InTheLoop | 11.10.2014
Computational Research Division Reorganized for Strategic Alignment & Clarity
Computational Research Division Director David Brown last week announced a reorganization of the Computational Research Division at the October CRD All-Hands meeting, creating four new departments aligned with CRD’s strategic directions. Three of the four departments recognize the Laboratory’s significant strengths in applied mathematics, computer science and data science and technology research. A fourth Department, Computational Science, reflects the increasing emphasis of high performance computing in scientific discovery, and the strong computational research efforts within CRD in a number of important scientific areas.
Brown also called out the new Center for Advanced Mathematics for Energy Research Applications (CAMERA), directed by James Sethian. CAMERA is an integrated, cross disciplinary center aimed at inventing, developing the fundamental new mathematics required to capitalize on experimental investigations at DOE Scientific User Facilities. »Read more.
Fourth International Workshop on Network-Aware Data Management at SC14
NDM'14: the 4th International Workshop on Network-aware Data Management, will be held Sunday, Nov. 16, in conjunction with the SC14 conference in New Orleans. The workshop will be held from 1:30-5:30 p.m. in room 274 of the Ernest N. Morial Convention Center.
The workshop, co-organized by Surendra Byna and Brian Tierney of Lawrence Berkeley National Laboratory and Mehmet Balman of VMware, Inc., who is also affiliated with Berkeley Lab, will feature keynote talks and technical presentations. »Read more.
This Week's CS Seminars
Visual Analytic Methods for Larger Volume of Scientific Imaged Datasets
Monday, Nov. 10, 10–11 a.m., Bldg. 50F, Room 1647
Chiwoo Park, Florida State University
Notable improvements in the temporal and spatial resolution of dynamic transmission electron microscopy allow atomic-scale observations of rapidly changing dynamic processes at high frame rate, such as electrolyte breakdown and dendrite growth in Li-ion batteries, metal organic framework formation for hydrogen gas storage of a hydrogen-powered vehicle, and protein folding. However, due to the lack of a big data analysis method for handling a few terabytes or petabytes of high frame rate data, the high frame rate microscopic observations have not been fully exploited for fundamental scientific studies. In this talk we present our new big data framework to automate the analysis of the high frame rate microscope data, which is composing of fast image segmentation algorithms and a large-scale data association solver. We will illustrate the use of the framework for analyzing nanoparticle self-assembly processes.
Applied Mathematics: A Constraint–based Formulation for Freely Moving Immersed Solid Bodies in Fluids
Wednesday, Nov. 12, 3:30–4:30 p.m., 939 Evan Halls, UC Berkeley
Neelesh Patankar, Northwestern University
Numerical simulation of moving immersed solid bodies in fluids is now practiced routinely. A variety of variants of these approaches have been published, most of which rely on using a background mesh for the fluid equations and tracking the body using Lagrangian points. In this talk, generalized constraint–based governing equations will be presented that provide a unified framework for various immersed body techniques. The key idea that is common to these methods is to assume that the entire fluid–body domain is a “fluid” and then to constrain the body domain to move in accordance with its governing equations. The immersed solid body can be rigid or deforming. The governing equations are developed so that they are independent of the nature of temporal or spatial discretization schemes. Specific choices of time stepping and spatial discretization then lead to techniques developed in prior literature ranging from freely moving rigid to elastic self-propelling bodies. To simulate Brownian systems, thermal fluctuations can be included in the fluid equations via additional random stress terms. Solving the fluctuating hydrodynamic equations coupled with the immersed body results in the Brownian motion of that body. The constraint–based formulation leads to fractional time stepping algorithms à la Chorin–type schemes that are suitable for fast computations of rigid or self-propelling bodies whose deformation kinematics are known. Application of this technique to interrogate aquatic locomotion to eventually enable an understanding of neural control of movement will be briefly summarized.
Link of the Week: Can Scientists Tweet Their Way to Success?
Researchers from the University of Wisconsin-Madison last week published a paper that found a correlation between professional social media exposure of scientists' work and a higher "h-index," a measure of the quality of a researcher’s work and influence. This study was limited to a small sample of university-based nano scientists, but the paper's authors say it opens the door to further inquiry. »Read more.