News Categories

NVIDIA's Omniverse gains traction and debuts second-gen OVX server for awesome digital twins simulation

By Vijay Anand - on 20 Sep 2022, 11:31pm

NVIDIA's Omniverse gains traction and debuts second-gen OVX server for awesome digital twins simulation

NVIDIA's full-service stack for Omniverse Cloud is now out.

By now, Omniverse is a critical piece of NVIDIA’s developer story at every NVIDIA GTC and who doesn’t deny that digital twins is fast becoming a valuable tool? Companies can simulate their entire factory operations virtually to fine-tune, improve processes and plan suitable investments to grow, predict weather behaviour for city-scale planning, and even co-create virtual worlds in the metaverse.


Omniverse momentum: Upsizing railway operations

Among the latest adopters of Omniverse include Digitale Schiene Deutschland (DSD), part of Germany’s national railway operator Deutsche Bahn, who is working with NVIDIA to increase their train network’s capacity without building new tracks and to optimally steer trains through its network. To do this, DSD is building the first country-scale digital twin to fully simulate automatic train operation across an entire network. That means creating a photorealistic and physically accurate emulation of the entire rail system. Using the AI-enabled digital twin created with NVIDIA Omniverse and feeding it with real-world data stream updates such as station platform measurements and vehicle sensor data, DSD can develop highly capable perception and incident prevention and management systems to optimally detect and react to irregular situations during day-to-day railway operation.

It is expected that future computer vision-powered systems could continually perform route observation and incident recognition, automatically warning of and reacting to potential hazards.


Digital Twins momentum: Enabling telcos to to plan, build and operate true 5G networks

Heavy.AI builds the world's first digital twin framework on Omniverse that is expressly targeted for telco usage.

Even telcos are getting a big boost for 5G network planning through Heavy.AI’s digital twin solution, HeavyRF, an industry first to boost the future of communications. HeavyRF simulates city-scale deployments faster and in a more energy-efficient manner by optimizing cellular tower and base station placements for the best possible coverage.

5G infrastructure is fundamentally very different from 4G infrastructure: It requires far more towers and antennas to provide the highest level of coverage and hence customer satisfaction. Because of that heightened demand, telcos can’t spend the same amount of time planning and locating each piece of 5G infrastructure. --- Jon Kondo, CEO of HEAVY.AI

According to consultancy and researcher firm Analysys Mason, 5G deployments require up to three times more base stations than 4G due to the higher frequency spectrum bands. As a result, 5G sites could consume up to three times more power and would cost four times more than an LTE one if deployed in the same manner. This is why virtualizing and planning the needs through accurate Digital Twins will help telcos deliver the next level of performance with the right level of infrastructure deployment and manage operational costs by tightly integrating key business information such as mobility and parcels data, and customer experience data within RF planning workflows.

Using an RF-synchronized digital twin would enable telcos to optimize capacity and coverage, plus interactively see how changes in deployment patterns translate into customer acquisition and retention at the household level.  The goal is to use machine learning and big data pipelines to continuously mirror existing real-world conditions.

The digital twin will use the parallel computing capabilities of modern GPUs for visual simulation, as well as to generate physical simulations of RF signals using real-time RTX ray tracing, powered by NVIDIA Omniverse’s RTX Renderer.


Omniverse Cloud is now in full swing

A sample of the Enter the Omniverse Cloud in action. (Image source: NVIDIA)

Announced back in March for early access use, Omniverse Cloud is now available with a full suite of cloud services for artists, developers and enterprise teams to design, share and collaborate 3d works, models, simulations and even build metaverse applications anywhere.

Using Omniverse Cloud, individuals and teams can experience in one click the ability to design and collaborate on 3D workflows without the need for any local computing power that was previously a prerequisite. Previously, you needed an NVIDIA Studio Laptop or NVIDIA RTX Workstation or an NVIDIA-certified Server to engage in the Omniverse. While it’s still the best way to experience and unleash your creative prowess without restraint, the Omniverse Cloud frees up the burden of having certified hardware grades to collaborate in the Omniverse and leverages the NVIDIA GeForce Now platform powered by NVIDIA RTX GPUs in the cloud.

Now roboticists can train, simulate, test and deploy AI-enabled intelligent machines with increased scalability and accessibility, while autonomous vehicle engineers can generate physically based sensor data and simulate traffic scenarios to test a variety of road and weather conditions for safe self-driving deployment, all in the cloud.

RIMAC Group's Nevera 3D car configurator in use over the Omniverse Cloud. (Image source: NVIDIA)

Early supporters of Omniverse Cloud include RIMAC Group, WPP and Siemens where the first two are vesting in end-to-end design to marketing aspects of the automotive pipeline and deliver advanced 3D content and experiences to leading automotive brands. Meanwhile, Siemens is helping businesses roll out smarter and faster digital transformation solutions via their Simens Xcelerator by leveraging NVIDIA’s Omniverse Computing Systems (OVX) and Omniverse Cloud.

Omniverse Cloud services now include the following:-

  • Omniverse Nucleus Cloud – provides creators with the ability to collaborate, share and edit 3D scene data in live view. Learn more about the role of the Omniverse Nucleus Server and Nucleus Cloud.
  • Omniverse App Streaming – enables users without RTX GPU hardware to stream Omniverse Create, an app for technical designers, artists and creators to interactively build 3D worlds in real time; and View, an app for non-technical users to view Omniverse content for approvals.
  • Omniverse Replicator – helps researchers and developers generate massive amounts of synthetic data to train AI networks and test digital twin environments. Read more about Omniverse Replicator’s role here.
  • Omniverse Farm – to harness multiple cloud instances and scale out Omniverse tasks such as rendering or synthetic data generation.
  • NVIDIA Isaac Sim and NVIDIA Drive Sim

Omniverse Farm, Replicator and Isaac Sim containers are available today on NVIDIA NGC (read more about NVIDIA GPU Cloud here) for self-service deployment on AWS using Amazon EC2 G5 instances featuring NVIDIA A10G Tensor Core GPUs. In addition, Omniverse Cloud will be available as NVIDIA managed services via early access by application.


Building large-scale Omniverse Digital Twins with second-gen NVIDIA OVX servers

Meet the NVIDIA OVX 2, the second-gen NVIDIA Omniverse Computing System for the data centre.

To recap, the NVIDIA Omniverse Computing System or NVIDIA OVX for short, is purpose-built for data-centre scale Omniverse applications like industrial digital twin production and simulation needs. First launched in March this year, these will be delivered to leading design and engineering teams at companies like BMW and Jaguar Land Rover.

To anticipate even more complex digital twin simulation needs, NVIDIA is getting ready to power newer OVX systems with NVIDIA’s new L40 GPU, which is based on the next-generation Ada Lovelace RTX GPU that's based on the Hopper H100 GPU architecture. Unlike the H100 that’s data-centre oriented and lacks any Ray Tracing Cores, the L40 Ada Lovelace GPU will have third-gen RT Cores to accelerate ray-traced and path-traced rendering of materials to physically and accurately simulate photo-realistic 3D synthetic data.

The new Ada Lovelace GPU architecture (Image source: NVIDIA). Read more about it our linked feature article.

The newer OVX server will also be paired with an even faster HPC networking chip, the ConnectX-7. This provides double the data rate of ConnectX-6 featured in the previous OVX server for 100/200/400G networking and instantly doubles the performance of RDMA, GPUDirect Storage, GPUDirect RDMA and In-Networking Computing.

NVIDIA OVX server versions compared
Configuration 2nd Generation 1st Generation
Networking 3x ConnectX-7 SmartNIC 3x ConnectX-6 SmartNIC
CPUs 2x Intel Xeon Platinum 8362
Memory 1TB
Storage 16TB NVMe

The second-generation NVIDIA OVX systems will be available from Inspur, Lenovo and Supermicro by early 2023, with GIGABYTE, H3C and QCT offering them in the future.

Source: NVIDIA

Join HWZ's Telegram channel here and catch all the latest tech news!
Our articles may contain affiliate links. If you buy through these links, we may earn a small commission.