InTheLoop | 02.04.2008
The weekly newsletter for Berkeley Lab Computing Sciences
February 4, 2008
“The Register” Discusses LBNL’s Research on Energy-Efficient Computing
The Register, a British-based online news source that covers the information technology beat with a cheeky attitude — their motto is “Biting the hand that feeds IT” — sent reporter Ashlee Vance to Berkeley Lab last Wednesday to cover the Technology and Inventors Expo. Vance filed two stories:
In “Geeks fight the smelter with embedded processor-based box”, Vance writes, “So, we’re glad to hear that Horst Simon, a prominent computer scientist at Berkeley Lab, has renewed work around slotting low power chips into supercomputer class machines.” The article discusses energy-efficient computing in general as well as the Climate Computer that was conceptually proposed in the “NERSC 2016” white paper.
The other story, “Nobel Prize winner demands more honesty from peers in green debate”, begins, “Scientists hoping to educate the public about environmental concerns could do themselves and the public a favor by abandoning hyperbolic scare tactics in favor of straightforward talk, according to a prominent scientist.” The article presents Lab Director Steve Chu’s view that scientists should discuss the range of possibilities and risks and let the public decide how they feel about them.
Arie Shoshani Invited to Contribute to Database Encyclopedia
Arie Shoshani, head of SciDAC’s Scientific Data Management Center and leader of CRD’s Scientific Data Management Group, has written two essay-type entries for the Encyclopedia of Database Systems, to be published later this year by Springer Verlag. According to the publisher, “The Encyclopedia of Database Systems is designed to meet the needs of research scientists, professors and graduate-level students in computer science and engineering.… Topics for the encyclopedia were selected by a distinguished international advisory board, and written by world class experts in the field.”
Based on his work in the field, Shoshani was invited to write entries on “Logical Models of Temporal Data” and “Summarizability Properties of Statistical Databases.” To request copies of the entries, contact Arie at email@example.com.
Ali Pinar Elected Officer of SIAM Activity Group on Supercomputing
Ali Pinar, a researcher in the Scientific Computing Group, has been elected to serve as secretary for the SIAM Activity Group on Supercomputing (SIAG/SC). His two-year term began on January 1. Ali’s research focuses on combinatorial scientific computing, and he is particularly interested in tackling combinatorial problems that are directly associated with scientific and engineering goals. He has worked on vulnerability analysis of the electric power grid, interconnection networks for ultra-scale systems, energy- efficient disk systems, and supernova spectra analyses.
SIAM is the Society for Industrial and Applied Mathematics. The SIAM Activity Group on Supercomputing provides a forum for computational mathematicians, computer scientists, computer architects and computational scientists to exchange ideas on mathematical algorithms and computer architecture needed for high -performance computer systems. The activity group promotes the exchange of ideas by focusing on the interplay of analytical methods, numerical analysis and efficient computation.
Four Scientific Computing Seminars Scheduled This Week
Four seminars have been scheduled this week. Here are the details:
(1) Monday, February 4, 11:00 am–noon, 50B 4205 conference room
Speaker: Karl Fuerlinger, Innovative Computing Laboratory, University of Tennessee
Title: Profiling and Incremental Profiling of OpenMP Applications
Profiling is often the method of choice for performance analysis of parallel applications due to its low overhead and easily comprehensible results. However, a disadvantage of profiling is the loss of temporal information that makes it impossible to causally relate performance phenomena to events that happened prior or later during execution. The talk presents a simple but useful profiling tool for OpenMP applications, ompP, and describes its utility for overhead and scalability analysis. We also present techniques to add temporal dimension to profiling data by incrementally capturing profiles during the runtime of the application and discuss the insights that can be gained from this type of performance data. Application examples come from the SPEC OpenMP benchmark suite.
(2) Tuesday, February 5, 11:00 am–noon, 50F 1647 conference room
Speaker: Erika Fuentes, Department of Computer Science, University of Tennessee
Title: Statistical Learning and Data Mining Techniques for Algorithm Selection for Solving Sparse Linear Systems
There are many applications and problems in science and engineering that require large-scale numerical simulations and computations. The issue of choosing an appropriate method to solve these problems is very common, however it is not a trivial one, principally because this decision is most of the time too hard for humans to make, or a certain degree of expertise and knowledge in the particular discipline or in mathematics is required. Thus, the development of a methodology that can facilitate or automate this process and help to understand the problem would be of great interest and help. The proposal is to utilize various statistically based machine-learning and data mining techniques to analyze and automate the process of choosing an appropriate numerical algorithm for solving a specific set of problems based on their individual properties.
(3) Wednesday, February 6, 11:00 am–noon, 50F 1647 conference room
Speaker: Rong Ge, Scalable Performance Lab, Virginia Polytechnic Institute and State University
Title: Theories and Techniques for Efficient High-End Computing
As large-scale computing systems grow tremendously in size and capacity, improving power and performance efficiency becomes a compelling issue. Today it is common for a supercomputer to consume several megawatts of electric power but deliver only 10–15% of its peak performance for average applications. Such power consumption not only costs millions of dollars annually but also dissipates enormous heat that reduces system reliability and productivity. To address these issues, I have developed theories to model the performance and power in high-end computing systems as well as techniques to optimize power and performance efficiency. In this talk, I will present these theories and techniques, yet focus on the quantitative communication performance models (lognP and log3P) and their usage in improving high-end computing efficiency. Compared to previous models, these models explicitly quantify the cost of memory accesses and middleware communications in distributed systems, and thus provide more accurate performance prediction. Moreover, these models aid algorithm designs that improve performance and efficiency. Results show algorithms designed using the lognP and log3P models can outperform those designed by previous models and reduce execution time by up to 59%.
(4) Thursday, February 7, 11:00 am–noon, 50F 1647 conference room
Speaker: Yunrong Zhu, Department of Mathematics, Penn State University
Title: Robust Multilevel Preconditioners for Problems with Strongly Discontinuous Coefficients
Although there is a vast literature of multilevel and domain decomposition (DD) methods for the finite element discretization of elliptic (H(grad)) systems, it remains an open question how to make these efficient solvers converge (nearly) uniformly for the H(grad) systems with strongly discontinuous coefficients. Recently, we proved that the multilevel and DD preconditioners lead to nearly uniform convergent preconditioned conjugate gradient methods. In this talk, I will present the theoretical and numerical justification of these results. As applications, I will also present the auxiliary space preconditioners (Hiptmair and Xu 2007) for H(curl) and H(div) systems in the compatible discretization framework, which reduce the Maxwell equations and mixed formulation of elliptic equations into solving several H(grad) equations. I will give some numerical results for the H(div) system and its application.
Student Competition at CITRIS Offers $25K in Prizes
CITRIS (Center for Information Technology in the Interest of Society) has announced the third annual CITRIS White Paper competition, which will give away $25K in cash prizes for the best ideas that demonstrate the ability of information technology to address a major societal challenge. The IT for Society contest is open to students from all four CITRIS campuses: UC Berkeley, UC Merced, UC Santa Cruz, and UC Davis.
The prizes will be allocated as follows: First Place: $12,500; Second Place: $7,500; Third Place: $3,000; Fourth Place: $2,000. The cash can be used as scholarships or to support the proposed project or idea (e.g., research, travel, workshops, miscellaneous expenses).
NERSC News: Climate Edition
Check out the latest NERSC News, which features two climate research projects that showcase the ability of the new Cray XT4 to accelerate scientific discovery. One of the projects also enlisted Prabhat on the NERSC Analytics Team to create a series of animations for analyzing the data. The other project broke speed and other records in developing a weather model. In this issue, you also will find a list of the 2008 INCITE projects for NERSC and Kathy Yelick’s vision for the center.
So You Think You Know What to Do during an Earthquake?
CS Safety Coordinator John Hutchings has found a challenging ten-question interactive quiz on what to do during and after an earthquake. Even though the quiz is centered on the home environment, it makes several points that could also apply here at the Lab. If you’ve lived in the Bay Area for a while and think you know all the answers.
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.