In other words, jet-engine design is such an important thing not to get wrong that aerospace companies are willing to spend a whole lot of money to get it right. As a result, this use case remains an important part of Cray's business, even today.
The first component in an aircraft jet engine is the compressor, which raises the pressure of the air flowing through the engine. The efficiency of the engine is strongly dependent on the quality of the air flow through the compressor. Improper flow patterns can cause compressor-blade vibrations that can cause premature engine failure. A good engine design requires large investments in testing because the flows that occur in compressors are not now adequately predictable. A goal of the center is to be able to predict these complex flows, predict the compressor performance under all possible operating conditions, and investigate flow-driven blade vibration problems, thereby dramatically reducing the need for costly testing.
This Trippy Image Shows How Supercomputers Helped Build Better Jet Engines
It's not a psychedelic desktop wallpaper from the future; it’s a compressor blade simulation from the past, produced by a Cray supercomputer.
Re-Exposure is an occasional Motherboard feature where we look back on delightful old tech photos from wire service archives. Most people have likely never seen a supercomputer in person, let alone sat on one, as the vintage Cray supercomputers memorably allowed people to do. But for those unfamiliar with these multi-million-dollar supercomputers, perhaps one of the most awesome and interesting use cases for them is perhaps jet engine design. Above, computer scientist John Aldag (a Cray employee) is shown in 1983, using a Cray machine to virtually test the compressor blades on the front of a jet for any inconsistencies that might show themselves. (Aldag had deep familiarity with graphics related to Cray machines, by the way, once helping to produce a paper titled "Requirements, Status and Plans for Graphics on Cray Computers" that he presented to the Cray user group in 1986.)
So why a use an extremely expensive Cray supercomputer for this? Why not an IBM PC?A 1997 press release from Stanford University, announcing the launch of Department of Energy-funded supercomputer simulations for aircraft engine design, breaks down the benefits of testing compressor blades using high-powered supercomputing technology such as Cray's: