Research employing Argonne Leadership Computing Facility supercomputers aims to improve knowledge of two-phase flows (for example, fuel in gaseous and liquid forms) and their combustion in gas turbines. This understanding will allow a better assessment of the impacts of new designs and fuels on the efficiency of helicopter and airplane engines. In this image: Fields of temperature and pressure are shown in a 330-millionelements simulation of a complete helicopter combustion chamber performed on the IBM Blue Gene/P at the ALCF. Image courtesy Pierre Wolf, Turbomeca and CERFACS.
at the U.S. Department of Energy’s (DOE) Argonne National Laboratory
have received part of a planned $25 million grant from the DOE Office of
Science to tackle the problem of extracting knowledge from massive data
work is part of the DOE’s newly established Scalable Data Management,
Analysis, and Visualization (SDAV) Institute. Researchers in Argonne’s
Mathematics and Computing Science division will receive a planned $3.4
million over five years for the research.
computing advances are enabling researchers to attack important
problems, from increasing the fuel efficiency of vehicles to making more
aerodynamic airplane wings. The result is a veritable “tsunami of
data.” Many simulations and experiments already generate petabytes of
data—a single petabyte is 2,000 times more data than you can fit on a
typical laptop—and they will soon be generating exabytes.
task of handling this data is overwhelming, forcing scientists to spend
much of their time developing special-purpose solutions to store,
access and manage the information,” said Robert Ross, Argonne computer
scientist and deputy director of the new institute. “The SDAV teams will
develop the necessary tools and software so that scientists can use
their time more effectively for scientific investigation and discovery.”
institute will address challenges in three areas. Data management
enables query of scientific datasets; data analysis provides techniques
for both in situ and postprocessing data analysis; and data
visualization includes tools for identifying and understanding features
in multiscale, multiphysics datasets.
SDAV Institute was announced at the White House on March 29 as part of a
new $200 million Big Data Research and Development Initiative. Funded
under the DOE Office of Science’s Scientific Discovery through Advanced
Computing (SciDAC) program, the institute is led by Lawrence Berkeley
National Laboratory (LBNL). In addition to Argonne and LBNL, four other
national laboratories, as well as seven universities and one
visualization software company, are participating in the collaboration.
make all this possible, we will actively work with applications teams,
assisting them with the tools and ensuring that our efforts meet the
high standards needed to ensure correctness and performance of the
scientists’ codes,” said Ross. “In turn, we will gain critical feedback
about scientists’ needs in addressing mission-critical challenges.”
to successful deployment and adoption of SDAV tools are close ties to
leading computational facilities. The institute includes partners from
the Argonne Leadership Computing Facility, the National Energy Research
Scientific Computing Center at LBNL and the Oak Ridge Leadership
Computing Facility at Oak Ridge National Laboratory, who are responsible
for installing the new technologies developed by the SDAV teams. All
three supercomputing facilities are supported by DOE’s Office of
Science. These partners will also inform SDAV team members of upcoming
system architectures, guiding development of SDAV tools to ensure that
they will be effective as new systems come online.
reach an even broader community, the SDAV team plans to hold tutorials
and workshops to gather information from other researchers and train
potential users. These efforts will be coordinated with leading
conferences and DOE computing facility activities.
combines the expertise from three successful SciDAC Centers and
Institutes: the SciDAC Scientific Data Management Center for Enabling
Technologies, the Visualization and Analytics Center for Enabling
Technologies and the Institute for Ultra-Scale Visualization.
successes in those earlier SciDAC programs provide the knowledge needed
to achieve breakthrough science in this data-rich era,” said Ross.
Source: Argonne National Laboratory