Customer use Cases for NVIDIA Grace Hopper Superchip and Grace CPU Superchip
, Scientist 5, Los Alamos National Laboratory
, Associate Professor, Earthquake Research Institute, The University of Tokyo
, Director, Data Center Product Marketing, NVIDIA
The Grace CPU is NVIDIA’s first CPU, and it's a high-performance and power-efficient Arm CPU that includes a high-performance NVIDIA-designed fabric and power-efficient and high-bandwidth LPDDR5 memory optimized for the data center. The NVIDIA Grace CPU is used in multiple data center NVIDIA superchip products that are found in the DGX platform, communication service provider, and manufacturer configurations. The products include the NVIDIA GH200 Grace Hopper Superchip that pairs a power-efficient NVIDIA Grace CPU with an NVIDIA H100 or H200 Tensor Core GPU to maximize the capabilities for accelerated computing and generative AI workloads. The second is the NVIDIA Grace CPU Superchip, the first no-compromise CPU for HPC, demanding cloud, and enterprise computing workloads. Both superchips are made possible by NVIDIA NVLink Chip-2-Chip, a coherent high-bandwidth interconnect. Hear from some of the early customers working with these innovative products as they share their experiences transitioning to Grace CPU Superchip or GH200 and learn about the performance and energy efficiency of the platforms.
活動:
日期:
產業:
技術水平需求:
領域:
NVIDIA technology: Base Command,BlueField DPU,Cloud / Data Center GPU,DGX,Grace CPU,Hopper