InTheLoop | 07.15.2013
The weekly newsletter for Berkeley Lab Computing Sciences employees
July 15, 2013
Berkeley Lab Hosts ESnet Site Coordinators Meeting, Life Sciences Data Workshop
Berkeley Lab is hosting the twice yearly meeting of the ESnet Site Coordinators Committee (ESCC) beginning , Monday, July 15, though Tuesday, July 16, in the Bldg. 66 auditorium.ESCC meetings are held twice a year. The meetings provide a forum for the site coordinators at each of the national labs to coordinate and plan network activities with ESnet staff and each other on common networking issues. See the agenda.
Following the ESCC Meeting, ESnet and Internet2 will be hosting the inaugural Focused Technical Workshop: Network Issues for Life Sciences Research on Wednesday and Thursday, July 17-18. The meeting will bring network engineers, network researchers together with the life science research community at Berkeley Lab to discuss network and computing challenges and requirements with the purpose of developing next steps for supporting this data-intensive science community
Berkeley Science Review: New Computers Resolve Clouds While Keeping Cool
Like the planet they simulate, climate models themselves are changing. As our scientific computing technology improves, models are getting larger and more complicated in order to make better predictions. However, with greater predictive power comes greater demand for computational and electric power. In this story from the Berkeley Science Review, writer Anna Lieb, explores how computer scientists from Berkeley Lab’s Computational Research Division and the Department of Computer Science at UC Berkeley are putting together a new blend of hardware and software to run climate simulations that are both accurate and more energy efficient than ever before. Read more.
This Week’s Computing Sciences Seminars
Efficient parallelization of scientific applications for numerical simulations
applied to nuclear systems
Monday, July 15, 11 a.m.–12 noon, Bldg. 50F, Room 1647
Christophe Calvin, Group Lead of High Performance Computing and R&D, Commissariat a l'Énergy Atomique et aux Énergies Alternatives (CEA), Saclay, France
Abstract: The work presented traces my various research activities for efficient parallelization of scientific applications in the field of numerical simulation applied to nuclear systems. Various topics will be discussed, such as the study, design and implementation of software architectures for efficient parallelization of numerical simulations applications in the field of thermal hydraulics, but also researches on the issue of dynamic meshes and load balancing for applications in fluid mechanics.
Another major area of numerical simulation for nuclear systems which I am interested is reactors physics. I particularly focused on the issues of the use of supercomputing in the field of numerical simulation for nuclear systems in order to identify the contribution of HPC in the field, but also future challenges.
Beyond the parallelization of applications themselves, basic research on parallel numerical algorithms have been conducted both on iterative methods for linear systems resolutions and on Krylov methods for eigenproblems. These numerical methods are extremely useful in scientific applications in the nuclear field (resolution of the Navier-Stokes equation or neutron transport) and allow not only to progress in parallel numerical algorithms but also to test and compare new approaches to high performance both in terms of computing hardware and programming languages (CUDA, OpenCL and PGAS languages).
A Data Management System with Web Interface for Pre-clinical
Multimodality Imaging: ReMI
Monday, July 15, 10–11 a.m., Bldg. 50F, Room 1647
Dr. Debasis Mitra and Hui Pan, Florida Institute of Technology and visiting scholars at LBNL
Abstract: Molecular imaging research is an inter-disciplinary activity, where expertise may be distributed over multiple locations. A well-designed database with web access is very useful for this purpose. We are designing a database for pre-clinical imaging research in order to prevent data loss or corruption, and to enable re-use of the same data in different research projects while data may grow dynamically. Well-managed metadata is a key to the reusability of any experimental data. A relational database stores the metadata. It points to, as well as manages, a file archive on a large server storing the experimental data. Typically, experimental data are binary images and very large in size, of the order of a few megabytes each. A scalable model-view-controller architecture is used for integrating the system with a web-based GUI.
The database ReMI (Research in Medical Imaging), accessible at http://remi.lbl.gov, is being populated with data that can be shared within the community, and is ready to add new users and contributors. A clone of the database may be accessed at http://126.96.36.199:3000 with (guest-guest account-password), where all metadata is similar to the actual database ReMI, but without the pointers to the actual data files as in the latter. ReMI is fully functional now. (1) However, entering metadata is a tedious and error-prone process. Even though some amount of user-personalization is available currently, we would like to automatically extract as much metadata as possible from relevant files by borrowing techniques from artificial intelligence. (2) A user should be able to quickly sample image data before deciding to go through the computationally expensive downloading process. Efficient online multi-modality image registration capability based on organ-specific atlas may be needed for this purpose.
Link of the Week: Physics, Computer Scientists and the Birth of the Emoticon
When physics combines with computer science, the result can be as groundbreaking as the recent discovery of the Higgs boson. And while not everyone may avail themselves of this discovery, another product of a physics-computer scientist mashup has become a key part of global online life – the emoticon. In this posting, Carnegie Mellon University alumni trace the origin of the smiley to a group of computer scientists discussing a physics puzzle in 1982. Read more.
A Note on InTheLoop Submissions
With the June 27 retirement of InTheLoop editor John Hules, news items for the weekly Computing Sciences staff newsletter should be sent to Jon Bashor at firstname.lastname@example.org.
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.