IMPACT May 2013
What's New in XSEDE
"What's New in XSEDE" is a monthly e-newsletter providing information on scientific discoveries made possible by the Extreme Science and Engineering Discovery Environment, as well as the people, places, and programs involved. XSEDE is a five-year high-performance computing project supported by the National Science Foundation.
Crunching Large Hadron Collider Data
 |
| This image of a supersymmetry event shows the transverse momentum imbalance due to dark matter particles escaping the detector (direction indicated by red arrow). Red and blue rectangles indicate energy deposited in the electromagnetic and hadronic calorimeter respectively; green tracks in the center show charged particles with transverse momentum larger than 2GeV. Yellow-outlined triangles indicate jet cones or the presence of subatomic particles called quarks. Image courtesy of Matevz Tadel, UC San Diego/CMS |
The XSEDE-allocated Gordon supercomputer at San Diego Supercomputer Center (SDSC) has been working on one of its most data-intensive tasks to date: rapidly processing raw data from almost 1 billion particle collisions as part of a project to help define the future research agenda for the Large Hadron Collider (LHC). Under a partnership between a team of UC San Diego physicists and the Open Science Grid, Gordon has been providing auxiliary computing capacity by processing massive data sets generated by the Compact Muon Solenoid (CMS), one of two large general-purpose particle detectors at the LHC used by researchers to find the elusive Higgs particle, a discovery that has garnered international attention.
While the CMS has significant compute resources, partnering with XSEDE to incorporate Gordon into the CMS workflow allowed analysis of the parked data to be completed months ahead of time, letting scientists review the results immediately to help provide input for future plans for the LHC. The around-the-clock data processing run on this XSEDE-allocated resource was completed in about four weeks' time, making the data available for analysis several months ahead of schedule. About 1.7 million core hours—or about 15 percent of Gordon's total compute capacity—were dedicated to this task, with more than 125 terabytes of data streaming through the supercomputer's nodes and into the XSEDE-allocated Data Oasis storage system at SDSC for further analysis.
Read more about XSEDE's role in crunching LHC data
Cleaning Up Air and Energy Technologies
Chemists at the University of South Florida and King Abdullah University of Science and Technology have discovered a more efficient, less expensive and reusable material for carbon dioxide (CO2) capture and separation. With the assistance of XSEDE-allocated resources, the international group of scientists has identified a previously underused material known as SIFSIX-1-Cu. The discovery addresses one the biggest challenges of capturing CO2 before it enters the atmosphere: energy costs associated with the separation and purification of industrial commodities currently consumes around 15 percent of global energy production. Also, water normally interferes with CO2 capture, but the SIFSIX-1-Cu resists it, making it a promising candidate for real-world applications.
Predicting the exact behavior of even small numbers of molecules requires a huge amount of computer memory—more than one terabyte—and such calculations are a specialty of the XSEDE-allocated Blacklight supercomputer at Pittsburgh Supercomputing Center, the largest "shared memory" computer in the world. Porous SIFSIX materials are built from combinations of inorganic and organic chemical building blocks and are part of a general class of materials known as Metal-Organic Materials (MOMs). The researchers used the results from Blacklight to simulate the behavior of the gasses and the MOMs in bulk on the XSEDE-allocated Ranger supercomputer at the Texas Advanced Computing Center and Trestles—a dedicated XSEDE cluster—at the San Diego Supercomputer Center.
The group believes the previously underused material has three potentially significant applications: carbon-capture for coal-burning energy plants; purification of methane in natural gas wells; and the advancement of clean-coal technology. Some 20 to 30 percent of the power output at a clean-coal plant is consumed by cleaning process. The new material could make those plants more efficient and put more power into the grid, the scientists predict.
Read more about what this discovery means for CO2 capture and separation
Swirling Secrets
 |
| Diagram of a protoplanetary disk around a young star. Angular momentum is transported outward and mass is transported inward and onto the star via disk turbulence. This turbulence is studied using local simulations of a small patch of disk (the 3D box on the lower right; this shows the gas density in the turbulent state). The presence of an external magnetic field similar to the one shown enhances this turbulence in the outer disk regions. |
Many newly formed stars are surrounded by swirling masses of warm dust and gas called protoplanetary disks. These disks may become celestial bodies such as planets and asteroids. But how they make that transformation will remain a mystery until researchers can get a grasp on the disordered movement, or turbulence, that characterizes the constituent gases of the disks. Jake Simon of the University of Colorado has turned to XSEDE to face two primary challenges in this research. XSEDE is aiding in the quest to understand protoplanetary disk turbulence through both developing correct models for the simulations and designing algorithms that capture the nature of the Hall effect—which refers to a voltage difference that occurs across an electrical conductor.
Simon and his team have used more than 4 million service units (compute hours) on the XSEDE-allocated Kraken supercomputer at Oak Ridge National Laboratory, including an average of approximately 585 cores per run and a single-run high of 18,432 cores. The research team now understands ambipolar diffusion, in which electrons and ions in protoplanetary disks are dragged along by a magnetic field. "If the ions and electrons don't collide with the neutrals frequently enough, ambipolar diffusion acts to damp out the turbulence," Simon explains. "The degree to which this happens has been explored with our high-resolution numerical simulations that we have run on the Kraken supercomputer. We believe we now have a much better understanding of how disks behave in their outer regions, far from the central star."
Learn more about the turbulence of gases in planet-forming protoplanetary disks
An Earth-shaking Computing Breakthrough
 |
| The image shows a snapshot of ground motion of the 2008 magnitude-5.4 Chino Hills earthquake in an east-to-west direction; the red-yellow and green-blue colors depict the amplitude of shaking. The simulation indicates that small-scale heterogeneities (causing the highly irregular pattern of shaking in the image) may significantly affect ground motion in geologic basins. Simulation by Efecan Poyraz/UC San Diego and Kim Olsen/San Diego State University. Visualization by Efecan Poyraz; map image courtesy of Google. |
With the help of XSEDE-allocated resources a team of researchers at the San Diego Supercomputer Center (SDSC) and the Department of Electronic and Computer Engineering at UC San Diego has developed a highly scalable computer code that promises to dramatically cut both research times and energy costs in simulating seismic hazards. The team, led by SDSC computational scientist Yifeng Cui, performed GPU-based benchmark simulations of the 5.4 magnitude earthquake that occurred in July 2008 below Chino Hills, near Los Angeles. Compute systems used in the project included Keeneland, an XSEDE-allocated resource managed by Georgia Tech, Oak Ridge National Laboratory (ORNL), and the National Institute for Computational Sciences (NICS).
By delivering a significantly higher level of computational power, researchers can provide more accurate earthquake predictions with increased physical reality and resolution, with the potential of saving lives and minimizing property damage. The project is part of a larger computational effort coordinated by the Southern California Earthquake Center (SCEC). A technical paper based on this work will be presented June 5-7 at the 2013 International Conference on Computational Science in Barcelona, Spain.
XSEDE13 schedule now available online
The schedule of tutorials, talks, birds-of-a-feather meetings, and more for XSEDE13 is now available online. XSEDE13, the annual conference focused on science, education, outreach, software, and technology related to XSEDE, will bring together hundreds of scientists, technology experts, educators, and students July 22-25, 2013, at the Marriott Marquis & Marina in San Diego.
Among the new activities planned for XSEDE13 is a weeklong robotics camp, on-site at the conference hotel, for students in grades 5-8. Ready, Set, Robotics! is offered as an option for attendees traveling with their children. There also are many off-site camps for children within a few miles of the hotel, including the San Diego Zoo, San Diego Children's Museum, and many museum camps in Balboa Park. XSEDE13 is presented in cooperation with ACM SIGAPP, the Association for Computing Machinery's Special Interest Group on Applied Computing, and is supported by corporate and non-profit sponsors.
Read more about the schedule for XSEDE13
XSEDE in a nutshell
Following are events, deadlines and opportunities related to XSEDE.
Extreme Scaling Workshop 2013 Call for Presentations due May 15. Workshop Aug. 15-16 in Boulder, CO