ALCC programme awards 1 billion hours on OLCF resources
Past ALCC project recipients have contributed to scientific discovery in the studies of energy efficiency, physics, materials science, and computer science.
8 Aug 2017 Oak Ridge - Each year, projects at the Oak Ridge Leadership Computing Facility (OLCF) - a US Department of Energy (DOE) Office of Science User Facility located at DOE's Oak Ridge National Laboratory (ORNL) - are awarded computing hours through the DOE Office of Advanced Scientific Computing Research Leadership Computing Challenge (ALCC). The projects selected are aligned with ALCC's mission to provide computing resources to high-risk, high-return simulations within energy-related areas of study.
The ALCC programme provides high-performance computing resources such as America's fastest supercomputer, Titan, to projects that align with DOE's energy mission. The programme allocates up to 30 percent of the computational resources at the OLCF and the Argonne Leadership Computing Facility, as well as up to 10 percent at the National Energy Research Scientific Computing Center.
These 1 year awards assist scientists from industry, academia, and national laboratories who work to advance scientific and technological energy research. Past ALCC project recipients have contributed to scientific discovery in the studies of energy efficiency, physics, materials science, and computer science.
ALCC allocations for 2017 continue in the tradition of innovation and discovery with projects awards ranging from 2 million to 300 million processor hours. Specific awards on Titan went to projects such as the following:
- Particle Physics: A team led by Gabriel Perdue of the College of William & Mary has been allotted 110 million hours on Titan to increase understanding of neutrino masses. One of the fundamental particles that make up our universe, neutrinos are copiously produced in high-energy collisions, most notably, the Big Bang. By studying these tiny particles, scientists can shed light on how our universe was formed and further develop theory residing within quantum physics.
The DOE Particle Physics Project Prioritization Panel (P5) has highlighted the physics of neutrino masses as one of five "Science Drivers", or lines of inquiry that show great promise for discovery. Consistent with P5's push to promote understanding of the nature and ordering of neutrino masses, Gabriel Perdue and his team seek to establish whether neutrinos and their counterparts, antineutrinos, differ in oscillation.
Thanks to advances in computing hardware - in this case, Titan's GPUs - computers have surpassed humans in their ability to recognize patterns in high-energy event collisions. Gabriel Perdue's team is using Titan's deep learning capability to improve the reconstruction of high-energy events, allowing for more detailed and accurate analysis of neutrinos.
- Energy: More than one-quarter of the electrical power generated in the United States begins with the operation of gas turbines. These systems are an efficient and flexible source of electric power that can be operated on fuels such as natural gas, a fuel that is considered more environmentally friendly than traditional fossil fuels such as coal because if its lower carbon dioxide emissions. General Electric (GE), a world expert in industrial power generation technology and the world's largest supplier of gas turbines, is using Titan to model one of the more complex problems in operating such turbines - turbulence.
A team led by Gregory Laskowski of GE Aviation is using Titan to trace the intricate flow of air that moves through the energy-creating turbine blades. Even a small reduction in turbulence enabled through modelling can greatly increase operational efficiency, further saving fuel costs and reducing CO2 emissions.
Gregory Laskowski's team is using machine learning techniques on Titan to develop more affordable turbulence simulation methods to be used in future turbine designs.
- Exascale Computing: In 2016, DOE created the Exascale Computing Project (ECP) with the goal of developing an exascale system by the year 2021. "Exascale" refers to computing systems at least 50 times faster than the nation's most powerful supercomputers in use today. The project's work encompasses the development of an entire exascale ecosystem: applications, system software, hardware technologies, and architectures, along with critical workforce development.
The 2017 ALCC allocation for ECP awards time at all three of the Advanced Scientific Computing Research programme's user facilities: the OLCF, the National Energy Research Scientific Computing Center, and the Argonne Leadership Computing Facility. This allocation allows the project to support approximately 100 sub-projects dedicated to realizing exascale capability through new system software, programme models, and algorithms. The accomplishments of these sub-projects will ensure the meeting of development milestones of the ECP.