NVIDIA Grace Hopper superchip designed to accelerate generative AI fully put into production

The NVIDIA Grace Hopper superchip, designed to accelerate generative AI, has been put into full production

Systems powered by GH200 will be integrated into over 400 system configurations from global system manufacturers based on NVIDIA's Grace, Hopper, and Ada Lovelace architecture.

COMPUTEX - May 29, 2023 - NVIDIA announced today that the NVIDIA® GH200 Grace Hopper superchip is now in full production, providing the power needed for complex AI and HPC workloads for systems launching worldwide.

These GH200-powered systems will be integrated into over 400 system configurations from global system manufacturers based on different combinations of NVIDIA's latest CPU, GPU and DPU architectures such as NVIDIA Grace™, NVIDIA Hopper™, NVIDIA Ada Lovelace and NVIDIA BlueField®, to meet the surging demand for generative AI in the market.

At COMPUTEX, NVIDIA founder and CEO Jensen Huang announced new systems, partners and more details about the GH200 Grace Hopper superchip. The product uses NVIDIA NVLink®-C2C interconnect technology to connect Arm-based NVIDIA Grace CPU and Hopper GPU architectures, achieving up to 900GB/s of total bandwidth, seven times higher than standard PCIe Gen5 channels in traditional acceleration systems, providing astonishing computing power to meet the most demanding generative AI and HPC applications.

NVIDIA's Vice President of Accelerated Computing, Ian Buck, said, "Generative AI is rapidly driving business transformation, unlocking new opportunities in industries such as healthcare, finance, and business services, and accelerating R&D in these industries. With the full production of the Grace Hopper superchip, manufacturers worldwide will soon provide the acceleration infrastructure needed for businesses to build and deploy generative AI applications using proprietary data."
Global hyperscale cloud service providers and supercomputing centers in Europe and the United States will be among the customers accessing GH200-powered systems.

Hundreds of acceleration systems and cloud instances

Numerous system manufacturers worldwide are launching various systems with different combinations of NVIDIA accelerators and processors. Taiwanese manufacturers including AAEON, Advantech, Aetina, ASRock Rack, ASUS, GIGABYTE, Ingrasys, Inventec, Pegatron, QCT, Tyan, Wistron and Wiwynn were mentioned as major partners in NVIDIA founder and CEO Jensen Huang's COMPUTEX keynote speech today.
In addition, global server manufacturers such as Cisco, Dell Technologies, Hewlett Packard Enterprise, Lenovo, and Eviden, a subsidiary of Atos Group, are providing various NVIDIA acceleration systems.
NVIDIA's cloud partners for the H100 include AWS, Cirrascale, CoreWeave, Google Cloud, Lambda, Microsoft Azure, Oracle Cloud Infrastructure, Paperspace, and Vultr.
The NVIDIA L4 GPU is now available on Google Cloud.

Cross-Accelerator Full-Stack Computing

The upcoming series of products accelerated by NVIDIA's Grace, Hopper, and Ada Lovelace architectures provide broad support for NVIDIA's software stack, including NVIDIA AI, NVIDIA Omniverse™ platform, and NVIDIA RTX™ technology.
NVIDIA AI Enterprise, as the software layer of NVIDIA AI platform, provides over 100 frameworks, pre-trained models, and development tools to simplify the development and deployment of production AI, including generative AI, computer vision, and speech AI.
The NVIDIA Omniverse development platform is used to build and run metaverse applications, enabling individuals and teams to work across multiple software suites and collaborate in real-time in a shared environment. The platform is based on the Universal Scene Description framework, an open and scalable 3D language for virtual worlds.
The NVIDIA RTX platform combines ray tracing, deep learning, and rasterization technologies and fundamentally changes the creative process for content creators and developers by supporting industry-leading tools and APIs. Applications built on the RTX platform bring real-time photorealistic rendering and AI-enhanced graphics, video, and image processing capabilities, enabling millions of designers and artists to create their best work.