NVIDIA GTC SPRING EDITION 2022
NVIDIA GTC SPRING EDITION 2022
The developer conference for the AI age
Join us March 21-24, 2022 for NVIDIA's virtual GPU Technology Conference (GTC22). Traditionally, GTC kicks off with NVIDIA CEO Jensen Huang's keynote. This is one you don't want to miss.
Choose from over 500 conference sessions with over 900 personalities from AI, High Performance Computing and Graphics, among others, giving talks and web sessions to educate, share and keep up with the latest developments.
sysGen, as an NVIDIA Elite Partner, will be there to bring you the most interesting news directly from the keynote. Learn from us what will help you do your best work tomorrow and move your business forward.
Always discover something new at GTC for four days. Learn what's driving transformation in your industry - from AI power to the collaborative virtual worlds of NVIDIA Omniverse and beyond - and discover the technological advancements and groundbreaking research that can help solve the world's biggest challenges together.
We expect NVIDIA Enterprise and NVIDIA Omniverse to receive exciting updates that will enable the mid-market in particular to benefit from the use of AI, and NVIDIA CEO Huang is expected to unveil new technologies, products and collaborations with some of the world's leading companies. Just keep coming back for updates during GTC22.
Our favorites of the GTC22
The spring edition of the GTC was also in no way inferior to the previous ones in terms of exciting and outstanding announcements. Here we present our favorites and under GTC22 Highlights you will find all the announcements from the keynote summarized for you to read shortly.
Important updates for developers
using the NVIDIA Omniverse real-time 3D design collaboration platform
using the NVIDIA Omniverse real-time 3D design collaboration platform
Updates to the Omniverse apps Create, Machinima, and Showroom, as well as an upcoming release of View. Powered by GeForce RTX and NVIDIA RTX GPUs, they dramatically accelerate creative 3D workflows.
New Omniverse Connections expand the ecosystem and are now available in beta: Unreal Engine 5 Omniverse Connector and Adobe Substance 3D Material Extension, followed closely by Adobe Substance 3D Painter Omniverse Connector. Cinema 4D from Maxon now supports Universal Scene Description (USD). Unlocking Cinema 4D workflows via OmniDrive brings deeper integration and flexibility to the Omniverse ecosystem.Using the Hydra render delegate feature, artists can now use Pixar HDStorm, Chaos V-Ray, Maxon Redshift, and OTOY Octane Hydra render delegates in the viewport of all Omniverse apps, with Blender Cycles coming soon.
Whether refining 3D scenes or exporting finished projects, artists can switch between the lightning-fast Omniverse RTX Renderer and their preferred renderer, giving them the ultimate freedom to work the way they want.
New Omniverse Connections expand the ecosystem and are now available in beta: Unreal Engine 5 Omniverse Connector and Adobe Substance 3D Material Extension, followed closely by Adobe Substance 3D Painter Omniverse Connector. Cinema 4D from Maxon now supports Universal Scene Description (USD). Unlocking Cinema 4D workflows via OmniDrive brings deeper integration and flexibility to the Omniverse ecosystem.Using the Hydra render delegate feature, artists can now use Pixar HDStorm, Chaos V-Ray, Maxon Redshift, and OTOY Octane Hydra render delegates in the viewport of all Omniverse apps, with Blender Cycles coming soon.
Whether refining 3D scenes or exporting finished projects, artists can switch between the lightning-fast Omniverse RTX Renderer and their preferred renderer, giving them the ultimate freedom to work the way they want.
NVIDIA announces DGX H100 systems - the world's most advanced enterprise AI infrastructure
Combined with the new NVLink switch system, each DGX SuperPOD delivers 1 exaflops of AI performance; the new NVIDIA Eos supercomputer is expected to be the fastest AI system in the world.
the DGX H100 systems provide the scalability needed to meet the massive computational demands of large-scale language models, recommendation systems, health research, and climate science
.
With eight NVIDIA H100 GPUs per system, interconnected via NVIDIA NVLink®, each DGX H100 delivers 32 petaflops of AI performance at new FP8 precision - 6x more than the previous generation.
DGX H100 systems are the building blocks of the next generation of NVIDIA DGX POD™ and NVIDIA DGX SuperPOD™ AI infrastructure platforms. The latest DGX SuperPOD architecture features a new NVIDIA NVLink switch system that can connect up to 32 nodes with a total of 256 H100 GPUs.
the DGX H100 systems provide the scalability needed to meet the massive computational demands of large-scale language models, recommendation systems, health research, and climate science
.
With eight NVIDIA H100 GPUs per system, interconnected via NVIDIA NVLink®, each DGX H100 delivers 32 petaflops of AI performance at new FP8 precision - 6x more than the previous generation.
DGX H100 systems are the building blocks of the next generation of NVIDIA DGX POD™ and NVIDIA DGX SuperPOD™ AI infrastructure platforms. The latest DGX SuperPOD architecture features a new NVIDIA NVLink switch system that can connect up to 32 nodes with a total of 256 H100 GPUs.
New NVIDIA RTX GPUs open up endless creative possibilities
Professional users have new, powerful ones with faster speeds and larger memory options: RTX A5500, RTX A4500 , RTX A3000 12GB, RTX A2000 8GB and NVIDIA RTX A1000. These GPUs feature the latest RTX and Max-Q technology and deliver exceptional performance.
NVIDIA's new flagship the RTX A5500 with 16GB of memory, is capable of handling the most demanding 3D and video workloads with up to twice the rendering performance of the previous generation RTX 5000, perfectly suited for the most complex, advanced and creative workflows.
with the new NVIDIA RTX GPUs, artists can create photorealistic, immersive digital experiences, scientists can make the latest breakthrough discoveries, and engineers can develop innovative technologies that will take us into the future.
NVIDIA's new flagship the RTX A5500 with 16GB of memory, is capable of handling the most demanding 3D and video workloads with up to twice the rendering performance of the previous generation RTX 5000, perfectly suited for the most complex, advanced and creative workflows.
with the new NVIDIA RTX GPUs, artists can create photorealistic, immersive digital experiences, scientists can make the latest breakthrough discoveries, and engineers can develop innovative technologies that will take us into the future.
Jetson AGX Orin
Server-level AI performance in the periphery
get started with development with the NVIDIA® Jetson AGX Orin Developer Kit™. It includes a high-performance, energy-efficient Jetson AGX Orin module and can emulate the other Jetson Orin modules. With up to 275 TOPs to run the NVIDIA AI software stack, this developer kit lets you create advanced robotics and edge AI applications for manufacturing, logistics, retail, service, agriculture, smart city, healthcare, and life sciences.
get started with development with the NVIDIA® Jetson AGX Orin Developer Kit™. It includes a high-performance, energy-efficient Jetson AGX Orin module and can emulate the other Jetson Orin modules. With up to 275 TOPs to run the NVIDIA AI software stack, this developer kit lets you create advanced robotics and edge AI applications for manufacturing, logistics, retail, service, agriculture, smart city, healthcare, and life sciences.
NVIDIA launches data center-scale Omniverse computing system
for industrial digital twins
for industrial digital twins
NVIDIA OVX is specifically designed to run complex digital twin simulations that run within NVIDIA Omniverse™, a real-time physically accurate simulation and 3D design collaboration platform.
The OVX system combines high-performance GPU-accelerated compute, graphics, and AI power with high-speed memory access, low-latency networking, and precise timing to provide the performance needed to create digital twins with realistic accuracy. OVX is used to simulate complex digital twins for modeling entire buildings, factories, cities, and even the entire world.
oVX will enable designers, engineers and planners to create physically accurate digital twins of buildings or create huge, realistically simulated environments with precise time synchronization between the physical and virtual worlds. Companies can evaluate and test complex systems and processes with multiple autonomous systems interacting in the same space-time to optimize, expand or make factories and warehouses more efficient, or train robots and autonomous vehicles before deploying them in the real world.
The OVX system combines high-performance GPU-accelerated compute, graphics, and AI power with high-speed memory access, low-latency networking, and precise timing to provide the performance needed to create digital twins with realistic accuracy. OVX is used to simulate complex digital twins for modeling entire buildings, factories, cities, and even the entire world.
oVX will enable designers, engineers and planners to create physically accurate digital twins of buildings or create huge, realistically simulated environments with precise time synchronization between the physical and virtual worlds. Companies can evaluate and test complex systems and processes with multiple autonomous systems interacting in the same space-time to optimize, expand or make factories and warehouses more efficient, or train robots and autonomous vehicles before deploying them in the real world.
LEARN MORE
NVIDIA launches AI computing platform for medical devices and
computational sensor systems
computational sensor systems
Clara Holoscan MGX™, a platform for the medical device industry to develop and deploy real-time AI applications at the edge, specifically designed to meet required regulatory standards.
Clara Holoscan MGX extends the Clara Holoscan platform, providing an all-in-one medical-grade reference architecture and long-term software support to accelerate innovation in the medical device industry. Processing high-throughput data streams to produce real-time insights, it takes edge computing to a new level of sensor innovation. From robotic surgery to exploring new approaches in biology, surgeons and scientists need medical devices that evolve into continuous sensor systems to explore and treat disease.
Clara Holoscan MGX extends the Clara Holoscan platform, providing an all-in-one medical-grade reference architecture and long-term software support to accelerate innovation in the medical device industry. Processing high-throughput data streams to produce real-time insights, it takes edge computing to a new level of sensor innovation. From robotic surgery to exploring new approaches in biology, surgeons and scientists need medical devices that evolve into continuous sensor systems to explore and treat disease.
LEARN MORE
Interest aroused? Then please do not hesitate to contact us.
GTC22 Highlights
Here we keep you "up-to-date" with the announcements in the Keynote GTC22 Spring Edition.
Omniverse Cloud
Omniverse Cloud enables real-time collaboration between 3D designers using different applications and hardware from different locations around the world. Access to the Omniverse Cloud is also available through other endpoints via NVIDIA Geforce Now.
Jetson Orin Modules and Dev Kits
At GTC 2021, NVIDIA debuted the Jetson AGX Orin module and dev kit, the latest and most powerful GPU in the Jetson family to date. Jetson AGX Orin is ready for use in next-generation autonomous machines, robots, or drones used in retail, manufacturing, transportation, agriculture, smart city, aerospace, and more. Learn here more.
Spectrum-4
The NVIDIA Spectrum-4 ASIC and the SN5000 switch family form NVIDIA's most innovative Ethernet platform to date, based on a 5nm process and containing over 100 billion transistors. With an ASIC total bandwidth of 51.2 Tbps, enabling port speeds of 400 and 800 GbE, and adaptive routing protocols, Spectrum-4 delivers the performance and security required for state-of-the-art data centers. In addition, Spectrum-4 innovations accelerate and simplify applications across a wide range of industries, including cloud, storage, artificial intelligence, deep learning and 5G solutions. Spectrum-4 is also available on NVIDIA Air for simulating data center infrastructures down to the ASIC level.
Deep Learning
Triton Inference Server streamlines AI inference by enabling teams to deploy, run, and scale trained AI models from any framework on any GPU- or CPU-based infrastructure. It gives AI researchers and data scientists the freedom to choose the right framework for their projects without impacting production deployment. It also helps developers deploy high-performance inference in the cloud, on-premises, on edge and embedded devices
Accelerated Computing
- NVIDIA SDK's for accelerated computing of AI workloads.
- NVIDIA H100 with 80 trillion transistors, with a new Ampere architecture for smooth processes between data and applications and the first 5th generation PCI-E interface and much more.
- NVIDIA Grace Hopper chip-to-chip connection of CPU and GPU. 144 CPU cores. 1 terabyte of memory bandwidth. Only 500 watts.
LaunchPad
Accelerate your AI journey with NVIDIA LaunchPad.With NVIDIA LaunchPad, you get instant access to NVIDIA AI with free curated labs - from AI-powered chatbots with Trition Inference Server to image classification with TensorFlow and more.
Networking
- NVIDIA Switch Systems 20TB HMB3, 192 Teraflops
- Quantum-2 INFINIBAND 57 trillion transistors, 400GBps
- H100 CNX CX7 network and graphics card combined. Offloads workloads from the CPU.
Digital Twin
- With EARTH-2 Supercomuter, NVIDIA relies on AI to tackle climate change. The system creates a digital twin of the Earth in the Omniverse.
- FourCastNet for predicting climate change, using 10 TB of data from a 40-year collection.
Quantum Computing
- NVIDIA Hopper GPU Architecture accelerates dynamic programming by up to 40x with new DPX instructionsDynamic
programming algorithms are used in healthcare, robotics, quantum computing, data science and many other fields
. - NVIDIA unveils entry into hybrid quantum computing NVIDIA cuQuantum debuts growing ecosystem and collaboration to develop programming model for tomorrow's most powerful systems.
- Announcing NVIDIA Eos - the world's fastest AI supercomputerNVIDIA
will be the first company to build a DGX SuperPOD with the groundbreaking new AI architecture to support the work of NVIDIA researchers advancing climate science, digital biology and the future of AI. The "Eos" supercomputer is expected to be the fastest AI system in the world when it comes online later this year. It has a total of 576 DGX H100 systems with 4,608 DGX H100 GPUs. NVIDIA Eos is expected to offer AI computing power of 18.4 exaflops, making AI processing four times faster than the Japanese supercomputer Fugaku, which is currently the fastest system in the world. For traditional scientific computing, Eos is said to offer performance of 275 petaflops.
Healthcare
- Clara Holoscan leverages much of the real-time computing power used in DRIVE to support medical instrumentation and real-time sensors such as RF ultrasound, 4K surgical video, high-throughput cameras, and lasers. Holoscan development platforms are available to Early Access customers today, generally available in May, and ready for medical-grade deployment in the first quarter of 2023.
Artificial intelligence
- The NVIDIA AI Platform receives major updates, according to Huang, including the Triton Inference Server, the NeMo Megatron 0.9 framework for training large language models, and the Maxine framework for improving audio and video quality.
- The platform includes NVIDIA AI Enterprise 2.0, an end-to-end, cloud-native suite of AI and data analytics tools and frameworks optimized and certified by NVIDIA and now supported by all major data centers and cloud platforms.
- NVIDIA Omniverse software for creating digital twins and new NVIDIA OVX systems at data center scale will be critical to "actionable AI." "Omniverse is at the heart of NVIDIA's robotics platforms" will soon receive new releases and updates. In robotics and industrial automation, the importance of digital twins and Omniverse is great." OVX will use Omniverse's digital twins for large-scale simulations with multiple autonomous systems operating in the same space-time, Huang explained. The backbone of OVX is the network fabric, Huang said, announcing the NVIDIA Spectrum-4 high-performance data network infrastructure platform.
- NVIDIA Spectrum-4 is the world's first end-to-end 400 Gbps networking platform and consists of the Spectrum-4 switch family, NVIDIA ConnectX-7 SmartNIC, NVIDIA BlueField-3 DPU and NVIDIA DOCA Data Center Infrastructure software.
- To bring Omniverse to even more users, Huang announced Omniverse Cloud announced. With just a few clicks, employees can now connect to the cloud through Omniverse.
- The"next wave of AI": robots and autonomous vehicles. New silicon, software and simulation capabilities will unleash "the next wave of AI," Huang said, robots that can "plan and act."
NVIDIA Avatar, DRIVE, Metropolis, Isaac and Holoscan are robotics platforms built end-to-end and entirely on "four pillars": Data Generation, AI Model Training, Robotics Stack and Omniverse Digital Twins, Huang explained.
The NVIDIA DRIVE autonomous vehicle system is essentially an "AI chauffeur."
And Hyperion 8 - NVIDIA's self-driving car hardware architecture on which NVIDIA DRIVE is built - can enable full self-driving with a 360-degree camera, radar, lidar and ultrasonic sensors. - Huang announced that NVIDIA Orin, a centralized AV and AI computer that acts as the engine of the new generation of electric vehicles, robotaxis, shuttles and trucks, will begin shipping this month.
- NVIDIA's robotics platforms consist of Metropolis and Isaac - Metropolis is a stationary robot that tracks moving things, while Isaac is a platform for things that move, Huang explained. To help robots navigate indoors - such as factories and warehouses - NVIDIA announced Isaac Nova Orin, which builds on Jetson AGX Orin, a state-of-the-art compute and sensor reference platform to accelerate the development and deployment of autonomous mobile robots.
Data Center
- Transforming data centers into "AI Factories", NVIDIA CEO unveils new silicon - including new Hopper GPU architecture and new H100 GPU, new AI and accelerated computing software, and powerful new data center-scale systems. "Enterprises are processing and refining their data, making AI software, and becoming makers of intelligence." Omniverse will bring all of this together to accelerate collaboration between humans and AI, better model and understand the real world, and serve as a testing ground for new kinds of robots, "the next wave of AI."
- New Silicon - NVIDIA H100: , built on the new Hopper architecture is considered the "new engine of the world's AI infrastructures." "Hopper H100 is the biggest generational leap ever - 9x training performance compared to A100 and 30x inference throughput on large language models." H100 is in production and will be available in the third quarter, according to NVIDIA.
- Announcing the new Grace CPU Superchip an, NVIDIA's first discrete data center CPU for high-performance computing. It consists of two CPU chips connected via a 900 gigabyte per second NVLink chip-to-chip interconnect to form a 144-core CPU with 1 terabyte per second of memory bandwidth, making it the ideal CPU for the world's AI infrastructures.
- Announcing the new Hopper GPU-based AI supercomputer on - DGX H100, H100 DGX POD and DGX SuperPOD. To connect everything together, NVIDIA's new NVLink high-speed interconnect technology will be used in all future NVIDIA chips - CPUs, GPUs, DPUs and SOCs, Huang said.
For developers
- NVIDIA MAXINE Accelerated SDK with state-of-the-art AI capabilities for building virtual collaboration and content creation applications. Language barriers are overcome with MAXINE.
- NVIDIA Merlin™ is an open source framework for building powerful recommendation systems at scale.
- NeMo Megatron is a new capability in the NeMo framework that enables developers to effectively train and scale language models to billions of parameters. With NeMo Megatron, you can train different variants of GPT-3 models and scale them to multiple nodes on Superpods. This deep learning software stack is optimized for NVIDIA DGX A100 SuperPODs with NVIDIA InfiniBand to provide efficient on-premises compute performance for training and deriving complex workloads.
- The RAPIDS suite of software libraries, based on CUDA-X AI, gives you the freedom to run end-to-end pipelines for data science and analytics entirely on GPUs. It is based on NVIDIA® CUDA® primitives for low-level computational optimization, but makes GPU parallelism and high memory bandwidth accessible through easy-to-use Python interfaces.
- With GPU acceleration for data science pipelines in Apache Spark 3.0 - available without code changes - you can accelerate data processing and model training while significantly reducing infrastructure costs.
- NVIDIA cuOpt™ is an AI logistics software API that enables near real-time routing optimizations. cuOpt empowers logistics and operational research developers to leverage larger data sets and faster processing, delivering new capabilities like dynamic-rerouting, simulations, and sub-second solver response time for last-mile delivery, supply chain, warehouse picking, and food delivery.
- Deep Graph Library (DGL) is a frame-independent, easy-to-use, and scalable Python library for implementing and training Graph Neural Networks (GNN). Because DGL is framework-independent, it can be easily integrated into an existing PyTorch, TensorFlow, or Apache MXNet workflow.
- NVIDIA cuQuantum is an SDK with optimized libraries and tools to accelerate quantum computing workflows. With NVIDIA GPU Tensor Core GPUs, developers can use cuQuantum to accelerate quantum circuit simulations based on state vector and tensor network methods by orders of magnitude.
- The NVIDIA Aerial™ SDK helps developers create high-performance, software-defined, cloud-native 5G applications that are programmable, scalable, and energy-efficient.
- Sionne's AI framework for 6G research
- NVIDIA Modulus is a neural network framework that combines the power of physics in the form of partial differential equations (PDEs) with data to create parameterized surrogate models with near real-time latency. Whether you want to start with AI-driven physics problems or design digital twin models for complex nonlinear, multiphysics systems, NVIDIA Modulus can help you get the job done.
- Monai for Medical Research with Artificial Intelligence.
- NVIDIA FLARE™ (NVIDIA Federated Learning Application Runtime Environment) is a domain-independent, open source, and extensible SDK for federated learning. It enables researchers and data scientists to adapt existing ML/DL workflows to a federated paradigm and empowers platform developers to create a secure, privacy-compliant offering for distributed multi-participant collaboration.
- The NVIDIA DRIVE® family of products for autonomous vehicle development. It covers everything from the vehicle to the data center. DRIVE Hyperion is the solution for the car, a vehicle architecture that includes sensors, DRIVE AGX for computing power, and software tools required for robust self-driving and intelligent cockpit capabilities. In the data center, NVIDIA provides the hardware and software you need for AV development, including NVIDIA DGX™ to train DNN for perception and DRIVE Sim to generate datasets and validate the entire AV stack.
- NVIDIA Clara™ Holoscan is the AI computing platform for medical devices that combines hardware systems for low-latency sensor and network connectivity, optimized libraries for data processing and AI, and core microservices to run streaming, imaging, and other applications, from embedded to edge to cloud. Clara Holoscan enables medical device manufacturers to develop and bring to market AI-enabled next-generation medical devices.
- NVIDIA Isaac Develop and deploy commercial AI-powered robots. The NVIDIA Isaac SDK™ is a toolkit of building blocks and tools that accelerate the development of robots that require the advanced perception and navigation capabilities enabled by AI.
- NVIDIA Metropolis consists of an application framework, a set of developer tools, and a partner ecosystem that brings together visual data and AI to improve operational efficiency and safety across a variety of industries. it brings order to the data deluge of trillions of sensors for frictionless retail, efficient inventory management, traffic engineering in smart cities, visual inspection on factory floors, patient care in medical facilities, and more. With cutting-edge technology and an extensive developer ecosystem, enterprises can now build, deploy and scale AI and IoT applications - from the edge to the cloud.
Topic excerpt
GTC22 will focus on accelerated computing, deep learning, data science, digital twins, networking, quantum computing, and computing in the data center, in the cloud, and at the edge. Missed the keynote live? No problem, just watch the recorded keynote with us.
Review: Keynote March 22, 2022
Discover AI, Omniverse and more at the GTC Keynote
Learn how NVIDIA's new computing platforms, advances in robotics, and breakthroughs in virtual collaboration are transforming every industry
Our top news from the GTC22
NVIDIA AI Enterprise 2.0
NVIDIA AI Enterprise 2.0 brings enterprise-grade AI to Enterprise-grade for anyone, anywhere, on any platform. With the 2.0 update, we can again expect some improvements.
AI Enterprise
VMware + NVIDIA AI-enabled platform
InfoNVIDIA AI Enterprise is an end-to-end, Cloud-native AI and data analytics software suite optimized, certified and supported by NVIDIA, certified and supported to run on VMware vSphere with NVIDIA Certified Systems™. It includes key NVIDIA technologies for rapid provisioning, management and scaling of Scale AI workloads in the modern hybrid cloud. For Everyone, but of particular interest to the mid-market.
NVIDIA Jetson AGX Orin Module and Dev Kit: the best AI performance of Jetson Family
At GTC 2021, NVIDIA debuted the Jetson AGX Orin Module and dev kit, the latest and most powerful GPU to date of the Jetson family. Jetson AGX Orin is ready for use in autonomous Machines, robots or next-generation drones used in retail, in manufacturing, transportation, agriculture, smart city, aerospace, etc. Update GTC 2022: available in Q2/2022.
NVIDIA Fleet Command
Deploy AI from any location to any destination.
NVIDIA Fleet Command is a managed cloud service that securely deploys, manages, and scales AI applications across the distributed edge infrastructure.
NVIDIA LaunchPad
Accelerate your AI journey with NVIDIA LaunchPad.
With NVIDIA LaunchPad, you get instant access to NVIDIA AI with free curated labs - from AI-powered chatbots with Trition Inference Server to image classification with TensorFlow and more.
NVIDIA Spectrum-4
The NVIDIA Spectrum-4 ASIC and the SN5000 Switch family form NVIDIA's most innovative Ethernet platform to date, based on a 5nm process and contains over 100 billion transistors transistors. With a total ASIC bandwidth of 51.2 Tbps, enabling port speeds of of 400 and 800 GbE, and adaptive routing protocols, Spectrum-4 provides Spectrum-4 provides the performance and security required for state-of-the-art data center required. In addition, Spectrum-4 innovations accelerate and simplify Spectrum-4 innovations accelerate and simplify applications across a wide range of industries, including Cloud, Storage, Artificial Intelligence, Deep Learning and 5G solutions. Spectrum-4 is also available on NVIDIA Air for simulation of Data center infrastructures down to the ASIC level.