On May 14th at the NVIDIA GTC 2020, the new NVIDIA Jetson Xavier NX Developer Kit was officially released. And now it is available at Seeed Studio online store. NVIDIA® Jetson Xavier™ NX Developer Kit includes an NVIDIA® Jetson Xavier™ NX Module which features a GPU of NVIDIA Volta architecture with 384 NVIDIA CUDA® cores and 48 Tensor cores. The CPU is based on 6-core NVIDIA Carmel ARM®v8.2 64-bit CPU 6 MB L2 + 4 MB L3.
In this article, we will take you to have a close look at the kit NVIDIA Jetson Xavier NX Developer Kit and also compare it with Jetson TX2.
JETSON XAVIER NX Developer Kit is at the same price as JETSON TX2 Developer Kit, however, compare with Jetson TX, JETSON XAVIER NX has improved a lot.
Compare NVIDIA Jetson Xavier NX Developer Kit with Jetson TX2
|Jetson TX2 Developer Kit||Jetson Xavier NX|
|Compute Performance||1.3 TFLOPS (FP16)||6 TFLOPS (FP16)|
21 TOPS (INT8)
|GPU||256-core NVIDIA Pascal™ GPU architecture with 256 NVIDIA CUDA cores||NVIDIA Volta architecture with 384 NVIDIA CUDA® cores and 48 Tensor cores|
|CPU||Dual-Core NVIDIA Denver 2 64-Bit CPU|
Quad-Core ARM® Cortex®-A57 MPCore
|6-core NVIDIA Carmel ARM®v8.2 64-bit CPU 6 MB L2 + 4 MB L3|
|Memory||8GB 128-bit LPDDR4|
1866 MHz – 59.7 GB/s
|8 GB 128-bit LPDDR4x @ 51.2GB/s|
Jetson can be used to deploy various popular DNN models and ML frameworks to the edge with high-performance inference to perform tasks such as real-time classification and object detection, pose estimation, semantic segmentation, and natural language processing (NLP).
A smaller size of NVIDIA® Jetson Xavier™ NX Developer Kit.
In size, the Jetson NX developer kit is much smaller than the Jetson TX2 developer kit. With such a small size, it is suitable for applications of autonomous cars and drones.
More MIPI CSI camera connectors
The Jetson Xavier NX development kit has two MIPI CSI camera connectors to easily connect two HD Raspberry Pi V2 cameras.
Four onboard USB 3.1 interfaces
Compared with the TX2 development kit with only one USB 3.0 interface, the Xavier development kit has four USB 3.1 interfaces. In terms of data transmission speed: USB3.1 is much faster than USB 3.0.
3.1 can reach 10G / s. However, 3.0 can only reach 4.8G / s.
The TX2 development kit cannot be externally connected to an SSD. Developers can only expand Jetson TX2 through an SD card or an external SATA hard drive. However, an M.2 Key-M NVMe slot attached at the bottom of the NX developer kit, which can be easily connected to an SSD to expand high-speed storage.
Support the latest software package, Jetpack 4.4
From NVIDIA’s introduction, it is said that “NVIDIA JetPack SDK is the most comprehensive solution for building AI applications. Use NVIDIA SDK Manager to 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, and documentation needed to jumpstart your development environment.”
This release adds support for the Jetson Xavier NX module and includes new versions of CUDA, TensorRT, and cuDNN.
In addition, DeepStream 5.0 Developer Preview is supported in this release.
JetPack (Jetson SDK) is an on-demand all-in-one software package that bundles development software for NVIDIA® Jetson embedded platform. JetPack includes support for the latest L4T BSP software package. Use the latest BSP to automatically refresh your Jetson development kit and install the latest software tools needed to build and configure Jetson embedded platform applications.
The latest release of Jetpack 4.4:
JetPack 4.4 Developer Preview Highlights:
- Support for upgrading JetPack and L4T using debian package management tool
- Support for Generic Timestamping Engine (GTE) for Jetson AGX Xavier and Jetson Xavier NX
- Support for Dynamic Frequency Scaling (DFS) for Video Image Compositor (VIC) using actmon
- SE (Security Engine) samples to demonstrate hardware backed authentication and encryption capabilities of Jetson TX2 series, Jetson AGX Xavier and Jetson Xavier NX modules
- Utility to fuse multiple Jetson modules simultaneously
- Option to specify APP partition size on the microSD card during initial configuration at first boot of Jetson Xavier NX and Jetson Nano Developer Kits
With multiple developments and deployment containers for Jetson hosted on NVIDIA NGC, containerized frameworks, and pre-trained AI models, it can be used as a building block for AI application development. The latest Jetson Xavier NX implements full-featured, multi-modal AI applications in the smallest possible size.
Inferencing performance across Jetson Nano, Jetson TX2, Jetson Xavier NX, and Jetson AGX Xavier
As the following picture shows, NVIDIA officially tested inferencing performance across Jetson Nano, Jetson TX2, Jetson Xavier NX, and Jetson AGX Xavier on popular DNN models for image classification, object detection, pose estimation and semantic segmentation. Below is the inference benchmark results for JetPack and NVIDIA’s TensorRT inferencing accelerator library that optimizes networks, trained in popular ML frameworks for realtime performance.