"NVIDIA Releases Jetson Xavier NX: The Smallest and Most Powerful Supercomputer, Focusing on AI Edge Computing, Priced at $399"

2019-11-08

On November 7th, according to foreign media reports, NVIDIA has released a new AI processor, Jetson Xavier™ NX, which can meet the computing needs of next-generation AI-driven autonomous vehicles, delivery robots, or drones. NVIDIA claims that Jetson Xavier™ NX is the world's smallest AI supercomputer. The latest member of the Jetson product family, Jetson Xavier NX, combines the compact size of Nano and the powerful performance of Xavier, all for just $399. With a compact form factor smaller than a credit card, the energy-efficient Jetson Xavier NX module provides up to 21 TOPS of server-level performance while consuming only 10 watts of power. Jetson Xavier NX opens the door for many embedded edge computing devices, which have high performance requirements but are limited by size, weight, power budget, or cost. These devices include small commercial robots, drones, intelligent high-resolution sensors (used in factory logistics and production lines), optical inspection, network video recorders, portable medical devices, and other industrial IoT systems. Deepu Talla, Vice President and General Manager of Edge Computing at NVIDIA, said, "AI is empowering modern robots and embedded devices, transforming entire industries. Many devices are limited by small size and low power, making it impossible to add AI capabilities. With Jetson Xavier NX, our customers and partners can significantly improve AI performance without increasing device size or power consumption." Jetson Xavier NX can deliver up to 14 TOPS of performance (at 10W power consumption) or 21 TOPS (at 15W power consumption), allowing for parallel execution of multiple neural networks and simultaneous processing of data from multiple high-resolution sensors in a compact form factor (70x45mm) similar to Nano. For companies already building embedded computers, Jetson Xavier NX can run on the same CUDA-X AI™ software architecture as all Jetson series products, ensuring fast time to market and lower development costs. As part of NVIDIA's software architecture approach, Jetson Xavier NX is supported by the NVIDIA JetPack SDK™. The NVIDIA JetPack SDK™ is a complete AI software stack that can run modern complex AI networks and is used for deep learning acceleration libraries, as well as computer vision, computer graphics, multimedia, and more. Ecosystem Support Jetson Xavier NX has received strong support from the robotics and embedded device ecosystem. "The NVIDIA Jetson embedded series of products has accelerated research, development, and deployment of embedded AI solutions on Lockheed Martin platforms," said Lee Ritholtz, Director of Applied AI and Chief Architect at Lockheed Martin. "With Jetson Xavier NX's outstanding performance, compact form factor, and low power consumption, we will be able to perform more real-time processing at the edge. Our goal is to significantly improve the quality and accuracy of optical inspection systems and accelerate the pace towards Industry 4.0." Otsuka Hiroshi, CEO of MuSAShi Seimitsu, said, "NVIDIA Jetson Xavier NX provides us with the computing power to enhance our visual inspection capabilities without increasing the size and power of our optical inspection systems." NVIDIA also announced today that it leads in all five benchmark tests measuring data center and edge AI inference workload performance. This means that while NVIDIA has achieved a leading position in recent AI training benchmark tests, it has further advanced its position. MLPerf Inference 0.5 is the industry's first independent AI inference benchmark test, and the results show strong inference performance when using NVIDIA Turing™ GPUs for data centers and NVIDIA Jetson Xavier™ system-on-a-chip (SoC) for the edge. The Jetson Xavier NX module is built based on the low-power version of the Xavier SoC used in these benchmark tests. "In this rapidly evolving world of AI chip development, I believe NVIDIA has raised the industry standard with Jetson Xavier NX. The small size, low power Jetson Xavier NX offers outstanding performance and a unified and powerful software architecture, which is crucial for embedded edge computing," said Patrick Moorhead, President and Chief Analyst at Moor Insights & Strategy. Jetson Xavier NX Module Specifications: - GPU: NVIDIA Volta with 384 NVIDIA CUDA cores, 48 Tensor cores, and 2 NVDLA - CPU: 6-core Carmel Arm 64-bit CPU, 6MB L2 + 4MB L3 - Video: 2x 4K30 encoding and 2x 4K60 decoding - Camera: Up to 6 CSI cameras (up to 36 via virtual channels), 12 lanes (3x4 or 6x2) MIPI CSI-2 - Memory: 8GB 128-bit LPDDR4x; 51.2GB/s - Connectivity: Gigabit Ethernet - OS Support: Ubuntu-based Linux - Module Size: 70x45mm Jetson Xavier NX is the latest member of the Jetson family, which also includes Jetson Nano™, Jetson AGX Xavier™ series, and Jetson TX2 series. Jetson Xavier NX provides a rich set of IOs, from high-speed CSI and PCIe to low-speed I2C and GPIO. It is compatible with many peripheral devices and sensors, and with its small form factor and powerful performance, it brings new capabilities to embedded AI and industrial IoT systems. In addition, Jetson Xavier NX is pin-compatible with Jetson Nano, supporting shared hardware designs. Users can upgrade to Jetson Xavier NX using the Jetson Nano carrier board and system. It also supports all major AI frameworks, including TensorFlow, PyTorch, MxNet, Caffe, and more. The Jetson Xavier NX module is priced at $399 and will be available from NVIDIA distribution channels in March next year, targeting companies that want to create edge systems in large quantities. Developers can now simulate Jetson Xavier NX by downloading software patches and using the Jetson AGX Xavier development kit for immediate development. Original article link: [https://nvidianews.nvidia.com/news/nvidia-announces-jetson-xavier-nx-worlds-smallest-supercomputer-for-ai-at-the-edge?ncid=so-twit-94734#cid=em02_so-twit_en-us](https://nvidianews.nvidia.com/news/nvidia-announces-jetson-xavier-nx-worlds-smallest-supercomputer-for-ai-at-the-edge?ncid=so-twit-94734#cid=em02_so-twit_en-us)

RECENT AI NEWS

RECENT AI TOOLS