NVIDIA Announces Its First CPU Codenamed Grace, Based on ARM Architecture & Neoverse Cores

NVIDIA has announced its first CPU codenamed Grace which is designed for the modern data center. The CPU is named after computer scientist Grace Hopper who was one of the pioneers of computer science and one of the first programmers of Harvard Mark 1 and inventor of the first linkers.

The Grace CPU is NVIDIA's first design featuring the next-generation Neoverse cores which have been under development for years. The whole SOC is a combination of several chips which come with three main components, the CPU, the GPU, and the memory/IO subsystems. NVIDIA hasn't revealed many details but the company did state that the Grace CPU will be able to deliver a score of 300 in SPECrate2017_Int_base.

NVIDIA states that its Grace is a highly specialized processor targeting workloads such as training next-generation NLP models that have more than 1 trillion parameters. When tightly coupled with NVIDIA GPUs, a Grace CPU-based system will deliver 10x faster performance than today’s state-of-the-art NVIDIA DGX-based systems, which run on x86 CPUs.

NVIDIA is introducing Grace as the volume of data and size of AI models are growing exponentially. Today’s largest AI models include billions of parameters and are doubling every two-and-a-half months. Training them requires a new CPU that can be tightly coupled with a GPU to eliminate system bottlenecks.

NVIDIA built Grace by leveraging the incredible flexibility of Arm’s data center architecture. By introducing a new server-class CPU, NVIDIA is advancing the goal of technology diversity in AI and HPC communities, where choice is key to delivering the innovation needed to solve the world’s most pressing problems.

Grace’s First Adopters Push Limits of Science and AI

CSCS and Los Alamos National Laboratory both plan to bring Grace-powered supercomputers, built by Hewlett Packard Enterprise, online in 2023.

“With an innovative balance of memory bandwidth and capacity, this next-generation system will shape our institution’s computing strategy,” said Thom Mason, director of the Los Alamos National Laboratory. “Thanks to NVIDIA’s new Grace CPU, we’ll be able to deliver advanced scientific research using high-fidelity 3D simulations and analytics with datasets that are larger than previously possible.”

Delivering Breakthrough Performance, Up To 10x Faster Than x86 Servers

Underlying Grace’s performance is fourth-generation NVIDIA NVLink interconnect technology, which provides a record 900 GB/s connection between Grace and NVIDIA GPUs to enable 30x higher aggregate bandwidth compared to today’s leading servers.

Grace will also utilize an innovative LPDDR5x memory subsystem that will deliver twice the bandwidth and 10x better energy efficiency compared with DDR4 memory. In addition, the new architecture provides unified cache coherence with a single memory address space, combining system and HBM GPU memory to simplify programmability.

Grace will be supported by the NVIDIA HPC software development kit and the full suite of CUDA and CUDA-X libraries, which accelerate more than 2,000 GPU applications, speeding discoveries for scientists and researchers working on the world’s most important challenges. Availability of Grace CPUs is expected in the beginning of 2023.