NVIDIA MGX Server (Grace / Grace Hopper)
NVIDIA Grace™ is a revolutionary CPU designed specifically for the needs of modern AI data centers. It is optimized to handle massive amounts of data and power a wide range of workloads, including AI, HPC, data analytics, digital twins, and hyperscale cloud applications.
Take a closer look at the superchip
The NVIDIA GH200 Grace Hopper Superchip combines the Grace and Hopper architectures using NVIDIA® NVLink®-C2C to deliver a CPU+GPU coherent memory model for accelerated AI and HPC applications.
- CPU+GPU designed for giant-scale AI and HPC
- New 900 gigabytes per second (GB/s) coherent interface, 7X faster than PCIe Gen5
- 30X higher aggregate system memory bandwidth to GPU compared to NVIDIA DGX™ A100
- Runs all NVIDIA software stacks and platforms, including the NVIDIA HPC SDK, NVIDIA AI, and NVIDIA Omniverse™
NVIDIA® Grace™ Hopper™
- Supermicro
- GPU ARS-221GL-NR
CPU | NVIDIA Grace™ CPU Superchip 500W TDP |
Memory | Up to 480GB DDR5-4800MHz ECC RDIMM |
GPU | Up to 4 NVIDIA H100 80GB |
Storage | Up 16x E.1S hot-swap NVMe drive slots |
Power Supply | Redundant (2+1) 2000W Titanium Level Power Supply |
4 GPU
- Supermicro
- ARS-111GL-NHR
CPU | NVIDIA GH200 Grace Hopper™ Superchip, up to 72C/144T, Supports up to 1000W TDP CPUs (Air Cooled) |
Memory | Onboard Memory, up to 480GB ECC LPDDR5X, Additional GPU Memory: Up to 96GB ECC HBM3 |
GPU | Up to 1 onboard GPU(s), NVIDIA: GH200 Grace Hopper™ Superchip (Air-cooled) |
Storage | 8x E1.S hot-swap NVMe drive slots, 2 M.2 NVMe |
Power Supply | 2x 2000W Redundant Titanium Level power supplies |
- QCT
- QuantaGrid S74G-2U
CPU | NVIDIA GH200 Grace™ Hopper™ Superchip 1000W TDP |
Memory | Up to 480GB LPDDRX embedded 96GB HBM3 GPU memory |
GPU | NVIDIA GH200 Grace™ Hopper™ Superchip 1000W TDP |
Storage | Up 4x E.1S NVMe drive slots |
Power Supply | (1+1) 2000W Titanium Level Power Supply |