[ad_1]
Tuesday at its GTC developer convention, Nvidia launched a brand new computing system for industrial digital twins, the Nvidia OVX.
Nvidia OVX was created for the aim of working digital twin simulations throughout the Omniverse, Nvidia’s industrial operations metaverse that’s “a real-time bodily correct world simulation and 3D design collaboration platform.”
“Simply as we’ve DGX for AI, we now have OVX for Omniverse,” mentioned Nvidia CEO Jensen Huang in his GTC keynote.
In fast-paced industries constrained by time, bodily area, and computing limitations, digital twin expertise is promising to remodel operations by means of superior designing, modeling, and testing capabilities.
“Bodily correct digital twins are the way forward for how we design and construct,” mentioned Bob Pette, vp of Skilled Visualization at Nvidia. “Digital twins will change how each trade and firm plans. The OVX portfolio of programs will be capable of energy true, real-time, always-synchronous, industrial-scale digital twins throughout industries.”
Click on to enlarge. Supply: Nvidia
Specs
A primary technology Omniverse OVX computing system is comprised of eight Nvidia A40 GPUs, three Nvidia ConnectX-6 Dx 200-Gbps NICs, twin Intel Ice Lake 8362 CPUs, 1TB system reminiscence, and 16TB NVMe storage. The OVX computing system could be scaled from a single pod of eight OVX servers as much as a SuperPOD of 32 OVX servers when linked with Spectrum-3 swap cloth. A number of SuperPODS will also be deployed for bigger simulation wants.
Capabilities
“OVX will allow designers, engineers, and planners to construct bodily correct digital twins of buildings or create huge, true-to-reality simulated environments with exact time synchronization throughout bodily and digital worlds,” mentioned Nvidia. With the ever-increasing complexity inside trade, the testing and analysis of intricate programs and processes that usually function concurrently and autonomously could be difficult or typically not possible. Making a digital twin of those programs, similar to these in a manufacturing facility or warehouse involving autonomous autos or robots, permits firms to digitally optimize operations for higher effectivity or innovation earlier than precise, real-world deployment of modifications or developments.
Click on to enlarge. Supply: Nvidia
In his keynote deal with, Huang famous that due to the complexity of business programs, “the Omniverse software program and pc have to be scalable, low latency and assist exact time,” and since knowledge facilities course of knowledge within the lowest doable time, however not in exact time, the corporate needed to create a “synchronous knowledge heart” with OVX. “Most significantly, the community and computer systems are synchronized utilizing precision timing protocol and RDMA minimizes packet switch latency,” Huang mentioned.
Origins of Digital Twins
Additionally in his keynote deal with, Huang gave the instance of the well-known Apollo 13 mission the place astronauts’ lives hung within the stability 136,000 miles from house after an oxygen tank exploded as a consequence of defective wiring. Again on Earth, NASA engineers used a bodily duplicate of the mission’s Odyssey spacecraft to check procedures for power-cycling and oxygen preservation which finally enabled NASA to securely return the crew. Huang famous that catastrophe was averted due to the duplicate and accompanying mathematical modeling, and the prototypical digital twin idea was born. NASA later coined the time period digital twin, defining it as “a residing digital illustration of one thing bodily.”
A screenshot from Huang’s GTC keynote displaying NASA engineers working with a duplicate of the Odyssey spacecraft. Supply: Nvidia
Whereas not all digital twin functions are as thrilling as saving the lives of astronauts, firms right this moment are discovering comparable worth in creating digital twins within the Omniverse. In Germany, DB Netze, an infrastructure arm of the nationwide rail holding firm, Deutsche Bahn, “is constructing in Omniverse a digital twin of Germany’s nationwide railway community to coach programs for automated practice operation and allow AI-enhanced predictive evaluation for unexpected conditions in railway operations,” in response to Nvidia. DB Netze is optimistic concerning the better outcomes a digital twin can facilitate.
“Utilizing a photorealistic digital twin to coach and check AI-enabled trains will assist us develop extra exact notion programs to optimally detect and react to incidents,” mentioned Annika Hundertmark, head of railway digitization at DB Netze. “In our present challenge, Nvidia OVX will present the size, efficiency, and compute capabilities that we have to generate knowledge for intensive machine studying growth and function these extremely complicated simulations and situations.”
First technology OVX programs have already been deployed at Nvidia and for some early prospects, and a second-generation system is at present in growth that may profit from Spectrum-4, a 51.2 terabits-per-second, 100 billion transistor Ethernet swap meant to tune precision timing even additional. The corporate says its Nvidia-Licensed OVX system will probably be out there by means of Inspur, Lenovo, and Supermicro later this 12 months, together with enterprise-grade assist and Omniverse software program to be collectively supplied by Nvidia and the OEM system builders. Clients can check the OVX system at one in every of Nvidia’s Launchpad areas worldwide.
To observe a replay of Tuesday’s GTC keynote, go to this hyperlink.
[ad_2]
