Five researchers at the
Department of Energy's Argonne National Laboratory will lead projects that
have been awarded a total of 80 million hours of computing time on Argonne's
energy-efficient Blue Gene/P. Using the computer allotments, the researchers
will conduct advanced simulation and analysis and will develop scalable system
software needed to fully utilize the power of leadership-class computing facilities.
Coolant-flow pressure distribution in a 217-pin wire-wrapped subassembly, computed on the Argonne Leadership Computing Facility's Blue Gene/P supercomputer. Image by members of the Argonne SHARP project, Paul Fischer, Aleks Obabko, Andrew Siegel, Dave Pointer and Jeff Smith.
The projects, selected competitively based on their potential to advance scientific
discovery, were awarded the supercomputer time through the U.S. Department of
Energy (DOE) Innovative and Novel Computational Impact on Theory and Experiment
(INCITE) program. Four of the projects are new, and one is a renewal.
Paul Fischer, a senior computational scientist, was awarded 30 million hours
on Argonne's Blue Gene/P as well as additional time at Oak Ridge National
Laboratory to conduct simulation and analysis of fast neutron reactors. Fischer
and his colleagues will study open questions concerning the thermal-hydraulic
performance – that is, the pressure drop and mixing induced by the coolant
flow – in these next-generation reactors. Thermal-hydraulic performance
issues figure prominently in understanding how to design a “closed”
nuclear fuel cycle that reduces the burden of geological disposal of nuclear
spent fuel. “Advanced simulation is viewed as critical in bringing fast
reactor technology to fruition in an economic and timely manner,” said
Andrew Binkowski, a structural biologist, will lead an Argonne team in a study
of protein-ligand binding interactions. Binkowski and his team will use the
25 million hours of computer time awarded on the Blue Gene/P to apply the most
advanced methods in biomolecular simulations and analysis to further understanding
of human health and disease. “A major obstacle to accurate biomolecular
modeling is the number of approximations necessary to reduce the complexity
of the system. Using the BlueGene/P will allow us to relax these constraints
and perform calculations in a matter of hours that would previously have taken
years,” said Binkowski.
The team will also evaluate and validate the predictive power of bimolecular
simulations through collaboration with the Center for Structural Genomics of
Jeff Greeley, a materials scientist, was awarded 10 million hours of supercomputing
time on Argonne's Blue Gene/P to investigate materials at the nanoscale.
(A nanometer is one billionth of a meter). Greeley will lead a collaboration
to understand the electronic and chemical properties of metal particles across
the nanoscale regime.
“We expect insights gained from having access to Argonne's supercomputing
resources to open significant new opportunities for manipulating and selecting
nanoparticles for groundbreaking technological applications,” Greeley
Ewing (“Rusty”) Lusk, director of Argonne's Mathematics and
Computer Science Division, was awarded 5 million processor hours on the Blue
Gene/P to improve the performance and productivity of key system software components.
Lusk will head a team investigating message-passing libraries, parallel input/output,
data visualization and operating systems on the leadership-class computer systems.
“Through rigorous experimentation, analysis and design cycles, we hope
to dramatically enhance the capabilities of these systems,” said Lusk.
Andrew Siegel, a computational scientist and leader of Argonne's nuclear
simulation activities, was awarded 10 million hours to continue his numerical
experiments of “thermal striping” – temperature fluctuations
generated by the mixing of flows in sodium-cooled fast reactors. Siegel is collaborating
with Argonne's nuclear engineers to study these fluctuations, which can
cause components to crack.
“Traditionally, designers have relied on data from instrumented experiments,
but this data is expensive and difficult to collect,” said Siegel. “Computation
has also been used, but with reduced models that greatly limit predictability.
With the INCITE resources, we can carry out the first detailed numerical experiments
of thermal striping on realistic reactor geometries.”
The 163,840-processor Blue Gene/P at Argonne has a peak performance of 557
teraflops. “By providing millions of hours of computing time on this powerful
system, the DOE INCITE awards allow us to address some of the nation's
most challenging scientific problems,” said Rick Stevens, associate laboratory
director of Computing, Environment, and Life Sciences at Argonne. “Argonne
researchers will be attacking these problems on two major fronts: simulation
of advanced technologies and development of the key system software needed to
take complete advantage of our nation's powerful leadership-class computers.”
The INCITE program was established by the U.S. Department of Energy Office
of Science seven years ago to support computationally intensive, large-scale
research projects. INCITE awards sizeable allocations (typically, millions of
processor-hours per project) on some of the world's most powerful supercomputers
at DOE's Leadership Computing Facilities at Argonne National Laboratory
in Illinois and Oak Ridge National Laboratory in Tennessee.