GPU powerhouse is seeking to muscle into the terriority of Intel and AMD, with the news it is developing an ARM-based server chip.
The new chip will be called ‘Grace’, and is Nvidia’s first data centre CPU that promises to deliver “10x the performance of today’s fastest servers on the most complex AI and high performance computing workloads.”
Nvidia of course already has a presence inside many data centres and supercomputers thanks to its GPU cards. In February it announced its new GeForce RTX 3060 card had reduced efficiency to mine the crypto-currency Ethereum.
That was a deliberate attempt to keep the card from being snapped by cryptocurrency miners, amid the global shortages of GPU cards for gamers and digital artists.
Grace chip
Nvidia says the ‘Grace’ CPU will be able to deliver a 10x performance leap, thanks to use of energy-efficient ARM CPU cores “with an innovative low-power memory subsystem to deliver high performance with great efficiency.”
It is being designed for the “fastest servers on the most complex AI and high performance computing workloads.”
“Leading-edge AI and data science are pushing today’s computer architecture beyond its limits – processing unthinkable amounts of data,” said Jensen Huang, founder and CEO of Nvidia.
Nvidia chief executive Jensen Huang at Siggraph. Credit: Nvidia
“Using licensed ARM IP, Nvidia has designed Grace as a CPU specifically for giant-scale AI and HPC,” he said. “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.”
The firm said that Grace is a highly specialised 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’s Grace chip has been named for Grace Hopper, the US computer-programming pioneer, and the firm accepts it may well serve a niche segment of computing.
First customers
The Swiss National Supercomputing Centre (CSCS) and the US Department of Energy’s Los Alamos National Laboratory are the first to announce plans to build Grace-powered supercomputers in support of national scientific research efforts.
“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 centre 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 said that 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.
Availability is expected in the beginning of 2023.