InTheLoop | 11.07.2011
November 7, 2011
ESnet Pushes Green Networking at SC11 Conference
ESnet staff are engaged in a number of activities at the upcoming SC11 conference to raise the visibility of efforts to make networking more energy efficient. One of the most prominent is the SC11 workshop: "Data Centers Have Gone Green (Or Haven't They?) When Will Networks Follow?" to be held from 9 a.m. – 5 p.m. Sunday, Nov. 13 in conjunction with SC11. Inder Monga, ESnet’s area lead for Research and Services, is co-organizer of the workshop. Read about other green networking sessions at SC11.
Unlike data centers where energy use is metered at a single site, networking equipment is widely distributed, making the power consumption harder to assess. One component of ESnet's Advanced Networking Initiative is the installation of equipment to measure energy use across the network. Read more about ESnet’s efforts in this area.
NERSC Science Gateway a 'Google of Material Properties'
New materials are crucial to building a clean energy economy—for everything from batteries to photovoltaics to lighter weight vehicles. But today the development cycle is around 18 years from conception to commercialization. To speed up this process, researchers from DOE's Berkeley Lab and MIT teamed up to develop a Google-like search engine for materials research. Called the Materials Project, the search engine is hosted on NERSC's Science Gateway infrastructure. Read more.
David Bailey's Household Earns Energy Bragging Rights
After the electrical portion of his utility bill soared over $400 in 2008, Berkeley Lab's Chief Technologist David Bailey worked tirelessly to improve the energy efficiency of his home. After some forensics work and a few improvements, his energy bill is now about $50 per month. He writes about his experience in the latest issue of Home Energy Magazine. Read the article.
PopScI Profiles NERSC's Franklin and Hopper
Last week, PopSci is peeked under the hood of some of the nation's biggest and baddest supercomputers—the machines that turn big data into big discoveries, big technologies, and big leaps forward. As part of the series, their editors contacted NERSC to find out what Franklin and Hopper were up to on a particular day. Read the entire series.
This Week’s Computing Sciences Seminars
Highly Parallel Applications and their Architectural Needs
Wednesday, Nov. 9, 11 a.m. - 12:30 p.m., 50B- 4205
Anand Deshpande, Intel Parallel Computing Research Lab
The "massive compute meets massive data" theme is especially applicable for scientific and engineering computations such as CFD. Further, with the convergence of Recognition, Mining and Synthesis (RMS) algorithms required for solving some of the complex problems, the need for massive compute on massive data is even more critical. Thus, the applications of today as well as tomorrow are driving the industry towards highly parallel, multicore and manycore computer architectures. In this talk, we will discuss multicore and manycore architectures and show examples of parallel computing applications using these architectures. We will give an overview of the research trends in this area and the strategies to address the compute needs of the future. Further, we will give an overview of the research happening in Parallel Computing Lab, Intel Labs, India and discuss potential opportunities for collaboration.
Multigrid Methods for Helmholtz Problems: Application of Complex Shifted Laplacian to Chemical Break Up Reactions
Tuesday, Nov. 8, 2 p.m. 2- 100B
Wim Vanroose, Universiteit Antwerpen, Belgium
The Helmholtz equation is ubiquitous in applications. It describes the propagation of seismic waves, electromagnetic scattering and chemical breakup reactions. After discretization, the problem becomes a complex valued and indefinite linear system Ax=b. In this talk we give an overview of recent progress and remaining challenges in the iterative solution methods for Helmholtz problems. In particular the class of "Complex Shifted Laplacian" iterative methods will be discussed. Here, the problem is solved by a preconditioner that is a shifted problem for which Multigrid can be used for the inversion. The methods will be evaluated on 3D model problems relevant for high performance simulation of high dimensional chemical break-up reactions. Examples from 3D seismic imaging will also be discussed.
Note: This seminar is presented by Dan Haxton of Chemical Sciences
Link of the Week: Local Community College Teacher Profiled as HPC Innovator
Tom Murphy, who teaches computer science at Contra Costa College in San Pablo, challenges his students by having them work on projects such as LittleFe, a 6-node cluster in a small crate, or LittleAl, a 16-core system in an aluminum attaché case. Murphy also works with the Lab’s IT Division in helping train his students for real-world jobs and last year brought his Parallel Programming Club to visit NERSC. As part of a lead-in to SC11, Computing Sciences’ Jon Bashor profiled Murphy for an HPC Innovator series published by Scientific Computing magazine. Read the profile.
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.