Jetpack 4.3 released for your Jetson Modules

For those who are using Jetson modules including Jetson AGX Xavier Series, Jetson TX2 Series, Jetson TX1 and Jetson Nano, you probably are familiar with Jetpack.

About Jetpack

For those who do not know what Jetpack is, NVIDIA Jetpack SDK is the most comprehensive solution for building AI applications. With the NVIDIA SDK Manager, you can flash your Jetson developer kit with the latest OS image, install developer tools for both host computer and developer kit and install the libraries and APIs, samples together with documentation needed to jumpstart your development environment

Newly Released Jetpack 4.3

With the newly released Jetpack 4.3, they include new versions of TensorRT, cuDNN, improved AI inference performance by 25% and full INT8 support for DLA on Jetson AGX Xavier which further improves the performance and efficiency of DLA, CSI and encode with containers. In addition, a new Debian package based installation mechanism for all JetPack components.

Jetpack 4.3 Key Features

Operating System

NVIDIA L4T provides the bootloader, Linux Kernel, necessary firmware, NVIDIA drivers, sample filesystem, etc.

Jetpack 4.3 Features:

• Support multiple Jetson devices at the same time
• Added USB 3.0 capabilities to Jetson TX1 / Nano and TX2 Homes.
• Support installation of JetPack components via Debian package archives. Makes installation easier and upgrading to future versions
• NVIDIA Container Runtime support for DLA, CSI and Encode from within containers
  • For Jetson AGX Xavier Series: NVIDIA Deep Learning Accelerator (DLA) engines are now supported from within containers.
  • For all Jetson products: Camera Serial Interface (CSI) and NVENC are now supported from within containers.
• Supports DeepStream 4.0.2
• Supports ISAAC SDK version 2019.3

TensorRT

Platform for high-performance deep learning inference. It includes a deep learning inference optimizer and runtime that delivers low latency and high-throughput for deep learning inference applications.

With TensorRT, you can optimize neural network models trained in all major frameworks, calibrate for lower precision with high accuracy, and finally deploy to hyperscale data centers, embedded, or automotive product platforms.

Jetpack 4.3 Features:

• New layers, operators and resize operations
• New samples
• New optimizations

cuDNN

Which stands for CUDA Deep Neural Network Library, provides high-performance primitives for deep learning frameworks. It
includes support for convolutions, activation functions, and tensor transformations.

Jetpack 4.3 Features:

• Performance enhancements for grouped convolution.
• Improved depth-wise convolution and grouped convolution.

Computer Vision

VisionWorks is a software development package for Computer Vision and image processing.

OpenCV is the leading open-source library for computer vision, image processing, and machine learning, and now features GPU acceleration for real-time operation.

VPI is a software library that provides Computer Vision / Image Processing algorithms implemented on PVA¹ (Programmable Vision Accelerator), GPU² and CPU².

Jetpack 4.3 Features:

• VPI now available as a developer preview for Jetpack 4.3

Developer Tools

CUDA Toolkit provides a comprehensive development environment for C and C++ developers building high-performance GPU-accelerated applications with CUDA libraries. 

Jetpack 4.3 Features:

• NVIDIA Nsight Systems for application profiling across GPU and CPU.
  • Includes timeline enhancements to improve identification of tiny events
  • Improved error reporting
  • Option to disable CPU periodic sampling and thread state backtrace
• NVIDIA Nsight Graphics for graphics application debugging and profiling.
  • Adds support for new Vulkan extension
  • Support for immediate mode
  • Improved visibility and detail in the Events View

For more information and details, you can check out the Jetpack 4.3 release notes!

Where can I get Jetpack 4.3?

To get Jetson Nano Developer Kit SD Card Image click here!

• For instructions, follow Jetson Nano guide here.

To get Jetson AGX Xavier Series, Jetson TX2 series, Jetson TX1 and Jetson Nano NVIDIA SDK manager click here!

• For instructions, follow Jetson Software guide for SDK manager here.

Want to use the all-new NVIDIA Jetpack but do not own a Jetson module?

You are in luck! Get a Jetson Nano now as we have just lowered the price of the NVIDIA Jetson Nano to only $89!

Get the NVIDIA Jetson Nano to get started on your Machine Learning and AI journey now!

Interested in more NVIDIA products? You can check out all of our NVIDIA products here!

About Author