Nvidia is expanding its Omniverse universe with cloud-based access to the metaverse creation platform and a new computing system to power the most complex industrial digital twins enabled by the platform, along with other Omniverse-related announcements the company made at its GTC Spring event this week.
During 2021 Nvidia founder and CEO Jensen Huang vigorously and frequently promoted the metaverse concept and the role of Omniverse in enabling it. At GTC Spring, the Omniverse news took second billing to the company’s launch of its Hopper architecture, Grace Superchip and other announcements, but there is no questioning Nvidia’s commitment to growing its Omniverse platform.
For starters, the company unveiled Omniverse Cloud, providing artists, creators, designers and developers instant access to Omniverse features, including 3D design collaboration and simulation, from almost anywhere and any device. While some new features are still under development, the company already is offering early access to Nucleus Cloud, a “one-click-to-collaborate” sharing tool that enables artists to access and edit large 3D scenes from anywhere, without having to transfer massive datasets.
Richard Kerris, vice president of the Omniverse platform at Nvidia, said during an Omniverse briefing this week that Omniverse Cloud users also can leverage Omniverse Create, an app for technical designers, artists and creators to interactively build 3D worlds in real time; and Omniverse View, an app for non-technical users to view Omniverse scenes streaming full simulation and rendering capabilities using the Nvidia GeForce NOW platform, which is powered by NVIDIA RTX GPUs in the cloud.
Richard Kerris, vice president of the Omniverse platform at Nvidia, said during an Omniverse briefing this week that
Jack Gold, principal analyst at J. Gold Associates, noted that Omniverse Cloud should help broaden access to capabilities that many firms are requiring as they begin to investigate what digital twins and other virtual model solutions can do for them.
“Omniverse is primarily a greenfield opportunity in that many orgs are looking for an integrated platform, and especially wanting to do things with digital twins, but it’s a heavy lift for many,” Gold said.”The advantage of moving to Omniverse Cloud will be the lack of need to deploy tech in on-prem data centers, and the ability to start small and experiment with the new technology.”
Companies that already have been using a variety of other tools to create such models may find it a little harder than those just starting out to move everything they are doing to a new full platform environment, but Gold said the service could prove effective “particularly [for] smaller scale companies who may want to go slow and build out the capabilities over time. It alleviates a major resource constraint that many have. But it still requires expertise to know how to build Omniverse and Digital Twin models. That might require going to a third party for help.”
Announcing OVX
Nvidia also announced OVX, a computing system architecture designed to power large-scale digital twins, nodding to the notion that the virtual worlds enterprises are creating with Omniverse are only going to get bigger, more complex and more difficult to run on computing architectures not purpose-built for the challenge.
The OVX server consists of eight Nvidia A40 GPUs, three Nvidia ConnectX-6 Dx 200-Gbps NICs, 1TB system memory and 16TB NVMe Storage. It scales from a single pod of eight OVX servers, to an OVX SuperPOD consisting of 32 OVX servers connected with Nvidia Spectrum-3 switch fabric or multiple OVX SuperPODs to accelerate massive digital twin simulations. OVX will be available later this year through Inspur, Lenovo and Supermicro, the company said.
Omniverse and Amazon
Kerris said the universe of Omniverse users and partners continues to expand, and that partners are bringing in greater access to technologies like “sensors and LiDAR-based scanning” to help users build richer, more accurate virtual scenes. More than 150,000 individuals have downloaded Omniverse, he said, even as Nvidia is making Omniverse easier to use without having to download it. Also, Nvidia has expanded its list of ways to connect to Omniverse to 82 Connector partners.
To further help users leverage Omniverse, Nvidia announced a multitude of new features and tools this week for developers. One of the most interesting is DeepSearch, a new AI-based search service that lets users quickly search through massive, untagged 3D asset libraries using natural language or images. The capability, now available in an early access stage, gives users a leg up navigating the expanding Omnivere universe, Kerris said.
One of the platform's big-name users is Amazon, Kerris said, adding, “Amazon has adopted Omniverse Enterprise and Isaac SIM [a robotics simulation application within Omniverse] to create full scale digital twins of their fulfillment centers, and to train the robots that operate in these warehouses.”
Deepu Talla, vice president of embedded and edge computing at Nvidia, added, “Using synthetic data generation from Omniverse and replicating it, Amazon is able to cut that training time from several days or weeks into hours.”