NVIDIA Announces New ARM Based CPU

NVIDIA has unveiled a new CPU based on the ARM architecture. It will power data centers to deliver 10x the performance of conventional servers on the most complex AI and high performance computing workloads. Named after Grace Hopper, the U.S. computer-programming pioneer, it’s designed to handle the world’s most demanding computational tasks.

This includes natural language processing, recommender systems and AI supercomputing. These tasks require the need to analyze vast datasets requiring both ultra-fast compute performance and massive memory. Grace uses energy-efficient ARM CPU cores and a low-power memory subsystem.

“As the world’s most widely licensed processor architecture, Arm drives innovation in incredible new ways every day,” said Arm CEO Simon Segars. “NVIDIA’s introduction of the Grace data center CPU illustrates clearly how Arm’s licensing model enables an important invention, one that will further support the incredible work of AI researchers and scientists everywhere.”

NVIDIA is taking advantage of ARM’s flexible data center architecture with the creation of its Grace CPU. This is serves as a testament to how capable the platform can be, and why many including Apple have flocked toward it by abandoning conventional x86 CPUs from the likes of Intel.

Todays largest AI models include billions of parameters and are doubling every two-and-a-half months. NIVIDIA belives that the best way to train them is by combining a new CPU that is coupled with a GPU in order to eliminate system bottlenecks.

When coupled with NVIDIA GPUS, Grace can deliver 10x faster performance than today’s state-of-the-art NVIDIA systems, which run on x86 CPUs. The ARM based CPU will utilize LPDDR5x memory subsystem, which delivers twice as much bandwidth and 10X better energy efficiency compared to current DDR4 memory technology.

Grace will have support from NVIDIA HPC software development kit and the full suite of CUDA® and CUDA-X™ libraries. They are responsible for accelerated computation of more than 2,000 GPU applications, which power the work of scientists and researchers working on the world’s most complex problems.

Hewlett Packard Enterprise will responsible for building the first Grace powered servers for the The Swiss National Supercomputing Centre (CSCS) and the U.S. Department of Energy’s Los Alamos National Laboratory. These new supercomputers will be used to support national scientific research efforts. Both systems are expected to be online by 2023.