High Performance Computing: Pursuing Innovation In Tough Times

By David Graff, U.S. industry solutions director, Microsoft Corp. Manufacturers can innovate during tough times by using high performance computing technology to simulate real-world scenarios and products -- saving materials and labor costs.

In today’s widespread economic downturn, manufacturing companies are in a delicate balancing act between reducing costs and maintaining their competitive edge. In times like these, a short-term focus on cutting costs naturally can take priority. But it’s important for manufacturers and their partners to look beyond the here and now and determine a strategy for long-term success. That strategy requires both cost-savings and innovation.

The crisis facing the automotive sector helps illustrate the rationale behind such a strategy. Automakers can use innovation to position themselves as leaders when the economy improves, while maximizing projects in the near term that generate cost savings and cash.

The Role of High Performance Computing

One way in which manufacturers can innovate during tough times is by using high performance computing (HPC) technology to simulate real-world scenarios and products, saving materials and labor costs. An HPC cluster contains a group of connected computers that simultaneously work on a job, delivering results more quickly and cost effectively than a standalone system.

Once confined to CIA code-breaking and Pentagon war games, HPC is now used by everyday manufacturers to save costs, drive major improvements in products, increase production efficiency and speed to market, and better meet changing customer needs.

HPC, which solves complex computational problems that are too difficult or time-consuming to compute with traditional computers, is emerging as a key enabler of innovation for manufacturers.

HPC breaks problems into smaller, manageable pieces and sends them to computers that reassemble the pieces and form a solution. Naturally, distributing the problem among multiple systems leads to faster problem solving. Industries as diverse as rail transportation, automotive manufacturing and aerospace are using HPC to replace or supplement physical experiments with computer-simulated modeling, tests and analysis. Moving from a physical to a virtual environment reduces cost and speeds time to market.

Reducing Costs

With HPC, problems that would take months using traditional computers can be reduced to days or even hours, resulting in enormous cost savings and increased productivity. In the automotive and transportation industries, for example, crash-testing real vehicles to determine safety characteristics is extremely expensive or even near impossible. Every time an automotive manufacturer or supplier crash-tests a real vehicle, the test run can total $250,000 or more.

Now imagine physically crash-testing high-speed trains. Determining potential problems by simulating crashes via computer before or in place of crashing real rail-vehicles is obviously more efficient and manageable, and HPC is an invaluable tool in reducing the time needed to complete such simulations.

Let’s look at how HPC benefited Bombardier Transportation, a global leader in the rail equipment manufacturing and servicing industry: The company was using computers to perform crash analysis of rail-vehicles when additional safety requirements emerged during the middle of a project, increasing the company’s need for more powerful computing. One requirement, a test and analysis of a head-on collision between two trains, would have taken more than a month to complete with the company’s hardware and software. With HPC, Bombardier Transportation completed the month-long simulation in only four or five days. HPC allowed the company to cut costs and deliver the project on time. And now the company is prepared to meet safety regulations more efficiently.

Improving Design and Production

While crash-test simulations may seem like an obvious use of HPC, the technology also can play an important role in everything from creating factory layouts to assembly sequencing to product design and production. Automotive designers and engineers can use HPC to calculate the airflow inside a vehicle to determine improvements in heating, ventilation and air-conditioning (HVAC) systems. An automaker may discover, for example, that the HVAC system is not cooling sufficiently, but that shaving one-sixteenth of an inch from the inside of the ductwork would make the cooling much more efficient.

Through HPC, engineers also can examine and minimize the variations that occur in designs, such as slight differences in dimensions or weight, and determine how a change in one system may affect the performance of other systems. Automakers can perfect designs and more readily integrate customer preferences on initial tries instead of analyzing a design after a product is produced, or even shipped, thus improving products and reducing recalls. With HPC, automakers can more accurately predict when and how a particular vehicle part will fail just from normal use, allowing them to produce safer vehicles. Manufacturers can complete designs faster, and deliver high-quality products with shorter lead times, giving them a competitive edge.

Meeting Customer Needs

Another part of maintaining a competitive advantage is being able to address changing customer needs. Companies like aerospace engineering firm a.i. solutions, Inc., are finding successful ways to do so with HPC.

When NASA requested that a.i. solutions analyze the potential impact of thousands of pieces of space debris on satellites in the same orbit over a 20-year period, a.i. solutions turned to HPC. With HPC software, the company completed the analysis in less than three days, instead of the month it would have taken with traditional software. The ability of the company’s analysts to run engineering simulations up to eight times faster than on the previous technology allows the company to look at many more possibilities than before, get faster time to insight and solve customers’ problems more quickly.

While the economic crisis facing manufacturers is a complicated issue, HPC is one cost-effective method that can help companies pursue innovation in tough times. With greater computing capabilities and reduced testing times, engineers and designers across manufacturing industries have greater flexibility to try new, creative ideas and methods, and save on costs. Finding new ways to cut expenses, improve products, increase efficiency and explore new ideas through HPC will differentiate manufacturers from competitors that maintain the status quo. Companies that continue to drive innovation and adopt technologies such as HPC will be better poised to manage the current economic crisis and prepare for future success.

For more information on Microsoft’s high performance computing solutions, visit http://www.microsoft.com/hpc