Race to Faster Fluent Results with Ansys Gateway Powered by AWS

Ansys Fluent users need exceptional computing speed and power to solve their numerically large models. Now they can leverage a purpose-built cloud platform: Ansys Gateway powered by AWS. 

Every day, in every part of the world, and in just about every industry, product development teams are using Ansys Fluent to model complex physical phenomena. They’re creating models with millions of elements and transient time scales as they replicate the effects of turbulence, single- and multiphase flows, combustion, conjugate heat transfer, and fluid-structure interactions. They’re also conducting system-level analyses for their product designs that consider trade-offs among cost, sustainability, durability, weight, and other factors.

The benefits are clear: These product development teams are driving time and costs out of the design cycle, as well as predicting real-world product performance with an incredibly high degree of accuracy via simulation. But they’re also creating numerically large problems that need to be solved quickly.

More Ansys customers are relying on cloud platforms to access the high-performance computing (HPC) resources they need to efficiently solve large, complex problems without investing in expensive on-premise clusters that require management and maintenance. A cloud delivery model simply makes good business sense for many product development teams, as it offers elastic, on-demand access to the most advanced chips, processors, memory, and storage technologies automatically. With its efficient HPC scaling and its ability to solve models over multiple processors, Fluent is especially well suited to cloud computing.

But, with so many cloud options available, which is the best choice for Fluent users? Ansys has simplified the cloud selection process by partnering with Amazon Web Services (AWS) to develop an innovative cloud solution that’s purpose built for product development teams leveraging engineering simulation. 

Ansys Gateway powered by AWS: Built for Simulation

The resulting cloud solution, Ansys Gateway powered by AWS, delivers dramatic improvements in computing power and solution speed for simulation. It’s also designed to accommodate the enormous data volumes, complex workflows, and cross-functional collaboration that characterize simulation today.

Available on AWS Marketplace, this solution marries the world’s most comprehensive and broadly adopted cloud platform with the deep expertise of Ansys in solving advanced engineering problems via HPC. Because Ansys experts know how to match a specific Ansys solution and problem type with an optimal HPC configuration, it delivers outstanding “plug-and-play” performance right out of the box. Its default virtual desktop infrastructure (VDI) and HPC settings have already been optimized for engineering simulation.

For more advanced users, Ansys Gateway powered by AWS also offers complete control over the cloud environment. Via a user portal, subscribers can configure their own unique VDI or HPC cluster that’s customized to their own simulation requirements. They can choose from a long list of AWS cloud instance types that represent various combinations of CPU, memory, storage, and networking capacity. Ansys Gateway powered by AWS also provides recommended instance types for each solver.

Cloud customization empowers Ansys Fluent users to choose the balance between solution speed and computing cost that’s right for their own specific needs. Some users may be facing an upcoming deadline and will want to choose the fastest runtime, regardless of cost. Other users may not need their CFD simulation results quickly and may choose a slower solution runtime that minimizes hardware costs. Only Ansys Gateway powered by AWS gives simulation users the capability to make their own informed choices.

Fueling Customized Performance for Fluent Users

To prove the value of Ansys Gateway powered by AWS for Fluent users, an expert team at Ansys recently conducted an exhaustive benchmarking study. The complete study is summarized in the white paper “Faster CFD Solution Times via the Cloud,” which is available for download. But a quick look at one component of the study illustrates the value of cloud customization via Ansys Gateway powered by AWS.

Ansys experts created a fairly typical Fluent simulation — modeling external aerodynamics for a Formula 1 racecar design — then solved the problem using different settings in Ansys Gateway powered by AWS. The car was modeled with a 140 million-cell hex-core mesh, using the realizable k-e turbulence model and the pressure-based pseudo-transient coupled solver in Fluent.

Ansys experts ran this simulation on five different AWS instance types with 10 different configurations that reflected a range of processors, core counts, RAM, and processing speeds. (It’s important to note that Ansys Gateway powered by AWS relies on Amazon EC2, or Elastic Compute Cloud.)​

In comparing cloud performance across these instances and configurations, the Ansys team focused on solver rating, which is defined as the number of times a given benchmark study can be run on a given machine in a 24-hour period. The solver rating is computed by dividing the number of seconds in a day by the number of seconds required to run a given simulation. A higher solver rating means better HPC performance.

The Ansys team also looked at the cost implications of achieving higher HPC performance, so Fluent users can use the results to make informed cost-performance trade-offs in running their own simulations.

The Formula 1 aerodynamics simulation via Fluent performed best on two instances: Amazon EC2 Hpc6a and Hpc7a. For these two AWS instances, per-core performance heavily depends on how many cores are utilized. Lower core utilization means more memory bandwidth and cache are available per core, which translates to better per-core performance and faster runtimes — but at a higher hardware cost. 

By fine-tuning core utilization, Fluent users can balance cost versus performance, whether their goal is the fastest solution, the most affordable solution, or achieving a good balance between speed and cost. Please read the white paper for a full discussion of the study results.

A Cloud Solution That Puts You in the Driver’s Seat

While this Ansys study revealed clear benefits in solution runtime for two instances, it’s critical to note that there is no single “best” cloud computing environment for Fluent, or any other Ansys software product.

There will always be a set of optimal settings for the software, as well as a multitude of AWS hardware instances that can be chosen based on availability, cost, and desired simulation turnaround time. But as HPC technologies continue to evolve and advance, different chip or RAM solutions will emerge as the optimal choice based on speed considerations, hardware costs, or some balance between the two.

So, what’s the key takeaway from the Ansys benchmark study focusing on Fluent? It’s simple: Cloud configuration makes an enormous difference.

By putting Ansys Fluent users in the driver’s seat, Ansys Gateway powered by AWS enables users to customize their simulation results in terms of both runtime and cost. By choosing the default “plug-and-play” settings recommended for them, Fluent users are already achieving a huge time and cost advantage over generic cloud settings, but they can take that advantage even further by easily customizing the cloud environment for their own specific requirements — and they can see the projected time and cost implications upfront, before they finalize their cloud settings.

Ansys will deliver new features and functionality in Fluent and its other solutions that will be instantly and seamlessly accessible to subscribers. AWS will offer new chips, core types, RAM capacities, and scaling behaviors that further accelerate runtimes.

But one thing won’t change: Ansys Gateway powered by AWS will always deliver the ideal cloud environment for engineering simulation while giving users full control over that environment. Learn more about how this cloud solution can optimize both your simulation costs and outputs by requesting a free trial today.