InTheLoop | 08.08.2011
August 8, 2011
NPR, NY Times Tap Expertise of Shalf and Patterson
Two Computing Sciences researchers recently talked to the national news media about computer chips’ unsustainable growth in energy use and the future of computing. NERSC’s John Shalf was interviewed by National Public Radio for the story “Curbing Supercomputers' Growing Drain on Energy,” and CRD researcher David Patterson was quoted in the New York Times article “Progress Hits Snag: Tiny Chips Use Outsize Power.” You can stay up to date on Computing Sciences in the news on this web page.
John Bell Gives Plenary Talk at 2011 ICiS Multiphysics Workshop
John Bell of CRD gave an invited plenary talk at the 2011 Multiphysics Workshop, held July 30–August 6 in Park City, Utah, and sponsored by the Institute for Computing in Science (ICiS). Bell spoke on “AMR Algorithms for Multiphysics Applications: Current Practice and Future Directions.” His talk is available here.
Yuan, Drummond Take Part in European-US HPC Summer School
The NSF TeraGrid, the Distributed European Infrastructure for Supercomputing Applications (DEISA), and the Partnership for Advanced Computing in Europe (PRACE) are sponsoring the second joint EU-US Summer School on High Performance Computing (HPC) Challenges in Computational Science this week, August 7–12, at Lake Tahoe, California.
Sixty US and European graduate and postdoctoral scholars, including NERSC Petascale Postdoc Xuefei (Rebecca) Yuan, were selected to participate in this expense-paid program. Leading American and European computational scientists and HPC technologists are giving presentations, including Tony Drummond of CRD, who will present “Numerical Libraries.”
Correction: 26 CS Researchers Presented Results at ICIAM 2011
One more addition to the list of Computing Sciences researchers who contributed to the 7th International Congress on Industrial and Applied Mathematics (ICIAM 2011): Xuefei (Rebecca) Yuan co-authored and presented a poster on “Implicit, Full Coupling of the Adaptive Grid and Physical Problems in Plasma Simulation with Additive Schwarz Preconditioned Inexact Newton.” You can see the complete list of CS conference contributors here.
Margie Wylie Named CS Web Content Coordinator
As of August 1, Margie Wylie is the new web content coordinator for all of Computing Sciences. In this new position (which also moves Margie from a term to a career appointment), she will be working with CRD, ESnet, and NERSC on various web projects, including implementation of the new CRD site, arranging training, etc. In her spare time she will also be handling other communications projects, albeit on a reduced scale. She joined the Computing Sciences communications team in January 2010 and up till now has been focusing on NERSC news and communications projects.
Call for Papers on Network-Aware Data Management
Mehmet Balman and Surendra Byna of CRD are organizing a Network-Aware Data Management Workshop to be held on November 14 in conjunction with SC’11 in Seattle.
Since the amount of data is continuously growing, traditional techniques of leaving the burden on the user of moving/storing data is not a viable option in many cases. In order to deliver true exascale performance to the application layer, network-aware data management services for resource provisioning, intelligent data-flow and coordination between the host systems is highly desirable. This workshop will seek contribution from academia, government, and industry to discuss emerging trends in use of networking for data management, novel abstraction techniques for data representation, simplification of end-to-end data flow by providing transparent data services, end-to-end resource coordination, and network-aware tools for the scientific community.
Researchers are invited to submit papers on topics such as:
- Challenges in data-intensive distributed computing
- End-to-end resource provisioning
- Optimization and development of data transfer protocols
- Scalable services for network-aware applications
- Re-configurable end-to-end data-access frameworks
- Network-aware scheduling and resource brokering
- Advanced services for managing data-flow
- On-demand bandwidth allocation and network reservation
- Application pipelines and workflow management
- Network-aware toolkits for data distribution
- Data replication and metadata management
- Data clouds, data scheduling, and data placement
- Heterogeneous resource management
- Performance evaluation of network-aware data management
Abstracts are due on August 22 and full papers on August 29. See submission guidelines here.
This Week’s Computing Sciences Seminars
Why We Can Expect Ever More Amazing Mobile Computing Devices in the Years Ahead
Monday, August 8, 12:00–1:00 pm, 90-4133
Jon Koomey, Stanford University
Information technology (IT) has captured the popular imagination, in part because of the tangible benefits IT brings, but also because the underlying technological trends proceed at easily measurable, remarkably predictable, and unusually rapid rates. The number of transistors on a chip has doubled more or less every two years for decades, a trend that is popularly (but often imprecisely) encapsulated as “Moore’s Law.”
This talk will explore the relationship between the performance of computers and the electricity needed to deliver that performance. Computations per kWh grew about as fast as performance for desktop computers starting in 1975, doubling every 1.5 years, a pace of change in computational efficiency comparable to that from 1946 to the present. Computations per kWh grew even more rapidly during the vacuum tube computing era and during the transition from tubes to transistors but more slowly during the era of discrete transistors. As expected, the transition from tubes to transistors shows a large jump in computations per kWh.
In 1985, the physicist Richard Feynman identified a factor of one hundred billion possible theoretical improvement in the electricity used per computation. Since that time computations per kWh have increased by less than five orders of magnitude, leaving significant headroom for continued improvements. The main trend driving towards increased performance and reduced costs in the microprocessor era, namely smaller transistor size, also tends to reduce power use, which explains why the industry has been able to improve computational performance and electrical efficiency at similar rates. If these trends continue (and we have every reason to believe they will for at least the next five to ten years), this research points towards continuing rapid reductions in the size and power use of battery-powered mobile computers, allowing further rapid progress in mobile sensors, computing, and controls.
The paper documenting the work to be summarized in this talk is Koomey, Jonathan G., Stephen Berard, Marla Sanchez, and Henry Wong. 2011. "Implications of Historical Trends in The Electrical Efficiency of Computing." In Press at the IEEE Annals of the History of Computing. July.
Uncertainty Quantification Discussion Group
Monday, August 8, 10:00–11:00 am, 50F-1647
Taghrid Samak, LBNL/CRD
The group will discuss two papers: “How to avoid a perfunctory sensitivity analysis” and “Ignorance is bliss: Or seven reasons not to use uncertainty analysis.”
NERSC Brown Bag: Data-Driven Web Pages on the New NERSC Web Site
Tuesday, August 9, 12:00–1:00 pm, OSF 943-238
Richard Gerber, NERSC
NERSC Brown Bag: ALS Applications and Magnetic Materials
Friday, August 12, 12:00–1:00 pm, OSF 943-238
Catherine Jenkins, ALS
Link of the Week: Why Exercise Makes Us Feel Good
Why does exercise make us happy and calm? Almost everyone agrees that it generally does, a conclusion supported by research. But just how, at a deep, cellular level, exercise affects anxiety and other moods has been difficult to pin down. But a recent animal study from researchers at the National Institute of Mental Health provides some intriguing new clues into how exercise intertwines with emotions, along with the soothing message that it may not require much physical activity to provide lasting emotional resilience. Read more.
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.