Hybrid supercomputers are opening new possibilities in research across the fields of science and engineering, and Oak Ridge National Laboratory (ORNL) has launched a new conference to discuss the advances made possible by extreme-scale hybrid supercomputers. The Accelerating Computational Science Symposium 2012 will take place March 28-30 in Washington, D.C. Continue reading
Japan’s K Computer is still the world’s most powerful supercomputer, according to the latest Top500 List.
The K Computer, installed at the RIKEN Advanced Institute for Computational Science (AICS) in Kobe, Japan, achieved 10.51 petaflop/s on the Linpack benchmark using 705,024 SPARC64 processing cores. The Japanese supercomputer maintained the top position thanks to a build-out that made it four times as powerful as the number two entrant, the Chinese Tianhe-1A system.
The K Computer is the first supercomputer to achieve 10 petaflop/s or 10 quadrillion calculations per second. The previous June Top500 list marked the first time that all of the top 10 achieved petaflop/s performance. Continue reading
When design engineers run a simulation in their favorite engineering software, massive amounts of number crunching occurs behind the scenes to simulate a particular event. Such simulation is critical to designers who can save time and costs by doing fewer real-world tests and more digital tests of their designs. But how do we know the simulations are accurate?
Let’s take a look at one example recently featured in ORNL Review. A team of mechanical engineers at Sandia National Laboratory was given 60 million processor hours this year on Oak Ridge Leadership Computing Facility‘s Jaguar supercomputer to conduct high-fidelity simulations of combustion in advanced engines.
The models they create are validated against benchmark experiments to simulate turbulent combustion at different scales. Once validated, the models can be used by design engineers, as the article explains:
These models are then used in engineering-grade simulations, which run on desktops and clusters to optimize designs of combustion devices using diverse fuels. Because industrial researchers must conduct thousands of calculations around a single parameter to optimize a part design, calculations need to be inexpensive.