NVIDIA Jetson Support from MATLAB Coder and GPU Coder
Generate and deploy optimized CUDA code on embedded NVIDIA Jetson boards
- NVIDIA GPU Support from GPU Coder
- NVIDIA Jetson Support from MATLAB Coder and GPU Coder
System requirements and installation options
Check the System Requirements
To install a hardware support package, you must have a supported product release, along with the required operating system and base product.
View the hardware support package system requirements table and confirm you have:
- A supported release
- A supported operating system
- The required base products
If you do not have a supported release or the required base products, you can get started with a free trial, or you can purchase products.
Continue with the Installation
After confirming you have the system requirements, proceed with either installation option listed below.
Option 1: Using the Toolstrip
Launch the Support Package Installer from your version of MATLAB and navigate to the correct support package.
Option 2: "Get Support Package" Button (.mlpkginstall file*)
Clicking the Get Support Package button provides the support package install file:
([filename].mlpkginstall)
There are several ways to initiate the Support Package Installer from these files:
- Open the
.mlpkginstallfile directly from your Internet browser. - Download and save the file (the default option for some Internet browsers), and open it directly from your Windows, Mac, or Linux operating system.
- Open the
.mlpkginstallfile directly from within MATLAB by navigating the Current Folder to the location of the downloaded file and double clicking it. That will initiate the installation for the MATLAB release that you have open.
The primary difference between using the .mlpkginstall file (option 2) rather than the toolstrip (option 1) to initiate support package installation is that the file allows the user to avoid the support package selection screen on the installer.
*These file types were introduced in MATLAB R2013a.
Users without an Internet connection can install hardware support packages by following the download instructions.
Capabilities and Features
MATLAB Coder support package for NVIDIA® Jetson™ and NVIDIA DRIVE™ platforms automates the deployment of MATLAB algorithms or Simulink models to embedded NVIDIA GPUs such as the Jetson platform. Prototype and develop your MATLAB algorithms and Simulink models with data from the Jetson peripherals, then use MATLAB Coder or GPU Coder to generate optimized C/C++/CUDA® code and deploy it to the Jetson board to run as a standalone executable.
Interactive communication: Remotely communicate with the NVIDIA target from MATLAB and Simulink to acquire data from supported sensors and imaging devices connected to the target and then analyze and visualize it in MATLAB. Log data from supported sensors to help fine-tune your algorithm for early prototyping.
Standalone execution: Deploy the generated C/C++/CUDA code as a standalone embedded application on the Jetson platform. Build and deploy the generated C/C++/CUDA code from your MATLAB algorithm and Simulink models, along with the interfaces to the peripherals and sensors, to the Jetson platform.
The support package supports the latest Jetson boards from NVIDIA, including the NVIDIA Jetson AGX Thor™, Jetson AGX Orin™, Jetson Orin Nano, and Jetson AGX Xavier™ developer kits. See the latest list of supported hardware for MATLAB Coder Support Package for NVIDIA Jetson and NVIDIA DRIVE Platforms.
