Microsoft and Altair are enabling transformative CAE via the cloud

 In today’s modern technology marketplace, manufacturers are under pressure to evolve their product development and accelerate innovation. They are being asked to iterate faster on design while also improving quality by simulating and replicating the real-world performance of physical products and incorporating feedback from IoT devices. In response to these challenges, manufacturers have turned to high-performance computing (HPC) to help drive the product development lifecycle. Altair’s technology collaboration with Microsoft helps manufacturers reliably leverage simulation tools rather than physical prototypes to improve product speed-to-market.

Microsoft’s Kurt Niebuhr, Principal Program Manager for Microsoft Azure Big Compute, sat down with Altair to discuss how manufacturers can transform their product design and engineering processes. Below is an interview between Kurt and Srikanth (Sam) Mahalingam, Chief Technology Officer, an expert with decades of experience in the manufacturing industry.

Kurt: Could you walk me through Altair’s approach to product engineering and how the Altair suite uniquely fosters innovation? What are some of the advantages that customers realize from using Altair’s engineering platform?

Sam: Sure, so as you all know Altair is a leader in the computer aided engineering (CAE) space. There are a tremendous number of areas where simulation can be applied today. These include 3D simulations used during the design phase of the product, simulations used to assist with design verification and validation, and also manufacturing simulations used to ensure whatever part of the product being designed can be feasibly manufactured. Nowadays, simulation can also be used in the individual operations of a particular product in order to determine the remaining usable life of the product and also to predict failure. Altair provides simulation tools for all of the use cases I just described.

This is just one way we are helping our customers – by ensuring they can reliably leverage simulation tools rather than physical prototypes to ensure products can be launched faster to the market.

Having said that, where Altair is truly unique is that our platform focuses on optimization – a specialization that began primarily with structural optimization but has since been expanded to multidisciplinary optimization. This differentiates us from our competitors, providing our customers with real value. Basically, rather than having our customers create a model and validate the model we can instead prescribe the engineering requirements that will lead to a final design in a very short period of time.

 

Kurt: Can you give a real-world example of a problem that is uniquely solved via multidisciplinary optimization?

Sam: Sure. One of the automotive companies in Europe is leveraging our multidisciplinary optimization tool to help design the vehicle’s body, chassis and powertrain, while also using our platform to verify and optimize to the performance targets required for vibration, fatigue, crash performance, etc.

In the past, there were two reasons why these things could not be done. One reason is that the tools require tremendous amounts of compute power. Another is that the data needs to be related in a consumable fashion to ensure that engineers can incorporate multiple design concepts into a single simulation cycle.

As an example, due to carbon emissions requirements, you may need to design a car to be more lightweight while at the same time ensuring the car still meets the desired performance characteristics and requirements for vibration and crash safety. Altair’s platform is ideal for this sort of multidisciplinary optimization, where engineers are enabled to yield a final design that effectively balances the multidisciplinary requirements.

 

Kurt:  What are your customers telling you about the new challenges they are facing?

Sam: Our customers are facing compressed timelines to get new products to market.

Data is playing a much bigger role. With the advent of Industry 4.0 and smart, connected products, real-time operational data and the real-time context in which products are operating is coming to a central system. Now, customers can simulate on real-world data and deliver optimized designs specific to operational requirements. Customers are asking for us to deliver this digital twin concept and drive the next product development lifecycle ensuring the products are more optimized and rightsized.

Another area where customers are pushing us is broader support for digital manufacturing scenarios. A customer recently came to us with a bill of materials marking which items could be manufactured digitally rather than through traditional fabrication. Now, when these customers come to us with parts they feel can be 3D printed, we can use the technology available today to go in and create designs for them that are 3D printable.

 

Kurt: In what other ways are your customers pushing you to innovate?

Sam: Our customers are pushing us to get results much faster. Specifically, with new lower power hardware designs (both GPU and CPU), we are pushing our products to become highly parallelized to leverage the new hardware innovations.

We are innovating by making software that is highly parallelized and highly scalable, which is important with multi-physics and multi-disciplinary design. Data visualization software is also becoming highly parallelized, as our customer must be able to visualize the massive amounts of data created by multi-physics simulations.

 

Kurt: What advantages and advances have you seen in simulation technology that computer-aided engineering (CAE) has enabled?

Sam: The areas where we have seen a lot of advances have been simulations with structure-fluid interactions and structure-thermal interactions–any scenario bringing in different physics and jointly simulating them rather than individually simulating them and increasing the iteration cycle. Solving for these multi-physics scenarios have been a big point of innovation for us. We have worked a lot in the past few years to ensure we can bring in as many physics solvers to talk to one another as we can, including electromagnetic, thermal structure, fluid structure, and electro magnetics vibration and acoustics.  We have seen this come in handy particularly with electric cars as you are solving for how the electric motor interferes with acoustics and the vibration of the car itself.

This is an area where we have advanced a lot though at the same time I think optimization is one of our key goals and objectives as well as we have been leaders in that space and want to maintain that position. We have been taking our structural optimization technology and we have been moving it to digital manufacturing where we can not only design structures that are 3D printable but also design structures with inbuilt components that are themselves 3D printable.

 

Kurt: What pain points do customers experience trying to run simulation and other CAE workloads on premise?

Sam: Limited resources. Customers needed to optimally utilize their compute resources, ensuring they’re rationed so that high priority jobs get the majority of what’s available. What is happening today is that although the software is capable of scaling, limited infrastructure and capital are in turn limiting the ability for designers and engineers to innovate. We have seen this at a lot of customer locations—infrastructure limiting innovation.

 

Kurt: Building from the prior scenario, given a situation where a customer was limited beyond their available on-premises infrastructure how would they typically respond in the past and how would they respond now?

Sam: In the past, when customers were limited by their on-premises infrastructure they would allocate all their existing resources to high-visibility processes while other ideas and innovations would be put on hold. The only option for the innovation cycle to proceed was to purchase and procure dedicated hardware, and then spend time configuring it to ensure that the software could run on it. This would typically take anywhere from eight weeks to six months for this new infrastructure to be usable for simulation.

 

Kurt: What advantages do your customers experience when leveraging the cloud to support Altair’s engineering platform beyond scale?

Sam: Today, if customers have the budget and want to innovate, they can simply burst to the cloud and leverage the HPC resources available. These cloud resources are effectively infinite if the customer is willing to invest to get them instantaneously—and leveraging these resources can stand to drastically increase the pace of innovation.

As an additional advantage, all of Altair’s products run easily on the cloud thanks to our licensing model. We always make sure to host our licenses on the cloud, so that whenever a customer leverages or executes our software on any cloud infrastructure, you can still use the same licenses since these are hosted on the public cloud.

The second advantage for our customers is that Altair is also a leader in the HPC space. We have set up our engineering platform as a true HPC infrastructure within the public cloud where it resides. This enables our customers to automatically leverage the HPC cluster that has been configured behind the scene when they use our tools. This is an important advantage that our customers can realize on our platform. We have a vast experience setting up these HPC clusters, and can ensure that the proper nodes are chosen for the different applications used by our customers. We do all this to ensure data doesn’t need to move from on-premises to the cloud and then back – instead, the data and licenses reside in a single location eliminating a potential data transfer bottleneck and enabling customers to access our entire suite from a single pane of glass. Our job is to let engineers be engineers not computer scientists. That’s why we have set up our platform to provide the simplest and most powerful toolset possible.

 

Kurt: You touched on this just briefly earlier, but do you see different advantages for different-sized customer businesses making the transitioning to cloud computing when running CAE?

Sam: Cloud is really changing the game for small and medium-sized businesses. In the past, these smaller businesses would need the IT expertise to set up their own HPC infrastructure in-house. This was a complicated process that could prove cost-prohibitive. With the cloud, these companies don’t have to worry about not owning a server room or a datacenter, they can simply leverage the engineering platform on the cloud and start innovating.

For smaller customers, cloud pricing is also an advantage. Both software and hardware pricing were once prohibitive for smaller customers. Now, with a subscription-based model, these customers can get a short, month-to-month subscription to the software and hardware they need to design a product prior to launch, without the initial capital expenditure.

With larger customers, there are a lot of great ideas for innovative products that once had to be shelved in favor of high priority projects due to resource constraints. Now, these customers can scale to the cloud to design, test and validate these ideas, helping them get started on their future-facing programs faster. For them, the only hang-up has been a fear that public clouds are somehow not secure enough to house their IP. What I have found is that public cloud options often provide better security than these customers’ own private datacenters.

 

Kurt: As a follow-up, how has the cloud increased productivity across your customers’ distributed engineering and design teams?

Sam: When collaborating with suppliers or vendors, our customer used to be very uncomfortable sending out designs, feeling that they were giving up their IP. Now, these collaborators are provided a single view with the data residing in a central location on a public cloud. This enables the vendors, engineers, analysts and OEMs to collaborate seamlessly without really moving the data around–effectively moving only pixels on screens. This helps solve for logistical challenges in collaboration while helping assuage concerns about data security.

Cloud computing has also enabled companies to collaborate internally. The cloud enables internal teams to collaborate seamlessly and instantaneously—with any changes made by one team becoming instantly visible to any teams affected by the change. Ultimately, this saves time that could be spent in long ponderous email chains, ensures there are no errors caused by data residing in multiple locations, and makes collaboration very visual by allowing engineers to work together within the tools, with which they are familiar. In addition, hosting data in the cloud provides security via access control mechanisms that ensure users can only access the data pertinent to their role.

 

Kurt: So now we are going to take a turn and talk a bit about Microsoft. What advantages does Microsoft Azure specifically offer as a public cloud option supporting simulation and other CAE workloads?

Sam: One of the first key requirements from a CAE standpoint is high performing compute resources, which is something that most cloud providers provide. But when we talk then about simulation workloads, or CAE workloads, one of the key things we also need is a high-performing, high-bandwidth network. I think Microsoft is the only cloud provider that has the Infiniband network which is an absolute must for running highly parallel CAE workloads. Additionally, for these workloads you need to be able to visualize the data without moving it from the cloud, requiring high performing graphics nodes with high performing graphics cards—both are components that Microsoft provides.

 

Kurt: From the business side, why did you choose to partner with Microsoft?

Sam: One of the biggest reasons was Microsoft’s market penetration and familiarity. Microsoft has agreements with many of our common customers in the automotive and aerospace markets, and we could see significant overlap between our customer bases. Microsoft is also a proven company for a number of these vendors; collaborating with Microsoft was beneficial as they already had good relationships with Microsoft and Altair. Microsoft and Altair can jointly go to market and sell to these mutual customers.

 

Get started now

With Microsoft Azure Big Compute, manufacturers are already realizing the powerful performance, scale, and flexibility of Microsoft’s global hyper-scale cloud. Don’t let a lack of computing power or networking and storage infrastructure stand between your engineers and the next generation of products. Unlock the full value of Big Compute for your business.

Learn more about our Big Compute solution portfolio today.