ARMdevices.net

Nvidia launes Jetson Xavier with 20x the performance of Jetson TX2 and 10x the energy efficiency with 512-core Volta GPU with Tensor Cores in an embedded module with more than 9 billion transistors it runs at under 30W, with multiple operating modes at 10W, 15W, and 30W. The Jetson Xavier ARM SoC has 6 kinds of high-performance processors on its SoC, a Volta Tensor Core GPU, an eight-core ARM64 CPU, dual NVDLA deep learning accelerators, an image processor, a vision processor and a video processor. Jetson Xavier has a peak performance of up to 30 TOPS (trillion operations per second) of mixed-precision FP32/FP16/INT8 performance. It can encode dual 4K60 H265 and decode dual 4K60 H265 at up to 12bit. Comes with 16GB LPDDR4x RAM with 137GB/s memory bandwidth, 32GB eMMC storage. It also has a dual NVDLA DL/ML Accelerator Engines which are open source available at http://nvdla.org and a 7-way VLIW Vision Accelerator. Nvidia Jetson Xavier runs using the Nvidia Isaac platform, a toolbox for the simulation, training, verification and deployment of Jetson Xavier. This robotics software consists of: Isaac SDK, APIs and tools to develop robotics algorithm software and runtime framework with fully accelerated libraries, Isaac IMX (Intelligent Machine Acceleration) applications, a collection of NVIDIA-developed robotics algorithm software, Isaac Sim, a highly realistic virtual simulation environment for developers to train autonomous machines and perform hardware-in-the-loop testing with Jetson Xavier. The Nvidia Jetson Xavier developer kit, which includes the Isaac robotics software, will be priced at $1,299, with early access starting in August from distributors worldwide.