David L Brown, Director of the Computational Research Division, LBNL
Title : Data-Intensive Science Research and Development at Berkeley Laboratory
Speaker: David L. Brown, Director of the Computational Research Division
Lawrence Berkeley National Laboratory
Berkeley, California, USA
Bio: http://crd.lbl.gov/about/staff/director/david-l-brown/
Date : July 3rd, 2015
Time : 2:00-3:00 pm
Location : Bâtiment 660 (Digiteo), Université Paris-Sud
Abstract:
Until the last decade, the world of high-performance computing at the U.S. Department of Energy (US-DOE) Laboratories was dominated by large-scale modeling and simulation. In a wide variety of scientific and engineering fields, high-performance computer codes have provided a mechanism for developing an understanding of scientific theories expressed as large systems of equations that can be solved with HPC. More recently, however, scientific experiments and observations have collected or produced quantities of data that can only be analyzed and understood with the aid of computers. As science becomes ever more data-intensive, the role of HPC has broadened to encompass both the theoretical and experimental/observational sides of science.
Computing has now become essential for most aspects of scientific discovery, and Berkeley Lab’s Computational Research Division is playing a key role in helping scientific research activities effectively leverage computing. In this talk, I will discuss the increasing role of computing in the scientific enterprise at our Laboratory as the data intensity grows. I will give examples of the scientific drivers for data-intensive computing, and present an overview of the broad set of research activities in our Division aimed at addressing the challenges that have resulted.
The following topics will be included in this presentation:
• Need for new mathematical developments to provide analysis capabilities for data types and quantities not previously encountered — I will discuss the CAMERA project at Berkeley Laboratory, which brings applied mathematicians and experimental scientists together to address new data analysis challenges at the large US-DOE scientific user facilities
• Combining modeling/simulation with data analysis to enhance scientific discovery — Cosmology and climate researchers routinely combine simulation and data analysis to reach scientific understanding; New mathematical methods for combustion science are under development for rigorously combining mathematical models and experimental data to more accurately determine unknown physical parameters such as chemical reaction rates
• Development of computing infrastructure to support complex data-intensive science workflows — The Tigres project is developing a programming library for composing and executing large-scale data-intensive scientific workflows; the SPOT Suite provides Advanced Light Source beamline scientists with a web-based interface for managing and analyzing a wide variety of experimental data types.
• User-centered software design — It is important to have significant interaction between domain scientists and data scientists in developing effective software tools for data-intensive science.
—————————————————————————
DIGITEO MOULON (build. 660 – Claude Shannon)
Physical address:
Digiteo Moulon, bâtiment 660
Rue Noetzlin
91190 Gif-sur-Yvette
GPS :
Lat : 48.712475 – Long : 2.167059
http://www.inria.fr/en/centre/saclay/overview/practical-info/how-to-reach-the-centre
