Nvidia is planning a high performance processor for a new generation of AI supercomputers. The chip, codenamed Grace, will use ARM’s second generation Neoverse processor core with Nvidia’s NV Link interconnect and low power LPDDR5 memory and is scheduled to ship in 2023.
The chip will be used for the Alps supercomputer at the Swiss National Computing Centre (CSCS) as a key tool for science throughout Europe when it comes online in the same year.
“We are thrilled to announce the Swiss National Supercomputing Centre will build a supercomputer powered by Grace and our next-generation GPU,” said Jensen Huang, CEO of Nvidia. “Using licensed Arm IP, Nvidia has designed Grace as a CPU specifically for giant-scale AI and HPC. Coupled with the GPU and DPU, Grace gives us the third foundational technology for computing, and the ability to re-architect the data centre to advance AI. Nvidia is now a three-chip company.”
ARM's Neoverse Z1 (Zeus) and N2 (Perseus) cores aim to provide data centre and supercomputer chip designers with higher performance on a 5nm process technology. The architectures will be detailed later this month in eeNews Europe.
Alps will be built by Hewlett Packard Enterprise using the new HPE Cray EX supercomputer product line as well as the NVIDIA HGX supercomputing platform, including NVIDIA GPUs and the NVIDIA HPC SDK as well as the new Grace CPU.
The Alps system will replace CSCS’s existing Piz Daint supercomputer which is one of the world’ s most power efficient supercomputers.
“Deep learning is just an incredibly powerful set of tools that we add to the toolbox,” said CSCS Director Prof Thomas Schulthess. ”Nvidia’s novel Grace CPU allows us to converge AI technologies and classic supercomputing for solving some of the hardest problems in computational science. We are excited to make the new CPU available for our users in