It can deliver 200 trillion operations per second and fits in the palm of your hand
Nvidia has introduced Jetson AGX Orin – a compute module which the company calls “the world’s smallest, most powerful and energy-efficient AI supercomputer for robotics and autonomous machines.”
Unveiled at the company’s GTC event, it reportedly boasts six times the processing power of its predecessor, the Jetson AGX Xavier.
Orin can deliver 200 trillion operations per second in AI workloads, and fit in the palm of your hand, the company said.
“As robotics and embedded computing transform manufacturing, healthcare, retail, transportation, smart cities and other essential sectors of the economy, the demand for processing continues to surge,” said Deepu Talla, VP and general manager of embedded and edge computing at Nvidia.
“Jetson AGX Orin addresses this need, enabling the 850,000 Jetson developers and over 6,000 companies building commercial products on it to create and deploy autonomous machines and edge AI applications that once seemed impossible.”
The module and developer kit for Jetson AGX Orin will be available in the first quarter of 2022.
AI performance: The next generation
Nvidia’s Jetson AGX upgrade comes after it launched an Industrial-focused module in June.
The latest offering is designed to allow developers to deploy more complex models to solve edge AI and robotics challenges in natural language understanding, 3D perception, and multi-sensor fusion.
It houses an Nvidia Ampere architecture GPU, as well as Arm Cortex-A78AE CPU.
Jetson AGX Orin users can utilize CUDA-X, the company’s accelerated computing stack, with pre-trained models from its NGC catalog that can be applied to customer datasets.
CPU - 12-core Arm Cortex -A78AE v8.2 64-bit CPU 3MB L2 + 6MB L3
GPU - Ampere, 2048 CUDA Cores and 64 Tensor Cores
DL Accelerators - 2x NVDLA v2.0
Memory - 32GB LPDDR5, 256-bit bus (204 GB/sec)
Storage - 64GB eMMC 5.1
AI Perf. (INT8) - 200 TOPS
Dimensions - 100mm x 87mm
TDP - 15W-50W
The AGX Orin powers Nvidia’s Clara Holoscan – a new computing platform for the healthcare industry.
Clara Holoscan is designed to allow developers to build software-defined medical devices which run low-latency streaming applications on the edge, such as robotic surgery tools.
Meanwhile, DRIVE AGX Orin is focused on transportation, and is the processor behind Nvidia’s newly announced DRIVE Concierge and DRIVE Chauffeur — two AI platforms covering in-vehicle AI assistants and autonomous driving, respectively.