SC 24: NVIDIA Reveals Blueprint for CAE Software Developers to Build Interactive Digital Twins

At Supercomputing 2024 (SC24) in November, NVIDIA unveiled an NVIDIA Omniverse Blueprint designed to help computer-aided engineering (CAE) software developers build real-time digital twins for their customers. The Omniverse Blueprint for real-time computer-aided engineering digital twins consists of reference workflows built around NVIDIA CUDA-X acceleration libraries, NVIDIA Modulus for physics-based machine learning and AI training, and interactive visualization in Omniverse via APIs. 

To achieve interactive simulation and visualization, traditional CAE software vendors such Ansys, Siemens, and Altair offer simulation technology—for example, finite element analysis (FEA) and computational fluid dynamics (CFD) software—for engineering projects. Now, CAE vendors can incorporate the NVIDIA Blueprint to their workflows to achieve real-time simulation with GPU acceleration, AI emulation and interactive visualization. 

According to NVIDIA, “Altair, Ansys, Cadence, and Siemens can use Omniverse Blueprint for real-time CAE digital twins to help their customers drive down development costs and energy usage while getting to market faster. The blueprint is a reference workflow that includes NVIDIA acceleration libraries, physics-AI frameworks, and interactive physically based rendering to achieve 1,200x faster simulations and real-time visualization.”

The announcement serves as a signal to CAE software developers  around the world that templates are now in place to enable manufacturers to easily develop, build, and operate CAE-driven virtual replicas of real-world products and processes. For CAE vendors, it’s an opportunity to add made-to-spec digital twins to their offerings. 

Tim Costa, NVIDIA senior director of CAE and EDA, said, “Our Omniverse Blueprint for real-time computer-aided engineering digital twins is revolutionizing the way industries approach design and simulation. By enabling the creation of interactive digital twins with physics-based AI and real-time visualization, we're empowering engineers to innovate faster, reduce costs, and push the boundaries of what's possible in aerospace, automotive, and manufacturing sectors.” 

To further these efforts, NVIDIA has worked closely with Dell Technologies to configure optimized Dell Precision professional workstations that enable developers and users of engineering applications to  leverage these new capabilities powered by NVIDIA Omniverse and AI technology.

Adaptive Real-Time CFD Demonstration

As a proof of concept, NVIDIA and its partner Luminary Cloud demonstrated a virtual wind tunnel at SC24 that allows real-time CFD simulation of airflows and fluid dynamics to test the aerodynamics of vehicles. The test showed near-instant recalculation of the flow fields in response to changes in the vehicle model. 

NVIDIA revealed it’s also partnering with Rescale, a high-performance computing service provider targeting CAE users, to incorporate the blueprint into the Rescale physics-AI platform. 

Costa continued, “This ability to achieve near-instantaneous feedback on design changes makes complex workflows more intuitive and accessible—marking a milestone in AI-powered engineering.”

The Omniverse Blueprint can be deployed on premises or in the cloud. Cloud options include NVIDIA DGX Cloud, Amazon Web Services, Google Cloud Platform, Microsoft Azure, and Oracle Cloud Infrastructure.

Ansys Adopts NVIDIA Blueprints

Ansys became one of the first simulation software developers to adopt this blueprint. Ansys adopted the blueprint with CUDA-X libraries for GPU acceleration of the solver, Modulus for AI emulation and Omniverse for interactive visualization. Ansys ran its CFD software Ansys Fluent at the Texas Advanced Computing Center on 320 NVIDIA GH200 Grace Hopper Superchips. “A 2.5-billion-cell automotive simulation was completed in just over six hours, which would have taken nearly a month running on 2,048 x86 CPU cores, significantly enhancing the feasibility of overnight high-fidelity CFD analyses and establishing a new industry benchmark,” according to NVIDIA.

“By integrating NVIDIA Omniverse Blueprint with Ansys software, we’re enabling our customers to tackle increasingly complex and detailed simulations more quickly and accurately,” said Ajei Gopal, president and CEO of Ansys. 

According to NVIDIA, Altair, Beyond Math, Cadence, Hexagon, Neural Concept, Siemens, SimScale, and Trane Technologies are also looking at the Omniverse Blueprint for integration into their own software. 

Foxconn Builds Digital Twin in Omniverse

Also at SC24, NVIDIA announced that Foxconn, the Taiwan-based contract manufacturing giant, is helping produce NVIDIA Blackwell GPUs. The company is also one of the manufacturers using NVIDIA Omniverse to develop industrial-scale  digital twins of its facilities.

Foxconn maintains a digital twin of its factory in Guadalajara, a hub of Mexico’s electronics industry. The factory engineers use the virtual factory to identify the best placement of robotic arms and to monitor the overall processes in real time with sensors and video cameras.

According to NVIDIA, Foxconn “has built digital twins with Omniverse that allow their teams to virtually integrate facility and equipment information from leading industry applications, such as Siemens Teamcenter X and Autodesk Revit. Floor plan layouts are optimized first in the digital twin, and planners can locate optimal camera positions that help measure and identify ways to streamline operations with Metropolis visual AI agents ... Using NVIDIA Omniverse and AI, Foxconn plans to replicate its precision production lines across the world. This will enable it to quickly deploy high-quality production facilities that meet unified standards, increasing the company’s competitive edge and adaptability in the market.”

Foxconn is producing and testing NVIDIA Blackwell systems at its factories in the U.S., Mexico, and Taiwan. The company is also building what is expected to become Taiwan’s largest supercomputer on NVIDIA’s GB200 NVL72 platform, fitting a total of 64 racks and 4,608 Blackwell GPUs. Check out this blog to learn more about Foxconn’s innovative use of NVIDIA AI and Omniverse to power factory scale simulations for facility planning and training autonomous systems. 

NVIDIA also announced CUDA-QX, an extension of CUDA-Q comprising optimized libraries. NVIDIA previously launched CUDA-Q, an open-source quantum development platform to orchestrate the hardware and software needed to run quantum computing applications. 

The company said, “CUDA-QX provides optimized kernels and APIs for key quantum computing primitives. This lowers the threshold for researchers and developers to access CUDA-Q’s GPU acceleration, leaving users with more time to focus on novel science and application development rather than code optimization.”

Discover how to create real-time digital twins by combining accelerated solvers, simulation AI, and virtual environments. For early access to NVIDIA Omniverse Blueprint for Real-time Computer-Aided Engineering Digital Twins, reach out to our sales team today.

More NVIDIA Coverage