NVIDIA Unveils Datacenter-Scale Omniverse Computing System for Industrial Digital Twins

NVIDIA OVX is purpose-built for simulation of large-scale, complex applications in robotics, AI, industrial automation, and more

SANTA CLARA, Calif. – GTC – March 22, 2022 – NVIDIA today announced the NVIDIA® OVX™ compute system for powering large-scale digital twins .

NVIDIA OVX is designed for running complex digital twin simulations in NVIDIA Omniverse™ , a real-time, physically accurate world simulation and 3D design collaboration platform.

The OVX system combines high-performance GPU-accelerated computing, graphics processing, and AI with high-speed storage access, low-latency networking, and precise timing, providing the performance required to create realistic digital twins. OVX will be used to simulate complex digital twins to build models of buildings, factories, cities, and even the world.

“Physically accurate digital twins are the future of design and construction,” said Bob Pette, vice president of Professional Vision at NVIDIA. “Digital twins will change the way every industry and company plans, and the OVX portfolio of systems will power realistic, real-time, always-synchronized, industrial-scale digital twins for every industry.”

With OVX, designers, engineers, and planners will be able to build physically accurate digital twins of buildings or create large-scale realistic simulation environments, and achieve precise time synchronization between the real and virtual worlds. Enterprises can evaluate and test complex systems and the interactive processes of multiple autonomous systems in the same space and time, thereby optimizing, expanding or creating more efficient factories and warehouses, or training robots and self-driving cars before deploying them into the real world.

Building a digital twin

Under the Digitale Schiene Deutschland initiative, DB Netze is building a digital twin of the German national railway network in Omniverse, which will be used to train automatic train operation systems and perform AI-enhanced predictive analysis of unexpected situations in rail transport.

“Using realistic digital twins to train and test AI trains will help us develop more accurate perception systems to optimally detect and respond to incidents,” said Annika Hundertmark, head of rail digitization at DB Netze. “NVIDIA OVX will provide the scale, performance and compute power we need for our current projects, allowing us to generate data for intensive machine learning development and operate these highly complex simulations and scenarios.”

Computing system specifications

The OVX Server consists of eight NVIDIA A40 GPUs, three NVIDIA ConnectX®-6 Dx 200Gbps network cards, 1TB of system memory, and 16TB of NVMe storage. The OVX compute system can scale from a single cluster of eight OVX Servers to one or more OVX SuperPODs (consisting of 32 OVX Servers) connected via the NVIDIA Spectrum-3 switch fabric to accelerate large-scale digital twin simulations.

Availability

OVX Solutions are NVIDIA-Certified Systems™ that have been tested and validated to deliver the necessary performance, manageability, security, and scalability. NVIDIA and OEM system builders will work together to provide full, enterprise-class support for NVIDIA OVX Solutions and Omniverse software.

NVIDIA OVX will be available later this year through Inspur, Lenovo and Supermicro.

To learn more about NVIDIA Omniverse, watch Jensen Huang’s GTC 2022 keynote. Register for free at GTC 2022 to attend sessions hosted by NVIDIA and industry leaders.