Crunching data with a million cores

In what is a record breaking achievement, researchers at Stanford University have just leveraged a million cores from the Sequoia IBM Bluegene/Q supercomputer to simulate the amount of noise produced by an experimental jet engine – one that they hope will allow for much quieter jet propulsion technology.

While there are technically over 1.5 million cores within the BlueGene supercomputer, this is still the first time that over a million has been used simultaneously for a single task. The particular code used had been tested on systems with 200,000 cores before, according to Wired, but never on such a large number. Because of this, the outcome of the experiment was up in the air until the moment of truth, with lots of potential for it to crash mid crunch. Fortunately for the record books, it didn't.

Researchers now believe they will be able to build a prototype for a much quieter jet engine, as well as allow others to make use of its mass computational software. Theoretically it could also be used to calculate other results, such as the full flow of air over an aircraft – thereby making expensive wind tunnel usage a thing of the past. It could also make calculations for the way an aircraft might react at super high velocities, something that'd difficult to do without actually putting the craft in the air.

KitGuru Says: With the growth of open source, distributed computing systems, it's nice to see good old supercomputers still popping their heads up now and again. There's nothing that quite says Skynet like a supercomputer and while no one wants that Terminator future, it's still cool to see these computing behemoths stretching their legs.

Become a Patron!