Engineering simulation is a critical tool used across various industries to accelerate product development, enhance safety, and optimize designs. However, these simulations typically require substantial high-performance computing (HPC) resources and can be time-consuming. Luminary Cloud, a member of the NVIDIA Inception program, aims to address these challenges by utilizing cloud and NVIDIA GPU technologies to expedite these processes.

Challenges in Computational Fluid Dynamics

Computational fluid dynamics (CFD) plays a significant role in engineering and scientific advancements. However, several challenges hinder its widespread adoption. A recent survey highlights these issues, including lengthy simulation turnaround times, the need for improved simulation robustness and accuracy, inefficient model preparation, and high costs.

• Simulation turnaround time: Many CFD projects require weeks or even months to complete.
• Simulation robustness: Improvements are needed in automatic meshing capabilities and price-performance ratios.
• Model preparation: Setting up CFD simulations can be time-intensive, taking months in some cases.
• Cost: High license costs are a significant barrier, with budget constraints being a major challenge for many organizations.

Accelerated Simulation with NVIDIA GPUs

NVIDIA GPUs have revolutionized various fields, including quantum computing, climate science, and artificial intelligence, by providing processing speeds up to 1,000 times faster than CPU-only processing. In CFD, the acceleration can be over 30 times faster than traditional CPU-based computing. This advancement has significantly reduced the time required for complex simulations.

High-Performance Computing in the Cloud

Access to HPC resources is a common challenge in CFD. The cloud offers a solution by providing scalable and efficient HPC resources that can be accessed as needed. This approach increases cost efficiency, reduces time-to-market, and allows companies to leverage the latest hardware, such as NVIDIA GPUs, without the need for in-house infrastructure.

Luminary Cloud’s Innovative Solution

Luminary Cloud offers a multi-physics solution supporting CFD for fluid-flow physics and conjugate heat transfer (CHT) for thermal management. The platform leverages cloud-based SaaS and multi-node NVIDIA GPU accelerated computing to provide a near-real-time engineering experience. Lumi AI, an AI-based engineering design copilot, further enhances efficiency by automating mesh generation and adapting computational meshes for higher accuracy.

In a demonstration, Luminary Cloud ran 20 parallel simulations with 40 million control volumes each, completing them in under two minutes using a modest number of GPUs. Another simulation involving a full aircraft geometry with 150 million control volumes was completed in about seven minutes using 40 NVIDIA A100 GPUs, a task that would have taken several hours with traditional methods.

Case Study: Joby Aviation

Joby Aviation has benefited significantly from Luminary Cloud’s platform and NVIDIA GPUs. The platform has enabled Joby to rapidly iterate on designs, reducing the time required for essential redesigns and expediting the regulatory approval process. This efficiency is crucial in the competitive market of electric air taxis.

“You can take complete aircraft configurations and run them in a matter of minutes. It allows a level of confidence that was unprecedented before. You can quickly say ‘Yes, this will or will not work,’” explained Joby Chief Aerodynamicist Gregor Mikić.

Benefits of Cloud-Based CAE

Luminary Cloud’s solution offers a novel, real-time engineering experience, facilitating highly interactive simulations that expedite engineering processes. The platform supports various modeling approaches, allowing users to select the appropriate accuracy level for their projects. Its user-friendly interface ensures a seamless experience, enhancing productivity and enabling exploration of new frontiers in computational engineering.

For more information about Luminary Cloud, visit the NVIDIA Technical Blog.

Image source: Shutterstock