SX220-1N

  • CPU+GPU Integration - 
    • Support the NVIDIA® MGX™ architecture with the powerful NVIDIA GH200 Grace Hopper Superchip
    • NVIDIA® NVLink®-C2C to deliver a CPU+GPU coherent memory model
  • Unified Memory Architecture -
    • Enabling massive AI/ML workloads with up to 576 GB of LPDDR5X and 96 GB of HBM3
  • AI/ML Optimized – 
    • Targeted for HPC, large-scale LLMs, scientific computing, and simulation workloads.

Downloads

Get the latest on AI infrastructure

Contact