Skip to navigation Skip to content
Careers | Phone Book | A - Z Index

CRD Staff Have Strong Presence at SIAM Conference

February 15, 2007

This year’s SIAM Conference on Computational Science and Engineering (CSE) featured a strong showing from CRD scientists, who spent a week in Costa Mesa, California this month to discuss topics that ranged from nanoscience simulations to scientific data management. SIAM is the Society for Industrial and Applied Mathematics.

In all, 13 researchers from the division attended the week-long conference, which also ran concurrently with the SIAM Workshop on Combinatorial Scientific Computing.

“The large number of contributions to one of the key conferences in the field shows the breadth of research and expertise we have in computational science and engineering,” said Horst Simon, Associate Lab Director of Computing Sciences and one of the conference attendees.

SIAM’s Activity Group on Computational Science and Engineering—currently chaired by CRD’s John Bell—organized this biennial conference. Bell also is chairing a session on fluid dynamics at the conference.

CRD scientists organized mini-symposiums, presented papers and discussed posters during the conference and workshop. They were:

Juan Meza, who discussed “Surface Structure Determination of Nanostructures Using a Mesh Adaptive Optimization Method.” The work examined a method of determining the atomic configuration of a nanostructure’s surface that serves as a cost-effective alternative to the common low-energy electron diffraction (LEED) method. Meza proposed the use of generalized pattern search methods in combination with a simplified physics surrogate for the full fidelity physics model.

Meza also will present a poster by CRD’s Zhengji Zhao titled, “Linear Scaling 3D Fragment Method for Petascale Nanoscience Simulations.” The poster described an O(N) method that is well suited for petascale nanoscience simulations. The method divides a large system into small fragments. It solves the electronic wavefunctions of each fragment independently, and then it patches the results together.

Esmond Ng, a featured speaker in the Combinatorial Scientific Computing workshop and gave the talk, “Towards Optimal Petacale Simulations.” The SciDAC-funded project aims to develop scalable solvers that will allow large-scale scientific simulation codes to run on petascale computer architectures.

Arie Shoshani, who was invited to speak on “Scientific Data Management: Essential Technology for Data-Intensive Science.” Shoshani talked about technological advancements that allow researchers to effectively mine for and analyze data when they have to wade through very large datasets. He also organized the mini-symposium, “Data Management for Scientific Applications,” which described the technology for, among other things, high bandwidth parallel file systems, efficient parallel statistical computing and automation of large scientific workflows.

Phil Collela, who moderated the panel, “Graduate Education Session: Career Path in Computational Science and Engineering – Part II.”

Andrew Canning, who discussed “New Eigensolvers for Large Scale Nanoscience Simulations.” He presented a comparison of some different eigensolvers for the solution of electronic structure calculations using semi-empirical potentials. He illustrated the approach with some applications in nanoscience, such as quantum dots and quantum wires. Canning also presented a poster by CRD’s Lin-Wang Wang titled, “Atomistic Psuedopotential Simulation of Nanometer Sized CMOS Devices.”

Xiaoye “Sherry” Li, who presented the work, “Extra-Precise Iterative Refinement for Least Squares Problems.” Linear least squares (LLS) fitting is the most widely used data  modeling technique and is included in almost every data analysis system (e.g., spreadsheets). With limited use of extra precision, for all but the most ill-conditioned LLS problems, the technique can eliminate the concerns of the floating-point calculation errors present in the solution. Li also chaired a track titled, “Sparse Direct Methods.”

Ali Pinar, who delivered a talk titled, “Vulnerability Analysis on the Electric Power Grid.” Pinar talked about a mixed integer nonlinear formulation for the vulnerability analysis of electric power systems, and how the special structure of this problem can be exploited to further educe complexity.

Christof Voemel, who presented “The Use of Bulk Information to Improve the Scalability of Parallel Band Gap Computations for Quantum Dot.” The calculation of photoluminescence properties of semiconductor quantum dots can be accelerated three- to four-fold by exploiting the physical relationships between the interior of a dot and an ideal crystal. This talk described the corresponding computational procedure in the framework of preconditioned parallel iterative eigensolvers for interior eigenstates of the quantum dot Hamiltonian.

John Shalf and Lenny Oliker, who organized the mini-symposium, “Beyond Petaflops: Specialized Architectures for Power Efficient Scientific Computing.” The discussion examined the potential of custom supercomputing platforms using embedded systems technology, which holds the promise of multi-petaflop computing at a fraction of the cost and power consumption of the conventional HPC design.

As part of this track, Oliker presented a talk titled, “Towards Ultra-High Resolution Models of Climate and Weather.” He discussed the use of embedded processor technology for building a supercomputer that is specifically designed to create kilometer-scale atmospheric models. Generating those simulations would require a sustained 10-petaflop computer, and using customized processors can lead to a more power-efficient system.

Panagiotis Stinis, who chaired the track on fluid dynamics and gave a talk on “Long Memory Mori-Zwanzig Models for the Euler Equations.” The talk presented a new class of reduced models for the 3D Euler equations. The models are based on the long memory caused by the appearance of organized vortical structures.

About Berkeley Lab

Founded in 1931 on the belief that the biggest scientific challenges are best addressed by teams, Lawrence Berkeley National Laboratory and its scientists have been recognized with 16 Nobel Prizes. Today, Berkeley Lab researchers develop sustainable energy and environmental solutions, create useful new materials, advance the frontiers of computing, and probe the mysteries of life, matter, and the universe. Scientists from around the world rely on the Lab’s facilities for their own discovery science. Berkeley Lab is a multiprogram national laboratory, managed by the University of California for the U.S. Department of Energy’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