NVIDIA’s Leap Into The Future: The Grace CPU Superchip
In the ever-evolving landscape of data centers, the demand for faster, more efficient processing power is relentless. NVIDIA, a titan in the field of accelerated computing, has once again pushed the boundaries with its latest innovation: the Grace CPU Superchip.
NVIDIA Grace™ is designed for a new type of data center—one that processes mountains of data to produce intelligence with maximum energy efficiency. These data centers run diverse workloads like AI, data analytics, hyperscale cloud applications, and high-performance computing (HPC). NVIDIA Grace delivers 2X the performance per watt, 2X the packaging density, and the highest memory bandwidth compared to today’s leading servers to meet the most demanding data center needs.
- High-performance CPU for data analytics, cloud, and HPC
- Superchip design with 144 Arm Neoverse V2 CPU cores with Scalable Vector Extensions (SVE2)
- NVIDIA Scalable Coherency Fabric with 3.2TB/s of bisection bandwidth
- World’s first LPDDR5X with error-correcting code (ECC) memory, 1TB/s total bandwidth
- 900 gigabyte per second (GB/s) coherent interface, 7X faster than PCIe Gen 5
- 2X the packaging density of DIMM-based solutions
- 2X the performance per watt of today’s leading CPU
- Runs all NVIDIA software stacks and platforms, including NVIDIA HPC SDK, NVIDIA AI, and NVIDIA Omniverse™.
Designed from the ground up, the Grace CPU Superchip is NVIDIA’s answer to the modern data center’s insatiable appetite for processing vast amounts of data to produce intelligence. It’s not just about speed; it’s about doing it with maximum energy efficiency. This superchip is poised to redefine the standards of performance and efficiency in today’s AI data centers, hyperscale cloud applications, and high-performance computing (HPC) environments.
At the heart of the Grace CPU Superchip lies the NVIDIA NVLink®-C2C technology, which interconnects two Grace CPU chips to deliver a staggering 144 cores and an unprecedented 1 terabyte per second (TB/s) of memory bandwidth. This is a game-changer for data analytics, cloud services, and HPC applications, offering 2X the performance per watt and 2X the packaging density compared to current leading servers.
The superchip’s architecture is a marvel, featuring 144 Arm Neoverse V2 CPU cores with Scalable Vector Extensions (SVE2), a NVIDIA Scalable Coherency Fabric with 3.2TB/s of bisection bandwidth, and the world’s first LPDDR5X with error-correcting code (ECC) memory. This results in a total bandwidth of 1TB/s and a 900 gigabyte per second (GB/s) coherent interface, which is 7X faster than PCIe Gen 5.
What sets the Grace CPU Superchip apart is its ability to run all NVIDIA software stacks and platforms seamlessly, including the NVIDIA HPC SDK, NVIDIA AI, and NVIDIA Omniverse™. This ensures that developers and data centers can leverage the full potential of NVIDIA’s ecosystem to meet their most demanding needs.
As we look to the future, the Grace CPU Superchip stands as a testament to NVIDIA’s commitment to innovation and its role in shaping the next generation of data centers. With its unparalleled performance, efficiency, and scalability, the Grace CPU Superchip is not just a leap forward; it’s a giant stride into the future of computing.
