At the most exciting and content-rich GTC yet, NVIDIA announced three new processors built around Arm IP.
The first, NVIDIA Grace, is our first data center CPU, which will take giant AI and high-performance computing (HPC) workloads to new levels.
Separately, the new NVIDIA® BlueField®-3 data processing unit (DPU), sporting Arm® cores, was also announced, opening doors to new, more powerful applications in data center networking.
NVIDIA DRIVE™ Atlan becomes our first processor for autonomous vehicles packing an Arm-enabled DPU, anticipating the need for processing speed complimented by high-performance networking as automakers push to SAE level 5.
We also announced several advances with partners, from cloud to laptops to wireless 5G technologies.
NVIDA Grace—Purpose-Built to Process the World’s Largest Models
At GTC21, we unveiled NVIDIA Grace, our first data center CPU.
Based on a future Arm® Neoverse™ architecture, Grace is a highly specialized processor targeting the largest, most data-intensive HPC and AI applications, such as the training of next-generation natural language processing models. When tightly coupled with NVIDIA GPUs, Grace will deliver 10X the performance of today’s fastest servers on the most complex AI and HPC workloads.
The Swiss National Supercomputing Centre (CSCS) and the U.S. 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.
CSCS’s new system will use Grace to enable breakthrough research in a wide range of fields.
“We are pleased and excited to be on this journey with NVIDIA. Expanding the laboratory’s compute capability with this year’s NVIDIA A100 [Tensor Core GPU] purchase allows our researchers and engineers to grow their skills and puts us in a good position to fully engage the Grace-based system in 2023,” said John Sarrao, deputy director for science, technology, and engineering at Los Alamos.
The performance delivered by today’s general-purpose CPUs is no longer sufficient to keep up with the exploding volumes of data being transferred, processed, and stored across data center networks. The newly announced BlueField-3 DPU packs 400 gigabits-per-second (Gbps) links and 5X the Arm compute power of the current BlueField-2 DPU.
Accelerating Arm Growth
NVIDIA’s collaboration with Arm extends well beyond hardware. The biggest challenge for a new architecture is to reach critical mass of applications that run well there. So, NVIDIA is making good on the commitment to elevate Arm to a first-class CPU architecture for HPC and AI with the new NVIDIA Arm HPC Developer Kit. The kit leverages Ampere Altra 80-core, Neoverse-based CPUs, plus dual NVIDIA A100 GPUs, a pair of NVIDIA BlueField-2 DPUs, and the NVIDIA HPC SDK, delivering an environment ready to compile and test applications.
In the cloud, we’re working with Amazon Web Services to offer NVIDIA GPUs to AWS EC2 instances based on the AWS Graviton2. We’re also working with Ampere Computing and Oracle to introduce NVIDIA GPU-powered instances to the Oracle Cloud Infrastructure running on Ampere Altra CPUs.
Arm cores will debut in next-generation, AI-enabled, autonomous vehicles powered by NVIDIA DRIVE Atlan, kicking out more than 1,000 trillion operations per second. Atlan marks the first time the DRIVE platform integrates a DPU, carrying Arm cores that will help it pack the equivalent of data center networking into autonomous vehicles.
And finally, at the edge, we’re working with MediaTek, the world’s largest supplier of smartphone chips, to create a new class of notebooks powered by an Arm-based CPU alongside an NVIDIA RTX™ GPU. We’re also partnering with Marvell Semiconductor to team its OCTEON Arm-based processors with NVIDIA GPUs.
The intersection of NVIDIA-powered AI solutions running on Arm-based hardware is just getting started, but we’re already witnessing its enormous potential for use in a wider market. We look forward to continued collaboration with Arm on bringing exciting new data center, edge, and IoT solutions to the world.
© 2021 NVIDIA Corporation. All rights reserved. NVIDIA, the NVIDIA logo, BlueField, DRIVE, and RTX are trademarks and/or registered trademarks of NVIDIA Corporation in the U.S. and other countries. Other company and product names may be trademarks of the respective companies with which they are associated. All other trademarks are property of their respective owners.
ARM, AMBA and ARM Powered are registered trademarks of ARM Limited. Cortex, MPCore and Mali are trademarks of ARM Limited. “ARM” is used to represent ARM Holdings plc; its operating company ARM Limited; and the regional subsidiaries ARM Inc.; ARM KK; ARM Korea Limited.; ARM Taiwan Limited; ARM France SAS; ARM Consulting (Shanghai) Co. Ltd.; ARM Germany GmbH; ARM Embedded Technologies Pvt. Ltd.; ARM Norway, AS and ARM Sweden AB.