Cadence brings AI supercomputing to digital twin simulations

Cadence Design Methods has launched a {hardware} and software program platform to convey supercomputing efficiency to digital twin and AI simulations.

San Jose, California-based Cadence has centered on instruments for digital techniques design. And immediately it’s asserting its Cadence Millennium Enterprise Multiphysics Platform, which it says is an enormous leap ahead in simulation.

The platform leverages GPU and CPU computing to ship as much as 100 occasions the design affect by means of accuracy, pace and scale.

GPU-resident solvers dramatically enhance the power effectivity of simulation by 20 occasions over
conventional high-performance computing. Cadence stated the digital twin, AI and HPC know-how mix to offer the very best multiphysics simulation answer for automotive, A&D, power, turbomachinery and information middle markets.

“All through our 35-year historical past, Cadence has been centered on growing efficiency with no lack of accuracy for probably the most difficult computational purposes. Algorithmic throughput stays a key precedence, and we’re now leveraging generative AI to reap data gained from large portions of design and simulation information,” stated Ben Gu, company vp of R&D for multiphysics techniques evaluation at Cadence. “Our revolutionary Millennium platform is a big leap ahead, delivering unprecedented acceleration and scalability of digital twin and AI purposes. CFD is poised to profit enormously from efficiency and effectivity positive factors, and the large energy of the Millennium M1 is disrupting industries that should discover extra design improvements and produce them to market sooner.”

The {hardware}/software program combo will make it simpler to create digital twins, or bodily correct digital imitations of real-life factories and different bodily creations.

The platform integrates graphics processing items (GPU) and central processing unit (CPU) computing with proprietary software program algorithms, ushering in higher efficiency and power effectivity for multiphysics system design and evaluation.

The primary-generation Cadence Millennium M1 focuses on accelerating high-fidelity computational fluid dynamics (CFD) simulations, catering to industries resembling automotive, aerospace and protection (A&D), power, and turbomachinery. The platform goals to handle the growing demand for designing mechanical techniques that obtain optimum efficiency and effectivity, aligning with trade priorities resembling bettering gas effectivity, decreasing emissions, and enhancing total system performance.

Options and advantages

Efficiency Revolution: Combines GPU-resident CFD solvers with devoted GPU {hardware}, attaining supercomputer-equivalent throughput per GPU of as much as 1, 000 CPU cores. It ensures optimum efficiency for advanced mechanical techniques with near-linear scalability on a number of GPU nodes.

Effectivity overhaul: Reduces simulation turnaround time from weeks to hours, boasting 20 occasions higher power effectivity in comparison with conventional CPU equivalents. Drives energy-efficient simulation, a important side for industries prioritizing sustainability.

Accuracy and constancy: Leverages Cadence Constancy CFD solvers, offering unmatched accuracy to deal with advanced simulation challenges. Ensures high-fidelity simulation information essential for constructing dependable digital twin visualizations.

AI digital twin integration: Employs generative AI know-how for speedy technology of high-quality multiphysics information. Permits the creation of quick and dependable digital twin visualizations for optimum system design options.

Turnkey answer: {Couples} GPU compute with trendy and scalable CFD solvers, providing an optimized setting for accelerated CFD and multidisciplinary design and optimization. Gives flexibility with deployment choices, obtainable within the cloud or on premises, accommodating numerous buyer wants.

The corporate acquired endorsements from massive corporations and lecturers, together with Honda, Corvid Applied sciences, Asus and Stanford College’s Parviz Moin.

“Honda and Cadence have labored collectively for greater than 10 years to speed up CFD design
processes,” stated Atsushi Ogawa, COO at Honda R&D Co., in an announcement. “Our deployment of Cadence’s Millennium M1 multiphysics supercomputing platform, with its LES solver, has demonstrated considerably sooner turnaround time and the scalability to simulate a lot bigger techniques for aerodynamics, combustion and aeroacoustics purposes for our automotive, aeropropulsion and eVTOL initiatives. With the potential for vital acceleration of high-fidelity CFD simulations, we sit up for persevering with our collaboration to develop next-generation CFD options.”