A-Z Index | Phone Book | Careers

InTheLoop | 10.27.2014

October 27, 2014

Water and Gold: A Promising Mix for Future Batteries

When a solid material is immersed in a liquid, the liquid immediately next to its surface differs from that of the bulk liquid at the molecular level. This interfacial layer is critical to our understanding of a diverse set of phenomena from biology to materials science. When the solid surface is charged, just like an electrode in a working battery, it can drive further changes in the interfacial liquid. However, elucidating the molecular structure at the solid-liquid interface under these conditions has proven difficult.

Now, for the first time, researchers at the US Department of Energy’s (DOE) Lawrence Berkeley National Laboratory (Berkeley Lab) have observed the molecular structure of liquid water at a gold surface under different charging conditions. Their findings were published October 23, 2014 in the journal Science.

Lead author Miquel Salmeron, a senior scientist in Berkeley Lab’s Materials Sciences Division (MSD) and professor in UC Berkeley’s Materials Science and Engineering Department, and his team have developed a method not only to look at the molecules next to the electrode surface but to determine their arrangement changes depending on the voltage. Calculations run at the National Energy Research Scientific Computing Center (NERSC) by co-author David Pendergast, a staff scientist in the Lab's Molecular Foundry and researcher in the Joint Center for Energy Storage Research (JCESR), helped the researchers to better understand the chemistry involved. »Read more.

Probing the Surprising Secrets of Carbonic Acid

Though it garners few public headlines, carbonic acid, the hydrated form of carbon dioxide, is critical to both the health of the atmosphere and the human body. However, because it exists for only a fraction of a second before changing into a mix of hydrogen and bicarbonate ions, carbonic acid has remained an enigma. A new study by Berkeley Lab researchers has yielded valuable new information about carbonic acid with important implications for both geological and biological concerns.

Richard Saykally, a chemist with Berkeley Lab's Chemical Sciences Division and a professor of chemistry at the University of California (UC) Berkeley, led a study that produced the first X-ray absorption spectroscopy (XAS) measurements for aqueous carbonic acid. These XAS measurements, which were obtained at Berkeley Lab's Advanced Light Source (ALS), were in strong agreement with supercomputer predictions obtained at the National Energy Research Scientific Computing Center (NERSC). »Read more.

CRD Researchers, NERSC Help Model 3D Map of Adolescent Universe

Using extremely faint light from galaxies 10.8 billion light years away, scientists have created one of the most complete, three-dimensional maps of a slice of the adolescent universe—just 3 billion years after the Big Bang.

The map shows a web of hydrogen gas that varies from low to high density at a time when the universe was made of a fraction of the dark matter we see today. It was created in part using supercomputing resources at the National Energy Research Scientific Computing Center (NERSC) by a team that included researchers from Berkeley Lab's Computational Cosmology Center (C3) in the Computational Research Division (CRD).

The new study, led by Khee-Gan Lee and his team at the Max Planck Institute for Astronomy in conjunction with researchers at Lawrence Berkeley National Laboratory (Berkeley Lab) and UC Berkeley, was published October 16, 2014 in Astrophysical Journal Letters. »Read more.

New perfSONAR Release Offers More Network Measurement Data, Increased Security

Since it was first released about five years ago, the perfSONAR network measurement toolkit has become increasingly important as research grows more collaborative and dependent on sharing large data sets. It gives network engineers with the ability to test and measure network performance and to archive data in order to pinpoint and solve performance problems that may span multiple networks and international boundaries.

The latest release of perfSONAR, version 3.4, provides more data about network performance and increased security protections. It will be discussed in a session called “perfSONAR 3.4: Not Just another Incremental Update” at the Technical Exchange conference being held from Oct. 19-25 in Indianapolis. The toolkit is developed by a collaboration between the Department of Energy’s ESnet, Internet2, Indiana University and GEANT, the pan-European research network. »Read more.

Berkeley Lab Women Energize Grace Hopper Conference

Women from Berkeley Lab's Computing Sciences area delivered talks, volunteered as mentors and helped organize and energize this year's Grace Hopper Celebration of Women in Computing. Elizabeth Bautista, of the National Energy Research Scientific Computing Center (NERSC), led the lab's diversity efforts as well as several efforts for Filipinas in Computing. Bautista also presented a talk entitled "Building Your Professional Network" and served on the panel "Reach for African and Asian Systers and Women in Computing Everywhere." Deb Agarwal, of the Computational Research Division (CRD), participated in two panels, "Managing Up" and "Latest Trends and Technical Challenges of Big Data Analytics," and co-chaired the LGBT committee. Daniella Ushizima, also of CRD, served on the poster committee, judged student posters, mentored graduate and undergraduate students and participated in the Senior Women in Computing program. Sowmya Balasubramanian and Mary Hester of the Energy Sciences Network (ESnet) also mentored during the Student Opportunities Lab session. NERSC interns Wilma Snider and Michelle Phung led two table top discussions: "Encouraging Filipinas to enter Computing Fields" and "Support for the Challenges of Computing Science Studies." Finally, all Berkeley Lab attendees helped recruit during the career fair. »Read more.

Cybersecurity Expert Jim Mellander Retiring from NERSC

Jim Mellander, senior cybersecurity engineer at NERSC, is retiring November 1. He’s been with NERSC since 2009 and affiliated with Berkeley Lab for nearly 15 years.

Mellander is well known in cybersecurity circles, having developed a number of innovative techniques to enhance cybersecurity at Berkeley Lab and NERSC. He’s also written several notable security software programs, including Update, a UNIX-based sniffer detector; Kazaa Obliterator, which disrupts many types of unauthorized peer-to-peer traffic; and Stomper, which prevents access to unauthorized websites. In addition, Mellander is co-author of the book Intrusion Detection and Prevention, published by McGraw-Hill, and he has received a Best Paper Award from Information Security Bulletin. »Read more.

Staff to Share Networking Expertise and Accomplishments at Tech Exchange Conference

Berkeley Lab staff from ESnet, NERSC and the IT Division will be among the presenters at the 2014 Technology Exchange, a leading technical event in the global research and education networking community. The annual meeting is co-organized by ESnet and Internet2. The conference will be held Oct. 27-30 in Indianapolis. Among the topics to be addressed by Berkeley Lab staff are ESnet’s recently announced 100 gigabits-per-second connections to Europe, the newest release of the perSONAR network measurement software library, Science DMZs and cybersecurity. »Read more.

This Week's CS Seminars

Exascale Seminar: Fast Exact Max-Kernel

Tuesday, Oct. 28, 2 p.m. – 3p.m., Bldg. 50B, Room 4205

Ryan Curtin, Georgia Tech

The celebrated fast multipole method has led to numerous algorithms exploiting trees for acceleration via approximation. One family of such algorithms are dual-tree algorithms, useful for problems including (but not limited to) as nearest neighbor search, minimum spanning tree calculation, kernel density estimation, and k-means clustering. I will describe a unifying tree-independent framework with which to understand, design, and improve dual-tree algorithms. Then, within the context of this framework, I will show a dual-tree algorithm for fast exact max-kernel search (argmax_{p \in S} K(q, p) for all q \in Q and some kernel K). This algorithm is the first fast algorithm to solve this problem exactly, and due to the tree-independent nature of the framework, linear runtime bounds (with respect to a dataset-dependent constant) can be shown when cover trees are used.

CITRIS: Universal Models for Large-scale Coherent Flow Structures in Turbulence

Wednesday, Oct. 29, 3:30 p.m. – 4:30 p.m., 939 Evans Hall, UC Berkeley

Eric Brown, Yale University

Many turbulent astrophysical, geophysical, and engineering flows are dominated by robust coherent structures on the largest scales, such as convection rolls in the atmosphere. One of the challenges in understanding such flows is that these structures have different shapes and dynamics depending on boundary conditions, so they defy universal statistical models of bulk turbulence. On the other hand, the organization imposed onto the turbulent flow by these structures holds some promise for the development of solvable models. I will present an approach to solvable universal models based on using empirically-known flow structures as approximate solutions to the Navier-Stokes equations, which leads to low dimensional dynamical systems models. As an example, I will present results of Rayleigh-Benard convection experiments, in which a container is filled with water and heated from below. Buoyancy drives a flow which organizes into a roll-shaped circulation which spontaneously breaks the symmetry of the system. As a consequence, this roll exhibits a wide range of dynamics including erratic meandering, spontaneous flow cessations, and several oscillation modes. A simple model consisting of stochastic ordinary differential equations quantitatively reproduces these observed flow dynamics. The effects of boundary geometry and different forcings are physically represented by different model terms. These results suggest that we may be able to develop general and relatively easy to solve models for a wide range of turbulent flows with potential applications to climate, weather, and other turbulent flow problems.

CITRIS: Pragmatic Translational Informatics for Health Care

Wednesday, Oct. 29, 12 p.m. – 1 p.m., Sutardja Dai Hall, 310 Banatao Auditorium, UC Berkeley

Nick Anderson, University of California Davis

As director of informatics research at the UC Davis Health System, Nick Anderson does research into the effective uses of biomedical data, information and knowledge for scientific inquiry, problem solving and decision making. Because informatics is at the core of many transformational efforts in health care, Anderson's expertise and leadership will help build the institution's capacity to optimize information technology and further develop this important field throughout the organization. The lecture is free and lunch is provided for those who register.

Exascale Seminar: Advancing HPC I/O to Enable Scientific Discovery

Friday, Oct. 31, 11 a.m. – 12:30 p.m., Bldg. 50B, Room 4205 (via ReadyTalk: OSF 943-236)

Prabhat, Data & Analytics Services, NERSC; Suren Byna, Scientific Data Management Group, CRD; and Allie Andrews, Storage Systems Group, NERSC

It is widely accepted that data size and complexity are impediments to modern computational and experimental science. Storing, reading, finding, analyzing, and sharing data are tasks common across virtually all areas of science, yet advances in data management infrastructure, particularly I/O, have not kept pace with our ability to collect and produce scientific data. This “impedance mismatch” between our ability to produce and to store/analyze data has been increasing as we move toward exascale. With the goal of solving this mismatch, we have been developing various techniques as part of the ExaHDF5 project in collaboration with The HDF Group and Argonne National Laboratory.

In this talk, we will present our recent progress in auto-tuning parallel I/O optimization parameters, developing a parallel querying framework, improving performance of high-level I/O libraries such as HDF5, and reorganizing data layout transparently to accelerate data analysis. Architectural improvements, such as Burst Buffers, offer opportunities for increasing bandwidth and reducing latency for moving data to/from storage. We will describe the Burst Buffer solution we are developing in collaboration with LANL, SNL and Cray for the Cori machine at NERSC, which is due to be delivered in late 2015.