Home Tech Updates NVIDIA Announces a Series of New Innovations at Computex 2022

NVIDIA Announces a Series of New Innovations at Computex 2022

by Helen J. Wolf
0 comment

Image, clockwise, from the top, left: NVIDIA VP for Accelerated Computing Ian Buck, Senior VP for Hardware Engineering Brian Kelleher, Director of Product Management for Accelerated Computing Ying Yin Shih, CTO Michael Kagan, Senior VP for GeForce Jeff Fisher, VP of Embedded and edge computing Deepu Talla.

NVIDIA announced its latest innovations in the data center, robotics, content creation, and gaming in a virtual keynote address on the opening day of Computex 2022 in Taipei.

Six NVIDIA leaders teamed up to deliver the keynote address, which covered advancements from robotics to AI and silicon to software and highlighted the work of the partner ecosystem.

NVIDIA Announces a Series of New Innovations at Computex 2022

Transforming data centers into AI factories

NVIDIA VP for Hyperscale and HPC Ian Buck started the keynote by explaining how data centers are turning into AI factories.

“This transformation requires us to reshape the data center at every level, from hardware to software, from chips to infrastructure to systems,” he said.

This will create huge business opportunities for NVIDIA’s partners in the data centers, HPC, digital twins, and cloud-based gaming, representing a “half-trillion market opportunity”.

Powering these modern AI factories requires end-to-end innovation at every level. As data centers become AI factories, data processing is essential.

These include NVIDIA Hopper GPUs, NVIDIA Grace CPUs, and NVIDIA BlueField DPUs as the building blocks interconnected by NVIDIA Quantum and Spectrum switches.

“The Bluefield DPU, along with the Quantum and Spectrum network switches, forms the infrastructure platform for the AI ​​factory of the future,” said NVIDIA CTO Michael Kagan.

NVIDIA technologies will be used in various server designs, including NVIDIA CGX for cloud gaming, OVX for digital twins, and HGX Grace and HGX Grace Hopper for simulation, data analytics, and AI.

NVIDIA has announced that the first wave of systems powered by the NVIDIA Grace CPU Superchip and Grace Hopper Superchip is expected in the first half of 2023.

“Grace will be great in AI, data analytics, scientific and hyperscale computing. And, of course, the entire suite of NVIDIA software platforms will run on Grace,” said Brian Kelleher, NVIDIA senior VP of hardware engineering.

Grace-powered systems from ASUS, Foxconn Industrial Internet, GIGABYTE, QCT, Supermicro, and Wiwynn will join x86 and other Arm-based servers to provide customers with various choices.

“All of these servers are optimized for NVIDIA-accelerated computer software stacks and qualify as part of our range of NVIDIA-certified systems,” said Ying Yin Shih, Director of Product Management for Accelerated Computing at NVIDIA.

liquid cooling

To empower businesses to deploy green data centers, NVIDIA also announced its first data center PCIe GPU with direct chip liquid cooling.

The liquid-cooled A100 PCIe GPUs will be supported by at least a dozen system builders on mainstream servers, with the first delivery in the third quarter of this year.

“All of these combined to deliver the future data center infrastructure that can handle these massive workloads,” Buck said. NVIDIA AI Enterprise software is needed to make all of these work seamlessly, which provides a robust 24/7 AI deployment. When it comes to reshaping the data center, NVIDIA has the complete, open platform of hardware and software for the AI ​​factories of the future,” he added.

A revolution in robotics with AI

NVIDIA VP of Embedded and Edge Computing Deepu Talla spoke about how the global push for automation is making robotics an important new application for AI.

More than 30 leading NVIDIA partners worldwide will offer the first wave of NVIDIA Jetson AGX Orin-powered production systems at Computex. New products from a dozen Taiwanese camera, sensor, and hardware suppliers will be used in edge AI, IoT, robotics, and embedded applications.

“We are entering the era of robotics — autonomous machines that are well aware of their environment and can make smart decisions about their actions,” Talla said.

Available since March, the NVIDIA Jetson AGX Orin developer kit delivers 275 trillion operations per second, with over 8x the processing power of its predecessor, NVIDIA AGX Xavier, in the same pin-compatible form factor.

Jetson Orin features the NVIDIA Ampere architecture GPU, Arm Cortex-A78AE CPUs, next-generation deep learning and vision accelerators, fast interfaces, faster memory bandwidth, and multimodal sensor support that can feed multiple concurrent AI application pipelines.

New Jetson AGX Orin production modules, which provide server-class performance for edge AI, will be available in July, while Orin NX modules will appear in September.

“Such modules are key to embedding smarter devices into the world around us,” says Talla.

The first pillar is about creating the AI, “a very time-consuming and difficult process that we make fast and easy,” Talla said, highlighting how tools like the Isaac Replicator for generating synthetic data, pre-trained NVIDIA models that available on NGC, and the NVIDIA TAO toolkit address this challenge.

The second pillar simulates how the robot works in the virtual world before deploying it in the real world with Isaac Sim.

The third is building physical robots.

And the fourth is about managing the fleet of robots throughout their lives, usually many years, if not more than a decade.

Part of that is Isaac Nova Orin, a reference design for state-of-the-art computers and sensors for autonomous mobile robots (AMR) – packed with technologies such as DeepMap, CuOpt, and Metropolis.

“This is the industry’s most comprehensive end-to-end robotics platform, and we will continue to invest in it,” said Talla.

NVIDIA senior VP for GeForce, Jeff Fisher, explained how NVIDIA works to deliver innovation to gamers and content creators.

“Over the past 20 years, NVIDIA and its partners have been committed to building the best platform for gaming and creation. Hundreds of millions now rely on it to play, work and learn,” he said.

Introduced in 2018, NVIDIA RTX has reinvented graphics thanks to advanced features like real-time ray tracing, and the momentum around it continues to grow. Over 250 RTX-compatible games and applications have doubled since the last Computex.

NVIDIA DLSS continues to set the standard for super-resolution with best-in-class performance and image quality and is now integrated into more than 180 games and applications.

DLSS is in the games gamers want to play, with 12 new games added to the ever-expanding library.

Developers of the critically acclaimed Hitman 3 are adding NVIDIA DLSS on May 24, along with ray-traced reflections and ray-traced shadows.

In addition, NVIDIA Reflex is now supported in 38 games, 22 screens, and 45 mice. With over 20 million gamers playing with Reflex ON monthly, Reflex has become one of NVIDIA’s most successful technologies.

The Reflex ecosystem continues to grow: ASUS introduced the world’s first 500Hz G-SYNC display, the ASUS ROG Swift 500Hz gaming monitor. Acer also launched the Predator X28 G-SYNC display. Meanwhile, Cooler Master introduced the MM310 and MM730 gaming mice with Reflex.

NVIDIA Studio, the RTX-powered platform that includes dozens of SDKs and accelerates the best creative apps and tools, and NVIDIA Omniverse, the enterprise platform for building interconnected 3D virtual worlds, are designed to enable collaboration and construction of these virtual worlds. To make.” said Fisher.

Gaming laptops remain the fastest-growing PC category, and the 4th Gen Max-Q Technologies — the latest iteration of NVIDIA’s thin-and-light laptop design — delivers a new level of power efficiency. GeForce RTX laptop models now total over 180.

These powerful systems are used to build massive, interconnected 3D destinations.

“These are our most portable, powerful laptops ever,” says Fisher.

You may also like