In the world of engineering and design, high-performance computing (HPC) has been used to solve complex tasks with ease. The popularity of HPCs has increased significantly, but many are still struggling to figure out the right configurations. In addition, the question often arises as to when and how they should be used. This struggle is partly due to the fact that HPC is confused with traditional supercomputers and the actual concept behind them. This article focuses on describing the concept of so-called HPC. By the end of the article you will have gained the following insights:
What is HPC?
As with most relatively new technical concepts, it takes some time for experts to reach consensus on definitions. This is also the case with HPCs, but there is still a basic general idea of HPC. In short, HPC systems can be defined as the pooling of computing resources to find solutions to the complex problems that a personal computer cannot solve.
This general definition is the reason why many still claim to define all HPC systems as supercomputers, but this is fundamentally wrong. The fact that personal computers can be integrated into a cluster to provide high performance computing capabilities nullifies this argument. While supercomputers can provide targeted HPC services that make them a niche example of an HPC system, not all HPC systems are supercomputers. Cray supercomputers are a perfect example of this phenomenon.
The early Cray computers of two to three decades ago were exclusively stand-alone supercomputers. These computers were equipped with custom functions required for handling complex simulations and calculations, which made them niche HPC solutions. Today, Cray computers have chosen HPC and now offer a cluster of supercomputers that use networked technologies to provide high-performance computing.
Another important difference to consider when evaluating supercomputers as HPCs is cost and customization. Most supercomputers are designed to run custom software applications and the cost of acquiring and operating one is in the millions. As a result, every supercomputer is inaccessible to 90% of the world’s companies. HPCs, on the other hand, are interconnected computer systems that use older software and are affordable. This means that the average HPC system consists of thousands of high-end personal computers that are used to solve complex problems.
How powerful computer systems work.
Every computer in an HPC system is called a node. Each node is generally equipped with several processors, called computing cores, which handle the computational aspect of solving problems. The processors, graphics processing units, and memory of each node are then interconnected through a high-performance computing network.
If you are familiar with 3D printing, you can compare HPC ecosystems to a 3D print farm or cluster. Here, multiple devices work together to achieve much greater performance or better results. If several people intend to use the HPC system, a scheduler comes into play. The scheduler allocates resources to multiple users and intuitively scales the allocations according to a user’s needs. In terms of storage, HPC projects require a lot of storage space. This is one of the reasons why cloud-based HPCs were developed.
Why is HPC needed?
In the early days of Bitcoin mining, HPC systems were used, which helped to make the concept known to SMEs. Today HPCs are mainly used in brick and mortar plants for some serious problem solving. In engineering, HPCs are used to handle complex simulations. These simulations are complex modelling phenomena that are small and transient, such as acoustics or combustion phenomena in engines.
In multiphysical simulations, the simultaneous effect of several (natural or artificial) phenomena on a product or structural design is investigated. An example is the analysis of how the effects of thermal stress affect a moving structure, as well as the turbulence caused by the moving structure at certain points in time. Here, paired partial differentiation and other complex mathematical models must be calculated to solve the problem. HPCs can accelerate the process and drastically eliminate errors that could affect the functionality of the structures.
In industrial design and manufacturing, HPCs have been used to solve structural and thermal design processes to optimize the production life cycle. One example was Rolls Royce’s attempt to eliminate the use of temperature sensors in its prototype engines for future cars. To accomplish this task, Rolls Royce intended to apply an interactive process to determine the heat flow and the effects of the removal of thermal sensors on the engine structure. The power plant giant turned to ANSYS Fluent and HPC Pack. Fluent’s CFD (Computational Fluid Dynamics) solver was used to determine the heat flow and CPU 24/7 HPC handled the computational workload.
In the end Rolly-Royce drastically reduced the time required for a coupled CFD structure simulation by 80 percent. Other benefits included avoiding costly design changes during prototype production and successfully developing a way to eliminate thermal sensors in its engines.
An example of generative design within an HPC system is the attempt by Airbus to revolutionize commercial aircraft. The idea was to develop a concept aircraft with reduced weight and an enlarged compartment structure without losing the strength of its commercial aircraft. To achieve this manufacturing goal, millions of variables and structural interactions were modeled simultaneously to obtain an optimal product. The scale of this computational task led Airbus to High Performance Computing. According to Gerd Buttner, HPC helps Airbus to redesign the future of flight.
HPCs can also be used to manage a complex manufacturing process that requires multiple industrial licenses throughout the production process. To achieve this, a supportive systematic approach to licensing must be adopted. The ANSYS Parametric Pack is such a solution. The Parametric Pack allows the simultaneous analysis of different design points using a number of licenses at an affordable price. This speeds up the production process and reduces the time needed to analyze and record multiple analyses.
Challenging popular misconceptions about HPC.
HPCs have been used predominantly by large companies in recent years, which has led to various misunderstandings or myths about their use or applicability. One of the misconceptions currently circulating is that these systems can only be attributed to highly complex simulations and the complexity of the underlying equations. However, HPCs are also used for structural and electromagnetic simulations.
Another myth that revolves around its use for SMEs is the cost. Most project managers and CTOs still believe that HPCs are too expensive and have a low ROI, but the facts say otherwise. With the advent of cloud-based HPCs, high costs cannot be an issue. Vendors like ANSYS and Microsoft are now offering scalable pricing models at lower cost for complex engineering tasks. This means that companies can choose to pay only for what they use and expand as needed.
The last myth is about usability. Many believe that a dedicated IT supervisor or manager is needed to handle HPC-related activities. Today, there are currently managed HPC systems that you can use.
Understanding the options.
The HPC field is growing rapidly and is becoming more user-friendly due to the various vendors offering integration services. These vendors offer integrated solutions that enable the deployment of HPCs in existing environments. The integration process can be done either via parallel file systems or a cluster management solution. Cloud computing also offers a more affordable way for businesses to take advantage of HPC services. The benefits of HPC cloud-based services include a scalable, on-demand and cost-effective option for running sophisticated simulations. As a result, around 64 percent of HPC platforms now integrate cloud computing and the market is expected to grow by 12 percent annually over the next 5 years.
Today, ANSYS cooperates with many HPC providers who offer optimized HPC services to SMEs and large enterprises.
These vendors also offer post-sale services to ensure that integration and deployment tasks are performed properly.
Summary.
The future of engineering and manufacturing may lead to an increased use of HPC. If Moore’s Law continues to apply, this future can be defined by small companies and entrepreneurs who want to take advantage of the potential offered by a cluster of smart devices. As more and more vendors expand their HPC as a service offerings, either on-premise or in the cloud, it is becoming accessible and affordable for everyone.
Thank you very much for your visit.