Never miss a story from Computational Fluid Dynamics. Subscribe for in-depth analysis and articles.
As we witness a surge in the need for computational resources accompanied by skyrocketing data center power, it becomes difficult for organizations to comply with and achieve net-zero targets. However, these challenges can be tackled with powerful tools such as accelerated computing and AI. NVIDIA GTC 2023 was all about the new chips and systems, acceleration libraries, cloud and AI services, and partnerships that open new markets.
In his keynote, NVIDIA’s CEO Jensen Huang shared his perspective on how acceleration libraries solve new challenges and open new markets. CFD tools benefit the aerospace and auto industries in turbulence and aerodynamic predictions, while CFD helps the electronics domain in effective thermal management design. For advanced CFD simulations, vast CPU/GPU acceleration is substantial. Huang mentioned, "At equivalent system cost, NVIDIA A100 is 9X the throughput of CPU servers, or at equivalent simulation throughput, NVIDIA is 9X lower cost or 17X less energy consumed.” He also presented a slide on our new CFD solver, CharLES, accelerated by NVIDIA CUDA (shown below).
From Cadence, Ben Gu, VP of R&D, discussed Optimizing Formula One with Fidelity. He explained how system design is undergoing a hyper-convergence of physics. The system layer of interest in his talk was an intelligent automobile. Aerodynamics, aeroacoustics, thermal management, and EMC/EMI requirements are substantial for overall vehicle design. Moreover, CFD is indispensable in simulating the former three physics!
He elaborated on Cadence's recent investment in CFD tools such as NUMECA, Pointwise, Future Facilities, and most recently, Cascade Technologies, as a step towards supporting system companies in this hyper-convergence of physics. Moreover, by acquiring Cascade Technologies, Cadence has extended its CFD capabilities to include a high-fidelity approach called Large Eddy Simulation (LES) that uses GPU acceleration.
Ben Gu mentioned that F1 does not currently allow GPU acceleration of the solvers due to compliance with the allocation unit hours (Auhs) definition. However, this transformation is impacting other areas of racing, commercial automotive, and many other industries. Watch Ben Gu's keynote on Optimizing Formula One with Fidelity for more insights on Cadence high-fidelity CFD solutions.
To experience and get hands-on with our high-fidelity CFD solutions, request a demo today!