Accelerating LS-DYNA Engineering Simulations up to 2.5x Faster

In the realm of engineering simulations, accuracy, speed, and scalability are paramount. Recognizing these requirements, Ansys, AMD with their EPYC™ processors, and Amazon Web Services (AWS) have collaborated to deliver an extraordinary performance boost for engineering simulations using Ansys LS-DYNA on Ansys Gateway powered by AWS. The study, “Ansys LS-DYNA on Amazon EC2 HPC6A Instances Featuring AMD EPYC^TM 7003 Series Processors,” showcases the remarkable capabilities of this partnership, revolutionizing the way engineers approach Ansys simulations. Let's dive into the key findings of this study and explore the benefits of this collaboration.

Unleashing Exceptional Performance

The study highlights the exceptional performance capabilities of AMD EPYC™ processors when running LS-DYNA simulations on Amazon EC2 instances. These processors are equipped with high core counts, excellent memory bandwidth, and robust floating-point performance, making them ideal for computationally intensive tasks like LS-DYNA simulations. The benchmarking results demonstrate significant performance improvements, enabling engineers to achieve faster simulation turnaround times, optimize design iterations, and enhance overall productivity.

Enhanced Scalability and Flexibility

Scalability is crucial with this kind of simulation, especially when dealing with complex models and large-scale simulations. The collaboration between Ansys, AMD, and AWS delivers enhanced scalability by harnessing the power of AMD EPYC™ processors on Amazon EC2 instances. With the ability to scale up to thousands of cores, engineers can tackle simulations that were previously challenging or time-consuming. This scalability empowers engineers to push the boundaries of their simulations, enabling the exploration of complex phenomena and capturing accurate real-world behavior.

Seamless Migration and Easy Adoption

The study emphasizes the seamless migration process from on-premises or other cloud environments to Ansys Gateway powered by AWS. Engineers using LS-DYNA can effortlessly transition their simulations to the cloud, without extensive code modifications or reconfigurations. This streamlined migration process ensures minimal disruption to ongoing simulation workflows, enabling engineers to take advantage of the enhanced performance and scalability promptly.

Streamlined Workflow and Improved Efficiency

The collaboration between Ansys AMD and AWS not only delivers enhanced performance but also streamlines the overall simulation workflow, leading to improved efficiency. AWS provides engineers with a robust and scalable cloud infrastructure, eliminating the need for costly on-premises hardware investments and maintenance. The study highlights the ease of provisioning instances on Ansys Gateway powered by AWS and managing simulation jobs, empowering engineers to focus on their simulations rather than infrastructure management. This streamlined workflow enhances productivity and allows engineers to devote more time to critical design decisions.

What are the Results?

This performance brief displays LS-DYNA 13.1.1 running the benchmarks on Ansys Gateway powered by AWS shown in Figures 1 and 2 on Amazon EC2 Hpc6a instances featuring AMD EPYC 7003 Series Processors. Each instance includes 96 physical cores with AMD Simultaneous Multithreading (SMT) disabled and a 100 Gbps Elastic Fabric Adapter networking for fast internode communication that supports scaling from one to four instances.

Ansys LS-DYNA scales very well on Hpc6a instances with a speed up of ~2.47x at 384 cores for both models.¹ FEA explicit solvers typically do not scale beyond two to four nodes, with production models running ~400 to 500 cores.

Amazon EC2 Hpc6a instances powered by AMD EPYC 7003 Series Processors offer excellent performance and scalability for running HPC workloads. Amazon EC2 Hpc6a instances introduce several targeted features to deliver cost and performance optimizations for customers running tightly coupled HPC workloads that rely on high levels of inter-instance communications. With 100 Gbps Elastic Fabric Adapter networking, and AWS Nitro System making all 96 cores available, the Amazon EC2 Hpc6a instances let you scale LS-DYNA explicit workloads at the best price performance in Amazon EC2.

Interested in learning more? Read the full benchmark. Or contact us if you want to try it for yourself.

References

1. 3rd Gen AMD EPYC™ processors deliver outstanding scale-out performance running Ansys® LS-DYNA® on one to four Amazon Web Services Hpc6a instances.