Supercomputer exceeding 1 million Cores performs Simulation
Researchers recently set a new precedent in the computer science community by using a supercomputer to determine the noise output of a jet engine traveling at supersonic speeds. This unique computer harnesses the power of over 1.5 million processors, and not only did the research performed garner important data that will help engineers make less noisy aircraft engines, but it also demonstrated the incredible possibilities of using million plus core computers to create fluid modules.
During takeoff and landing, high powered aircraft produce more noise than just about any other man-made source. This makes hearing damage a real threat for ground crews, even if they are wearing state-of-the-art hearing protection devices. Furthermore, such noise pollution also affects property value in surrounding communities, and is aggravating to those who have to deal with it on a regular basis.
It therefore stands to reason that engineers are concerned with quieting these engines, and one method they are looking into is the shape of the nozzles. To help with this, computers are being used to simulate the noise production of various designs. Computers with thousands upon thousands of cores break down computation into smaller amounts, in order to allow different parts of the simulation to be worked on simultaneously, resulting in faster data production.
So why not just jump straight to millions of cores and get the job done even faster? Unfortunately, at some point the number of processors starts to present other issues that actually slow things down. That’s why this latest research was so important – not just to engineers but to computer scientists too.
As with other massive supercomputers, this one has been dubbed with a name – Sequoia. Once the premier supercomputer in the world, this IBM machine earned its name from the massive redwood trees of California – now it is being used to study jet noise and other propulsion systems.