DOE awards 6 billion core hours for HPC research
The US Department of Energy’s Office of Science has announced that new 55 projects have been awarded time on HPC resources at Argonne and Oak Ridge National Laboratories through its Innovative and Novel Computational Impact on Theory and Experiment (INCITE) programme.
The projects will share 5.78 billion core-hours and support a broad range of large-scale research topics in areas such as sustainable energy technologies and next-generation aerospace designs.
Michael Papka, director of the Argonne Leadership Computing Facility (ALCF), a DOE Office of Science User Facility located just outside Chicago commented: ‘These computational projects represent the biggest challenges in science and engineering today, and can’t be done anywhere else.’ Papka stressed that it is not only the computing time but access to the national laboratories staff expertise which can be crucial to a project’s success.
The INCITE program is the primary means of accessing the DOE Leadership Computing Facilities at Argonne and Oak Ridge National Laboratories and is jointly managed by the two centres. Researchers from academia, government research facilities, and industry can submit INCITE project proposals, which are awarded on computing hours on a competitive basis. The average award is more than 105 million core-hours – with some awards of up to several hundred million core-hours – on systems capable of quadrillions of calculations per second.
However, the INCITE programme does not only provide computing hours on some of the largest supercomputers in the US. The programme also enables the allocation of domain and computational scientists at the centres to partner with each INCITE project. This helps to accelerate code and method development, optimisation, troubleshooting, and assisting with data analysis and visualisation.
Many of the project teams are building on findings from previous INCITE awards in their pursuit of significant scientific breakthroughs. DOE's world-class capabilities are also attracting new users. Fifteen of the 2017 INCITE projects are headed by principal investigators who have not previously received an INCITE award.
‘Once again this year, we’re delighted to have this opportunity to play a key role in enabling breakthrough science through the 2017 INCITE program,’ said James Hack, director of the National Center for Computational Sciences (NCCS), home to the Oak Ridge Leadership Computing Facility (OLCF), a DOE Office of Science User Facility in Oak Ridge, Tennessee. ‘Our user community continues to propose ever more ambitious and computationally demanding campaigns. Helping that community fully exploit the unique capabilities of our facilities is always a welcome challenge.’
The OLCF’s Titan supercomputer is a 27-petaflops Cray XK7 hybrid system employing both CPUs and GPUs in its 18,688 compute nodes. The ALCF’s Mira supercomputer is a 10-petaflops IBM Blue Gene/Q system with 49,152 compute nodes and a power-efficient architecture.
Despite continued upgrades, expansions, and advances in computing power, demand for leadership-class resources such as Mira and Titan continues to exceed availability, and, once again, more applications for time were made to INCITE than were awarded.
This year computing time was awarded Gorazd Medic of United Technologies Research Center which was awarded 35 million core-hours to assess high-order hybridized discontinuous Galerkin methods for large eddy simulation of complex turbulent flows in turbomachinery.
Another successful project came from Jean-Luc Vay of Lawrence Berkeley National Laboratory received 100 million core-hours to study ultra-high intensity lasers that will enable accurate, realistic modelling of plasma mirrors and ion acceleration.
A full list of the successful applications to the INCITE programme can be found on the DOE leadership computing website.
The INCITE program promotes transformational advances in science and technology through large allocations of time on state-of-the-art supercomputers. For more information, please visit.