Starting with the Grace CPU Superchip, it is a datacenter processor that NVIDIA claims will be the fastest CPU available when it launches in early 2023. It is an ARM processor with 144 cores that will directly rival AMD's EPYC CPUs. It runs on the 4N – 4nm-based process – from TSMC.
“A new type of data center has emerged — AI factories that process and refine mountains of data to produce intelligence,” said Jensen Huang, founder and CEO of NVIDIA. “The Grace CPU Superchip offers the highest performance, memory bandwidth and NVIDIA software platforms in one chip and will shine as the CPU of the world's AI infrastructure.”
As for the Hopper H100 GPU, it comes with 80 billion transistors and will work at supercomputer level AI computing grades. To achieve this, the chip supports 3 TB/s HBM3 bandwidth and 4.9 TB/s external bandwidth.
“Confidential Computing — H100 is the world's first accelerator with confidential computing capabilities to protect AI models and customer data while they are being processed. Customers can also apply confidential computing to federated learning for privacy-sensitive industries like healthcare and financial services, as well as on shared cloud infrastructures.”
Speaking of Confidential Computing, NVIDIA teamed with Microsoft to announce this week Confidential Computing GPUs for Microsoft Azure.
Creating a confidential computing environment relies on enabling trusted execution environments (TEEs) to drive data protection at a CPU-level. One of the problems with TEEs in their current form is that they are built in CPUs. Microsoft wants to extend their capabilities to GPUs and to make that happen is collaborating with NVIDIA
Tip of the day: Did you know that Windows now has a package manager similar to Linux called “Winget”? In our tutorial, we show you how to install and use this new tool that allows the quick installation of apps via PowerShell or a GUI.