Product Overview
Key Applications High Performance Computing AI/Deep Learning Training and Inference Large Language Model (LLM) and Generative AI Key Features High density 1U GPU system with Integrated NVIDIA庐 H100 GPU NVIDIA Grace Hopper驴 Superchip (Grace CPU and H100 GPU) NVLink庐 Chip-2-Chip (C2C) high-bandwidth, low-latency interconnect between CPU and GPU at 900GB/s Up to 576GB of coherent memory per node including 480GB LPDDR5X and 96GB of HBM3 for LLM applications
