A-Z Index | Phone Book | Careers

InTheLoop | 09.17.2012

September 17, 2012

Berkeley Lab Computational Scientists to Help Illuminate Dark Universe

As part of a new Department of Energy collaboration aimed at illuminating the 95 percent of the universe known as dark matter and dark energy, researchers in Berkeley Lab’s Computational Research Division will apply their scientific computing expertise in simulation and analysis to boost the success of next-generation cosmology experiments. Peter Nugent, co-leader of the Computational Cosmology Center, and Ann Almgren, an applied mathematician in the Center for Computational Sciences and Engineering (CCSE), are the leads for Berkeley Lab’s participation in the three-year project called Computation-Driven Discovery for the Dark Universe as part of DOE’s Scientific Discovery though Advanced Computing (SciDAC) program. Read more.


At 100 Gbps, ESnet Puts Network Research on Fast Track

When the DOE announced the Advanced Networking Initiative in 2009 to develop the first 100 gigabit-per-second (Gbps) production-ready science network, it also included funding for a 100 Gbps experimental testbed and a national dark-fiber testbed, both operated by ESnet. Recently, the testbeds received additional funding for continued operations, providing a unique resource for experimenting with potentially disruptive network technologies.

Brian Tierney, who leads ESnet’s Advanced Networking Technologies Group, talks more about the project in this Q&A for HPCwire with Jon Bashor of Berkeley Lab Computing Sciences Communications. Read more.


Wes Bethel and Hank Childs Co-Edit Book on High Performance Visualization

Wes Bethel and Hank Childs of CRD, along with Charles Hansen of the University of Utah, have edited the book High Performance Visualization: Enabling Extreme-Scale Scientific Insight, which is now available for order from CRC Press. The publication date is October 19.

The book focuses on algorithm design, implementation, and optimization for visualization and analysis tools on emerging multi- and many-core platforms. It collects some of the most seminal work in the field, including algorithms and implementations running at the highest levels of concurrency and used by scientific researchers worldwide.

Other CRD staff who contributed to the book include Oliver Rübel, Prabhat, Kesheng John Wu, David Camp, Mark Howison, Gunther Weber, and Hari Krishnan. Berkeley Lab Deputy Director Horst Simon wrote the foreword.


ESnet’s Greg Bell to Give Invited Keynote Talk at Scandinavian Networking Conference

Greg Bell, head of ESnet and director of Berkeley Lab’s Scientific Networking Division, will give the closing keynote address on Thursday, Sept. 20, at the 2012 NORDUnet conference in Oslo, Norway. NORDUnet is a joint collaboration by the five Nordic national research and education networks in Denmark, Finland, Iceland, Norway and Sweden. ESnet has collaborated with NORDUnet on common methodologies for reserving end-to-end bandwidth to accelerate large data transfers, and on standards development for the emerging Network Service Interface protocol.

In his talk, “Network as Instrument: The View from Berkeley,” Bell will argue that “it’s time to start thinking about research networks as instruments for discovery, not as infrastructures for service delivery.” He will explain what’s at stake in this distinction, and “how ESnet’s strategy is influenced by its unusual institutional context: embedded in a U.S. national laboratory, classified as a user facility, and located uphill from a famously audacious university.”


CRD Staff in the News

In an editorial for The Conversation, CRD’s Complex Systems Group Lead David Bailey and University of Newcastle Mathematics Professor Jonathan Borwein debunk myths about the dangers of Smart Meters.

Computer and Data Sciences Department Head John Shalf was quoted in a Wired article explaining why Intel and AMD are building new networking fabric directly into chips and servers.

Future Technologies Group Lead Erich Strohmaier was quoted in this CIO news story about Intel and HP teaming up to build the world’s most efficient supercomputer at DOE’s National Renewable Energy Laboratory.


Volunteers Still Needed for CS Exhibit at Berkeley Lab Open House

“Science at Warp Speed: A Journey Through Space and Time in 3D” is the theme of the Computing Sciences exhibit at this year’s Berkeley Lab Open House, which takes place Saturday, Oct. 13, from 10 am to 3 pm. Here is the full description:

Travel 14 billion years back to the beginning of space and time. Soar through burner flames and supernova explosions. Explore the microscopic pores of rocks buried deep inside Earth. And if you are still not tired, zip around the world at 100 gigabits per second. Sound impossible? It’s not, researchers go on this journey every day with the ultimate scientific instruments—supercomputers and networks. Part time-machine, spaceship, microscope and more, these tools allow humans to explore environments and events that are too big, too small, too fast, too slow, or too dangerous to study otherwise.

Volunteers who work a minimum of three hours will receive a free T-shirt and a voucher for lunch. Go here to register; under “Attendee Type,” choose “Volunteer: Assigned to an Exhibit,” and send an e-mail to Computing Sciences’ Open House coordinator Linda Vu to let her know that you’ve registered.


This Week’s Computing Sciences Seminars

Matrix-Valued Boltzmann Equation for the Hubbard Chain
Monday, Sept. 17, 2:00–3:00 pm, 50F-1647
Christian Mendl, Technische Universität München, Germany

We study, both analytically and numerically, the Boltzmann transport equation for the Hubbard chain with nearest neighbor hopping and spatially homogeneous initial condition. The time-dependent Wigner function is matrix-valued because of spin. The H-theorem holds. The nearest neighbor chain is integrable which, on the kinetic level, is reflected by infinitely many additional conservation laws and linked to the fact that there are also non-thermal stationary states. We characterize all stationary solutions. Numerically, we observe an exponentially fast convergence to stationarity and investigate the convergence rate in dependence on the initial conditions.

New Frontiers in Networking
Monday, Sept. 17, 3:00–4:00 pm, 50B-4205
Martin Swany, Indiana Center for Network Translational Research and Education (InCNTRE), Indiana University

The effective and efficient utilization of networks is a persistent challenge in today’s computing ecosystem. The trend towards increasingly distributed computing environments supporting data-intensive applications compounds the issue, making data movement a cross-cutting concern, affecting users of distributed computing and data-intensive science alike. This talk will discuss emerging network approaches and their implications for the future. Two key areas to be explored are software defined networking and remote direct memory access (RDMA) over Ethernet.

OSF Brown Bag: EVMS Overview
Wednesday, Sept. 19, 12:00–1:00 pm, OSF 943-238
Tanya Boysen, LBNL

An earned value management system (EVMS) is a set of best business practices, processes, and tools for project planning and control. The process includes integration of scope, schedule, cost, a performance measurement baseline, and earned value. A lot of us have heard about EVMS. Tanya Boysen from the Lab’s CFO office will give a brief, informal overview of EVMS and address a number of questions: Is it useful? Who uses it? Do I need to know more about it?

CITRIS Research Exchange: What Lies Beneath: The Foundations of Motivation in the Age of Social Media
Wednesday, Sept. 19, 12:00–1:00 pm, 250 Sutardja Dai Hall, Banatao Auditorium, UC Berkeley
Judd Antin, User Experience Researcher, Facebook

In this talk, I’ll focus on the deep motivations that drive so many to create, interact, and collaborate on Facebook, StackExchange, Wikipedia, Twitter, and countless other sites. As social media have become the Internet’s driving force, questions about motivation and incentives have come to the forefront. Motivation, however, is hard to talk about and harder to measure. In this talk I discuss some of the problems with current models of motivation that are enshrined in trends like “gamification.” However, I’ll argue that focusing on underlying social psychological motivations presents important opportunities for research and design, and finish by discussing how I am bringing this perspective to Facebook.

Planetary Convection: Heat Transfer by Rotating Flows
Wednesday, Sept. 19, 4:00–5:00 pm, 939 Evans Hall, UC Berkeley
Eric King, UC Berkeley

Fluids throughout the solar system are driven to turbulence by thermal and gravitational energy. These convective flows are subject to the rapid rotation of their host bodies through the Coriolis force, which can dramatically alter the nature of turbulent convection. Using simplified models of turbulent, rotating convection, we try to answer an age old question: how much heat can these flows transport? Following the seminal work by Malkus (1954), we leverage the important role of the thermal boundary layers to infer an exact scaling law for heat transport by rapidly rotating convection. The scaling law is then tested against results from experiments and simulations.

EECS Colloquium: CrowdSolving
Wednesday, Sept. 19, 4:00–5:00 pm, 306 Soda Hall, UC Berkeley
Adrien Trueille, Carnegie Mellon University

What happens when hundreds of thousands of non-experts play a protein folding game? They make important scientific discoveries, of course! Behind the scenes, such magical applications are being enabled by advances in simulation technology and by our deepening understanding of on-line social games. This talk will describe several large-scale experiments we have launched in which hundreds of thousands of gamers are playing with interactive biomolecular simulations and thereby making discoveries which have eluded the scientific community for decades. These experiences presage a new era of “CrowdSolving” with radical implications for scientific discovery, education, labor organization, and intellectual property law.

TRUST Security Seminar: The Calculus of Security
Thursday, Sept. 20, 1:00–2:00 pm, 254 Sutardja Dai Hall, UC Berkeley
George Cybenko, Dartmouth College

This talk will describe several concrete ideas and theoretical results for quantifying the security of composed computing and information systems. Some techniques described are security analogs of reliability theory while others are based on game theory models of attacking and defending complex systems.

Dissertation Talk: Parallel Application Library for Object Recognition
Thursday, Sept. 20, 4:00–5:00 pm, 380 Soda Hall, UC Berkeley
Bor-Yiing Su, EECS, UC Berkeley

Object recognition is a major branch of computer vision that identifies objects from images. To achieve the highest accuracy, state-of-the-art object recognition systems must extract features from hundreds to millions of images, train models with enormous data samples, and deploy those models on query images. As a result, these systems are computationally intensive. In order to make such complicated algorithms practical to apply in real life, we must accelerate them on modern massively parallel platforms. However, parallel programming is complicated and challenging, and takes years to master. In order to help object recognition researchers employ parallel platforms more productively, we propose a parallel application library for object recognition. To pave the way for such a library, we perform pattern mining on 31 important object recognition systems, and conclude that 15 application patterns are necessary to cover the computations in these systems.

In this talk, I will present how we optimize a subset of the application patterns by carefully exploring the design space of algorithms, parallelization strategies, and platform parameters. We have performed exhaustive search on three kernels: eigensolver for the normalized cut algorithm, breadth-first-search graph traversal on images, and contour histogram. To improve the portability and flexibility of the proposed library, we also initiate the OpenCL for OpenCV project. This project aims to provide a collection of autotuners that optimize the performance of application patterns on many different parallel platforms. We have developed two autotuners in this project. clSpMV is an autotuner for sparse matrix vector multiplication (SpMV) computation — it tunes the representation of a sparse matrix and the corresponding SpMV kernel. clPaDi is an autotuner for the pair-wise distance computation application pattern — it allows a user to customize his own distance function, and finds the best blocking size for that function. By employing these optimized functions in a state-of-the-art object recognition system, we have achieved 110–120x speedup compared to the original serial implementation. Now it takes only three seconds to identify objects in a query image — a much more practical and useful processing time. Our research makes it possible to deploy complicated object recognition algorithms in real applications.

With these encouraging results, we are confident that our methodology is applicable to optimizing all application patterns. If we expand the parallel application library to support all 15 application patterns, the library will be a key toolkit for both existing and future object recognition systems.


Link of the Week: Engineers Build Supercomputer Using Raspberry Pi, Legos

Computational engineers at the University of Southampton have built a supercomputer from 64 Raspberry Pi computers, using Lego bricks to build the racks. Raspberry Pi is an ARM GNU/Linux box that sells for around $25, so the whole system cost under £2,500 (excluding switches). The system can be powered on just a single 13 Amp main socket. 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.