Skip to content

Nvidia Embedded System

Nvidia Hardware Terminology

See in this page: Nvidia Hardware Terminology

Jetson series

It is an edge AI computing platform, within development board form.

  • Jetson Nano: for beginner
    • 128 cuda core
    • 4 core ARM CPU
    • 4GB RAM
    • 0.5 TOPS(FP16)
    • JetPack 4.6(Ubuntu 18.04)
    • CUDA 10.2
    • TensorRT7
  • Jetson Orin Nano: for small projects
    • 1024 cuda cores
    • 32 tensor cores
    • 6 core ARM CPU
    • 8GB RAM
    • 67 TOPS
    • JetPack6(Ubuntu 22.04)
    • CUDA 12.2
    • TensorRT 8.6
    • Pytorch 2.0, TensorFlow 2.x
  • Jetson Orin NX:
    • 1024 cuda cores
    • 64 tensor coresB RAM
    • 100 TOPS
  • Jetson AGX Orin
    • 2048 cuda cores
    • 64 tensor cores
    • 12 core ARM CPU
    • 275 TOPS
  • Jetson AGX Thor
    • 2560 cuda cores
    • 96 tensor cores
    • 14 core ARM CPU
    • 2070 TFLOPS(FP4 Sparse)

Drive AGX

NVIDIA DRIVE AGX gives you a scalable and energy-efficient AI computing platform designed to process the complex workloads required for autonomous driving.

Hardware

  • Nvidia Drive AGX Hyperion 10
    • 2 x Drive AGX Thor
    • Features 14 high-definition cameras, 9 radars, 1 lidar, and 12 ultrasonic sensors and a microphone array.
    • Includes DRIVE™ AV software, purpose-built for L4 autonomy.
  • Nvidia Drive AGX Thor
    • Delivers more than 1,000 INT8 TOPS (2,000 FP4 FLOPs).
    • Offers a scalable architecture supporting Level 2+ to fully autonomous driving.
    • Provides superior safety, with ASIL-D compliance and redundancy.
  • Nvidia Drive AGX Orin
    • Delivers up to 254 TOPS of AI performance.
    • Offers a scalable architecture supporting Level 2+ to fully autonomous driving.
    • Provides superior safety, with ASIL-D compliance and redundancy

Software

  • Nvidia DriveOS
  • Nvidia Drive AV

Clara AGX: AI for Medical Devices and Robotics