lawrence jenga
November 21, 2024 18:10
NVIDIA’s Grace and Grace Hopper CPUs deliver unprecedented energy efficiency and performance to Ansys workloads, transforming computing tasks in areas such as automotive safety and aerodynamics.
NVIDIA has unveiled its latest advancement in computing performance with the introduction of Grace and Grace Hopper CPUs specifically designed to power Ansys workloads. According to NVIDIA’s official blog, these cutting-edge processors are set to redefine performance benchmarks in high-performance computing (HPC), especially in industries such as automotive engineering and aerodynamics.
Innovative architecture for increased efficiency
Featuring 72 high-performance Arm Neoverse V2 cores, NVIDIA Grace CPUs are designed to maximize data flow efficiency with a Scalable Coherency Fabric that delivers 3.2TB/s of bipartite bandwidth. This architecture supports high-speed LPDDR5X memory, providing significant bandwidth at a fraction of the power required by traditional DDR memory systems.
In contrast, the NVIDIA GH200 Grace Hopper Superchip integrates the Grace CPU and Hopper GPU. This combination leverages the high-bandwidth NVLink Chip-2-Chip interconnect to allow seamless memory coherence between CPU and GPU threads, significantly improving performance and developer productivity.
A revolution in automotive engineering
Ansys LS-DYNA software, the standard for crash analysis in the automotive safety domain, runs primarily on the CPU. Grace CPUs offer a powerful alternative with significant energy and cost savings potential, which is critical for power-constrained data centers.
Benchmark tests show that Grace CPUs excel in performance and energy efficiency, outperforming existing x86 architectures. These efficiencies allow data centers to increase computing power without exceeding existing energy budgets.
Advances in aerodynamic simulation
In aerodynamics, Ansys Fluent software leverages the capabilities of the Grace Hopper superchip. Recent tests at the Texas Advanced Computing Center showed a 110-fold increase in the speed of large-scale simulations, producing results in hours rather than weeks. These performance improvements are accompanied by significant reductions in energy consumption and operating costs.
Future Outlook and Industry Impact
With the introduction of these new processors, NVIDIA is not only pushing the boundaries of what is possible in computing engineering, but is also setting a new standard for energy-efficient computing. The upcoming NVIDIA Vera CPUs are expected to further enhance these features, making Arm-based architecture the best choice for HPC environments.
As the industry continues to pursue efficient and powerful computing solutions, NVIDIA’s advancements with Grace and Grace Hopper CPUs represent an important step forward in meeting these needs, especially in sectors where computing performance and energy efficiency are critical.
Image source: Shutterstock