NVIDIA Grace Family: Revolutionizing Data Center Efficiency

The exponential growth in data processing demand is projected to reach 175 zettabytes by 2025, according to the NVIDIA Technical Blog. This surge contrasts sharply with the slowing pace of CPU performance improvements, highlighting the need for more efficient computing solutions.

Addressing Efficiency with NVIDIA Grace CPU

NVIDIA's Grace CPU family is designed to tackle this challenge. The first CPU developed by NVIDIA to power the AI era, the Grace CPU features 72 high-performance, power-efficient Arm Neoverse V2 cores, NVIDIA Scalable Coherency Fabric (SCF), and high-bandwidth, low-power LPDDR5X memory. The CPU also boasts a 900 GB/s coherent NVLink Chip-to-Chip (C2C) connection with NVIDIA GPUs or other CPUs.

The Grace CPU supports multiple NVIDIA products and can pair with NVIDIA Hopper or Blackwell GPUs to form a new type of processor that tightly couples CPU and GPU capabilities. This architecture aims to supercharge generative AI, data processing, and accelerated computing.

Next-Generation Data Center CPU Performance

Data centers face constraints in power and space, necessitating infrastructure that delivers maximum performance with minimal power consumption. The NVIDIA Grace CPU Superchip is designed to meet these needs, offering outstanding performance, memory bandwidth, and data-movement capabilities. This innovation promises significant gains in energy-efficient CPU computing for data centers, supporting foundational workloads such as microservices, data analytics, and simulation.

Customer Adoption and Momentum

Customers are rapidly adopting the NVIDIA Grace family for various applications, including generative AI, hyper-scale deployments, enterprise compute infrastructure, high-performance computing (HPC), and scientific computing. For instance, NVIDIA Grace Hopper-based systems deliver 200 exaflops of energy-efficient AI processing power in HPC.

Organizations such as Murex, Gurobi, and Petrobras are experiencing compelling performance results in financial services, analytics, and energy verticals, demonstrating the benefits of NVIDIA Grace CPUs and NVIDIA GH200 solutions.

High-Performance CPU Architecture

The NVIDIA Grace CPU was engineered to deliver exceptional single-threaded performance, ample memory bandwidth, and outstanding data movement capabilities, all while achieving a significant leap in energy efficiency compared to traditional x86 solutions.

The architecture incorporates several innovations, including the NVIDIA Scalable Coherency Fabric, server-grade LPDDR5X with ECC, Arm Neoverse V2 cores, and NVLink-C2C. These features ensure that the CPU can handle demanding workloads efficiently.

NVIDIA Grace Hopper and Blackwell

The NVIDIA Grace Hopper architecture combines the performance of the NVIDIA Hopper GPU with the versatility of the NVIDIA Grace CPU in a single Superchip. This combination is connected by a high-bandwidth, memory-coherent 900 GB/s NVIDIA NVLink Chip-2-Chip (C2C) interconnect, delivering 7x the bandwidth of PCIe Gen 5.

Meanwhile, the NVIDIA GB200 NVL72 connects 36 NVIDIA Grace CPUs and 72 NVIDIA Blackwell GPUs in a rack-scale design, providing unparalleled acceleration for generative AI, data processing, and high-performance computing.

Software Ecosystem and Porting

The NVIDIA Grace CPU is fully compatible with the broad Arm software ecosystem, allowing most software to run without modification. NVIDIA is also expanding its software ecosystem for Arm CPUs, offering high-performance math libraries and optimized containers for various applications.

For more information, see the NVIDIA Technical Blog.