InTheLoop | 05.24.2010
May 24, 2010
NERSC Puts Carver into Production, Dirac into Testing
A new computing system is in production at NERSC. Built on IBM iDataPlex technology, the new system is called Carver in honor of American scientist and inventor George Washington Carver. Carver replaces NERSC’s Opteron Jacquard cluster and IBM Power5 Bassi system, which were both decommissioned at the end of April. Carver has 3.5 times the theoretical peak performance capacity of Jacquard and Bassi combined. Read more.
NERSC has also installed an experimental, general-purpose graphics processing unit (GPU) computing testbed named Dirac in collaboration with the Computational Research Division at Berkeley Lab, using funding from the DOE/ASCR Computer Science Research Testbeds program. Dirac has the following objectives:
- Investigate applicability of GPUs to scientific calculation
- Gain experience with scheduling heterogeneous elements (such as GPUs) in a multi-user environment
- Investigate programming models (e.g., MPI + CUDA, OpenMP + CUDA, etc.) for multi-GPUs to test scalability of code
Simulations Reveal That Earth’s Silica Is Predominantly Superficial
Silica is one of the most common minerals on Earth. Not only does it make up two-thirds of our planet’s crust, it is also used to create a variety of materials, from glass to fiber optic cables. Yet new results generated by a team of physicists from Ohio State University, using NERSC supercomputers, show that this mineral only populates our planet superficially—in other words, silica is relatively uncommon deep within the Earth. Read more.
LBNL Scientists Build Software Framework for ATLAS Collaboration
Three thousand researchers in 37 countries are searching for the origins of mass, new dimensions of space, and undiscovered forces of physics in the head-on collisions of high-energy protons at the Large Hadron Collider’s ATLAS experiment. To facilitate this distributed workflow, these researchers are relying on a software framework called Athena, which was developed with the help of scientists in the Berkeley Lab’s CRD and Physics Divisions. Read more.
New Performance Management Process Meetings This Week
This year the NERSC and Computational Research divisions, along with 13 other scientific and operations divisions, including the Laboratory Directorate, will be piloting a new performance management process (PMP). Both Paul Alivisatos and Jim Krupnick would like this new process to be rolled out Laboratory-wide in 2011. There will be informational sessions held for CS employees at OSF (Room 238) at 3:00 pm on Tuesday, May 25; and in the Bldg. 50 Auditorium at 10:00 am on Wednesday, May 26, and at 11:00 am on Wednesday, June 2. Please plan to attend one of the sessions.
Last year, improving performance and career mentoring conversations was a definite theme in employee feedback to Division Directors across all Divisions. These conversations foster the kind of initiative and accomplishments that lead to discovery, innovation and operational efficiencies.
Annual performance review discussions are a critical part of these conversations. For the pilot, all non-represented CS employees (scientist and non-scientist) will participate in PMP’s annual performance review process and use its new ratings and forms. You’ll be asked to provide feedback on the process this fall. Our represented employees will continue to use PRD ratings and forms for their annual review process.
NERSC Staff Contribute to Cray User Group Conference
NERSC is well represented at CUG2010, the Cray User Group meeting being held May 24–27 at the University of Edinburgh, Scotland. Here are the presentations they will be giving:
- Jonathan Carter: “External Services on the Cray XT5 System Hopper” (with Katie Antypas and Tina Butler)
- Yun (Helen) He and Hwa-Chun Wendy Lin: “Franklin Job Completion Analysis” (with Woo-Sun Yang)
- Alice Koniges: “Application Acceleration on Current and Future Cray Platforms” (with Jihan Kim, Robert Preissl, and John Shalf, NERSC; and David Eder, Aaron Fisher, Nathan Masters, and Mlaker Velimir, Lawrence Livermore National Laboratory)
- Alice Koniges: “Analyzing the Effect of Different Programming Models upon Performance and Memory Usage on Cray XT5 Platforms” (with John Shalf, Hongzhang Shan, and Nick Wright, NERSC; Haoqiang Jin, NAS Systems Division; and Seung-Jai Min, CRD)
- Andrew Uselton: “File System Monitoring as a Window into User I/O Requirements” (with Katie Antypas, NERSC; Daniela Ushizima, CRD; and Jefferey Sukharev, University of California, Davis)
Nick Cardo is program chair and vice president of the CUG Board of Directors, and Jim Craw is chair of the XTreme Systems Special Interest Group.
ACTS Workshop Will Be Held August 17–20 in Berkeley
The 11th Workshop on the DOE Advanced Computational Software (ACTS) Collection, “High Performance Software Tools to Fast-Track the Development of Scalable and Sustainable Applications,” will be held in Berkeley, California, August 17-20, 2010.
The four-day workshop, organized by Lawrence Berkeley National Laboratory, will present an introduction to the DOE ACTS Collection for application scientists whose research demands include either large amounts of computation, the use of robust numerical algorithms, or combinations of these. The workshop will include a range of tutorials on the tools currently available in the collection, discussion sessions aimed to solve specific computational needs by the workshop participants, and hands-on practice using state-of-the-art supercomputers at NERSC. Presenters are tool developers from DOE national laboratories.
This workshop is open to computational scientists from industry and academia. Registration fees are fully sponsored by the DOE Office of Science. In addition, DOE will sponsor travel expenses and lodging for a limited number of graduate students and postdoctoral fellows. The deadline for registration is June 28. For more information on the workshop, please contact Tony Drummond at (510) 486-7624.
This Week’s Computing Sciences Seminars
OSF Brown Bag: Analytics using Aster Data’s nCluster
Tuesday, May 25, 11:30 am–1:00 pm, OSF-254
Jonathan Goldman, Aster Data
Aster Data is a leader in big data management and big data analysis for data-driven applications. Aster Data’s nCluster is the first MPP data warehouse architecture that allows applications to be fully embedded within the database engine to enable ultra-fast, deep analysis of massive data sets.
Jonathan Goldman, Director of Analytics and Applications, will present the overall architecture of Aster Data and discuss the vision for embedding advanced analytics within Aster Data’s nCluster. He’ll draw from some examples developed while at LinkedIn. He’ll discuss Map Reduce applications that can be run on the Aster system. The goal is to make this an informal and open discussion and learn more about the challenges facing the real-time classification of massive data streams.
Link of the Week: The Hilbert Hotel
Can a hotel with an infinite number of rooms accommodate an infinite number of guests? The answer depends on whether infinity comes in different sizes. Steven Strogatz contemplates this meeting planner’s nightmare in his New York Times math column “The Hilbert Hotel.” Watch out for Georg Cantor and David Hilbert arguing with Henri Poincaré in the lobby.
About Computing Sciences at Berkeley Lab
The Lawrence Berkeley National Laboratory (Berkeley Lab) Computing Sciences organization provides the computing and networking resources and expertise critical to advancing the Department of Energy's research missions: developing new energy sources, improving energy efficiency, developing new materials and increasing our understanding of ourselves, our world and our universe.
ESnet, the Energy Sciences Network, provides the high-bandwidth, reliable connections that link scientists at 40 DOE research sites to each other and to experimental facilities and supercomputing centers around the country. The National Energy Research Scientific Computing Center (NERSC) powers the discoveries of 6,000 scientists at national laboratories and universities, including those at Berkeley Lab's Computational Research Division (CRD). CRD conducts research and development in mathematical modeling and simulation, algorithm design, data storage, management and analysis, computer system architecture and high-performance software implementation. NERSC and ESnet are DOE Office of Science User Facilities.
Lawrence Berkeley National Laboratory addresses the world's most urgent scientific challenges by advancing sustainable energy, protecting human health, creating new materials, and revealing the origin and fate of the universe. Founded in 1931, Berkeley Lab's scientific expertise has been recognized with 13 Nobel prizes. The University of California manages Berkeley Lab for the DOE’s Office of Science.
DOE’s Office of Science is the single largest supporter of basic research in the physical sciences in the United States, and is working to address some of the most pressing challenges of our time. For more information, please visit science.energy.gov.