Ansys Partners with Supermicro and Nvidia to Accelerate Multiphysics Simulation Speeds by Up to 1,600X

Ansys has partnered with Supermicro and Nvidia to provide turnkey hardware solutions that deliver exceptional acceleration for Ansys multiphysics simulations.

By harnessing this synergy of top-tier hardware and software, Ansys users can tackle larger and more complex models, achieving speeds up to 1,600 times faster. The collaboration with Supermicro and Nvidia technology enables faster time-to-market and enhances design exploration across various applications, including automotive crash testing, aerospace gas turbine engines, 5G/6G antennas, and biopharmaceutical development. Notably, Synopsys is in the process of acquiring Ansys for $35 billion.

Maximizing the benefits of multiphysics simulations involves integrating diverse physics solvers with tailored hardware, each offering distinct performance advantages. However, selecting and configuring the appropriate hardware for multiphysics can be complicated, with significant implications for performance, cost, and productivity.

Customized turnkey hardware solutions, featuring central processing units (CPUs), graphics processing units (GPUs), interconnects, and cooling systems, enable engineers to conduct highly accurate simulations more efficiently.

Collaborative testing by Ansys and Supermicro found that replicating Ansys Fluent and Ansys Rocky performance on a single Nvidia GPU would require the equivalent of 1,500 and 480 CPU cores, respectively. Furthermore, Ansys Perceive EM on one Nvidia GPU matches the capability of over 1,000,000 CPU cores. The testing also revealed impressive acceleration metrics: Ansys optiSLang AI+ achieved a 1,600x speed-up; Ansys Fluent saw a 24x speed-up; Ansys Mechanical a 6x speed-up; Ansys HFSS an 11x speed-up; Ansys Perceive EM a 53x speed-up; Ansys Rocky a 17x speed-up; and Ansys LS-DYNA a 4x speed-up.

These speed enhancements were made possible by integrating or substituting CPU cores and GPUs with Nvidia technology, including Nvidia H100 GPUs, Nvidia L40S GPUs, and the Nvidia Grace CPU Superchip.

“The breadth and depth of the Ansys multiphysics portfolio necessitates a strategic approach to compute infrastructure with GPUs,” stated Vik Malyala, President of EMEA and SVP of Technology & AI at Supermicro. “We are actively collaborating with Ansys and Nvidia, leveraging a wide range of GPU and CPU systems (Hyper/CloudDC, scalable GPU systems: 4U-10U, and more) to accelerate simulations, eliminate guesswork, and enhance customer deployment times globally.”

With this innovative technology, engineers can minimize overhead costs and reduce energy consumption by utilizing fewer servers for equivalent workloads.

“The synergy between Ansys and Nvidia is ushering in a new era of technological innovation,” said Dion Harris, Director of Data Center Product Solutions at Nvidia. “Ansys simulation solutions are instrumental in developing our advanced AI superchips, while Nvidia’s accelerated data center AI and Digital Twin platforms enable Ansys to push the limits of simulation performance. Together, we are unlocking profound insights and paving the way for groundbreaking progress that will define the future of engineering and AI.”

Most people like

Find AI tools in YBX