{"id":521,"date":"2022-04-19T08:50:12","date_gmt":"2022-04-19T08:50:12","guid":{"rendered":"https:\/\/openzeka.com\/en\/?page_id=521"},"modified":"2022-04-27T06:12:38","modified_gmt":"2022-04-27T06:12:38","slug":"grace-cpu","status":"publish","type":"page","link":"https:\/\/openzeka.com\/en\/grace-cpu\/","title":{"rendered":"Grace CPU"},"content":{"rendered":"

NVIDIA GRACE CPU<\/h1><\/div>

Purpose-built to solve the world\u2019s largest computing problems.<\/p>\n<\/div><\/div><\/div><\/div><\/div>

Accelerating the Largest AI, HPC, Cloud, and Hyperscale Workloads<\/h1><\/div>

AI models are exploding in complexity and size as they improve conversational AI with hundreds of billions of parameters, enhance deep recommender systems containing tens of terabytes of data, and enable new scientific discoveries. These massive models are pushing the limits of today’s systems. Continuing to scale them for accuracy and usefulness requires fast access to a large pool of memory and a tight coupling of the CPU and GPU.<\/p>\n<\/div><\/div><\/div><\/div><\/div>

<\/div><\/div><\/div><\/div><\/div><\/div>

Products<\/h1><\/div><\/div><\/div>

NVIDIA Grace Hopper Superchip<\/h1><\/div>

NVIDIA Grace Hopper Superchip combines the Grace and Hopper architectures using NVIDIA\u00ae NVLink\u2122-C2C to deliver a CPU+GPU coherent memory model for accelerated AI and high performance computing (HPC) applications.<\/p>\n

    \n
  • CPU+GPU designed for giant-scale AI and HPC<\/li>\n
  • New 900 gigabytes per second (GB\/s) coherent interface, 7X faster than PCIe Gen 5<\/li>\n
  • 30X higher aggregate system memory bandwidth to GPU compared to DGX A100<\/li>\n
  • Runs all NVIDIA software stacks and platforms, including NVIDIA HPC, NVIDIA AI, and NVIDIA Omniverse\u2122<\/li>\n<\/ul>\n<\/div><\/div><\/div>
    <\/span><\/div><\/div><\/div>
    <\/span><\/div><\/div><\/div>

    NVIDIA Grace CPU Superchip<\/h1><\/div>

    NVIDIA Grace CPU Superchip uses the NVLink-C2C technology to deliver 144 Arm\u00ae v9 cores and 1 TB\/s of memory bandwidth.<\/p>\n

      \n
    • High performance CPU for HPC and cloud computing<\/li>\n
    • Super chip design with up to 144 Arm v9 CPU cores<\/li>\n
    • World\u2019s first LPDDR5x with ECC Memory, 1TB\/s total bandwidth<\/li>\n
    • SPECrate\u00ae2017_int_base over 740 (estimated)<\/li>\n
    • 900 GB\/s coherent interface, 7X faster than PCIe Gen 5<\/li>\n
    • 2X the packaging density of DIMM-based solutions<\/li>\n
    • 2X the performance per watt of today\u2019s leading CPU<\/li>\n
    • Runs all NVIDIA software stacks and platforms, including NVIDIA RTX, HPC, NVIDIA AI, and NVIDIA Omniverse<\/li>\n<\/ul>\n<\/div><\/div><\/div><\/div><\/div>

      Watch NVIDIA founder and CEO Jensen Huang unveil the NVIDIA Grace CPU Superchip during his keynote.<\/p>\n<\/div>