Stanford Researchers Conduct First-ever Million-core CFD Run on Sequoia Super

Print Friendly, PDF & Email

Andrew Myers writes that Stanford researchers using the Sequoia IBM Bluegene/Q system at LLNL have set a new record, harnessing a million compute cores to model supersonic jet noise.

These runs represent at least an order-of-magnitude increase in computational power over the largest simulations performed at the Center for Turbulence Research previously,” said Joseph Nichols, a research associate in the center. “The implications for predictive science are mind-boggling.” Sequoia once topped list of the world’s most powerful supercomputers, boasting 1,572,864 compute cores and 1.6 petabytes of memory connected by a high-speed five-dimensional torus interconnect. Because of Sequoia’s impressive numbers of cores, Nichols was able to show for the first time that million-core fluid dynamics simulations are possible—and also to contribute to research aimed at designing quieter aircraft engines.”

In addition to jet noise simulations, Stanford researchers in the DoE-sponsored Predictive Science Academic Alliance Program are using the CharLES code to investigate advanced-concept scramjet propulsion systems used in hypersonic flight and to simulate the turbulent flow over an entire airplane wing. Read the Full Story.

In related news, the HPC Advisory Council will host the Stanford HPC Conference on Feb. 7-8, 2013.