SciDAC at CSC

Homepage

Scientific Discovery through Advanced Computing (SciDAC) was recently renewed as a five-year program to develop the Scientific Computing Software and Hardware Infrastructure to support petascale computations for DOE's research programs.

ITAPS 
J. Glimm, R. Samulyak, Z. Xu

At Brookhaven we have formed an alliance with researchers at Argonne, Livermore, Oak Ridge, Pacific Northwest, Rensselaer, Sandia, and Stony Brook, to develop technologies that enable scientists to use complex mesh and discretization strategies easily and interchangeably, within a single simulation on terascale computers. The Interoperable Technologies for Advanced Petascale Simulations (ITAPS) Center, and its predecessor, the Terascale Simulation Tools and Technologies (TSTT) Center, provides interoperable tools to facilitate use of advanced mesh and discretization technologies. The Center has developed standardized interfaces to local mesh refinement codes. Existing codes of the Center partners and new codes are supported to create a plug and play capability, whereby an application user can experiment easily with alternative technologies. Insertion of these tools into targeted applications, including nuclear energy fusion, accelerator design, and ground water modeling, is part of the Center work plan.

CSC work within ITAPS currently focuses on jet breakup, spray formation, bubbly flow for nuclear energy applications, electromagnetic simulations for accelerator design, and on finite element and front tracking contributions to ITAPS technology.

The ITAPS website is http://www.tstt-scidac.org.

TOPS
D. Keyes

Terascale Optimal PDE Simulations (TOPS) is a SciDAC project connected to Brookhaven. TOPS software is being applied in CSC through a collaboration with one of its principal investigators. TOPS deploys a toolkit of open source solvers for partial differential equations, large systems of stiff ordinary differential equations, and linear and nonlinear algebraic systems, including eigenvalue problems, that arise in application areas such as accelerator design, biology, chemistry, magnetohydrodynamics, and particle physics. Scalable solution algorithms – primarily multilevel methods -- aim to reduce computational bottlenecks by one or more orders of magnitude on terascale computers, enabling scientific simulation on a scale heretofore impossible. Along with usability, robustness, and algorithmic efficiency, an important goal is to attain high computational performance by accommodating to distributed hierarchical memory architectures.

The convergence rates of solvers traditionally employed in PDE-based codes degrade as the size of the system increases. This creates a double jeopardy for applications -- as the cost per iteration grows, so does the number of iterations. Fortunately, the physical structure of PDE problems, such as Poisson's equation for electrostatic potential, provides a natural way to generate a hierarchy of approximate models, through which the required solution may be obtained efficiently.

The efforts defined for TOPS and its collaborations incorporate existing and new optimal algorithms into scientific applications through code interoperability behind a standard interface. TOPS provides support for the software packages Hypre, PETSc (which has powered three Gordon Bell Prizes in recent years), Sundials, SuperLU, TAO, and Trilinos. Some of these packages are in the hands of thousands of users, who have created a valuable experience base on thousands of different computer systems.

The TOPS project webpage may be found at: http://www.scidac.gov/math/TOPS.html.

Advanced Computing for 21^st Century Accelerator Science and Technology
R. Samulyak
 

The SciDAC Accelerator Modeling Project, “Advanced Computing for 21st Century Acceleratory Science and Technology,” was initiated in June 2001. Its primary goal is to establish a comprehensive terascale simulation environment for use by the U.S. particle accelerator community. Building upon a previous DOE Grand Challenge project as well as previous individual efforts at several national laboratories and universities, the SciDAC Accelerator Modeling Project represents the largest effort to date for the development of computer codes for accelerator design and analysis. The activities of the project are organized into three application-specific focus areas: Electromagnetics, Beam Dynamics, and Advanced Accelerators. Work in these areas is supported by collaboration with the SciDAC Integrated Software Infrastructure Centers (ISICs) and by personnel (including CSC/BNL) supported through the SciDAC Scientific Application Partnership Program (SAPP).

Research at CSC/BNL is conducted in close collaboration with the Beam Dynamics group. The primary goal of the study is the development of novel mathematical models and software modules for the computation of wake fields and their interaction with particle beams in high intensity accelerators. We have implemented our software in the MaryLie/Integrated Map and Particle Accelerator Tracking code, a parallel code that combines the magnetic optics capabilities based on the Lie algebraic technique with the 3D space charge capabilities and the Synergia framework.

The SciDAC website is http://www.scidac.org

Statistical Approaches to Aerosol Dynamics for Climate Simulation
R. McGraw and W. Zhu

The SciDAC climate modeling project, “Statistical Approaches to Aerosol Dynamics for Climate Simulation,” was recently awarded five years funding.

This research supports the goal of the DOE Climate Change Prediction Program (CCPP), which is to determine the range of possible climate changes over the 21st century and beyond using a more accurate climate system model. In partnership with the Applied Mathematics and Statistics Department at SUNY Stony Brook, we develop new statistical approaches for improving the representation of aerosols, aerosol microphysical processes, and aerosol-cloud interactions in the Community Climate System Model (CCSM). The new approaches, which track multivariate moments of the particle population, are highly efficient, yet provide the comprehensive representation of natural and anthropogenic aerosols, and of their mixing states and direct and indirect effects, that the CCSM will require. Findings from this study will be incorporated into a new aerosol microphysical module in time for CCSM5.

Last Modified: January 31, 2008
Please forward all questions about this site to: Claire Lamberti

One of ten national laboratories overseen and primarily funded by the Office of Science of the U.S. Department of Energy (DOE), Brookhaven National Laboratory conducts research in the physical, biomedical, and environmental sciences, as well as in energy technologies and national security. Brookhaven Lab also builds and operates major scientific facilities available to university, industry and government researchers. Brookhaven is operated and managed for DOE’s Office of Science by Brookhaven Science Associates, a limited-liability company founded by Stony Brook University, the largest academic user of Laboratory facilities, and Battelle, a nonprofit, applied science and technology organization.

Privacy and Security Notice  | Contact Web Services for help