Grace - AshokBhat/notes GitHub Wiki

About

  • NVIDIA Grace Arm server CPU
  • Target segment: HPC, AI in data-centers
  • General availablity: H1, 2023

Specification

  • Armv9 Neoverse Core
  • DDR5 with 500 GB/sec of bandwidth
  • Native NVLink 4.0 ports

Grace Hopper Superchip

  • Grace CPU (72 Armv9 CPU core) and Hopper GPU
  • NVLink-C2C interconnect
  • CPU+GPU coherent memory model
  • 900 gigabytes per second (GB/s) coherent interface - 7X faster than PCIe Gen 5
  • 30X higher aggregate system memory bandwidth to GPU compared to DGX A100

Grace CPU Superchip

  • Two Grace CPU units (with 72 Arm v9 CPU cores)
  • 1 TB/s of memory bandwidth.
  • LPDDR5x
  • 900 GB/s coherent interface, 7X faster than PCIe Gen 5

Deployment

CSCS Alps supercomputer

  • At Swiss National Supercomputing Center (CSCS)
  • In early 2023
  • Grace Arm server CPU + Hopper Nvidia GPU
  • 20 exaflops of AI performance
  • HPE Cray system

LANL Venado supercomputer

  • Grace Arm server CPU + Hopper Nvidia GPU
  • HPE Cray system
  • In early 2023.
  • 10 exaflops of AI performance

See also