier Industrial) are pin compatible.

Comparison between Developement Kits and Modules

JetPack#

NVIDIA JetPack SDK is the most comprehensive solution for building AI applications. All Jetson modules and developer kits are supported by JetPack SDK.

JetPack SDK includes the Jetson Linux Driver Package (L4T) with Linux operating system and CUDA-X accelerated libraries and APIs for Deep Learning, Computer Vision, Accelerated Computing and Multimedia. It also includes samples, documentation, and developer tools for both host computer and developer kit, and supports higher level SDKs such as DeepStream for streaming video analytics and Isaac for robotics.

JetPack is available at https://developer.nvidia.com/embedded/jetpack.

JetPack is included either in the Nvidia SD card Images which run on Jetson modules, or in SDK Manager which runs on a Host Machine (usually an Ubuntu OS machine).

System size

The default developement system images provided by Nvidia are quite large, at about 13 GB. SDK Manager can make a smaller system image, at about 5 GB, which will not include JetPack developement tools and libraries, and suitable to be flashed on eMMC storage.

If you want a smaller image size, you can build a custome image from a minial RootFS base.

Get started#

You should start with the provided image from Nvidia by following the starter guide, such as Jetson Nano Developer Kit.

After the first boot, you should start reading the documents, which are provided in the Jetson Download Center.

One important point you have to learn is to customize the system image which only contains what you need. Read Linux for Tegra.

Comments