DOE - Advanced Computational Testing and Simulation Research

DOE Advanced Computational Testing and Simulation Research Budget Code: KJ0101, KJ3501

This activity has three fundamental research components: Applied Mathematics, Computer Science, and Software Tools. In addition, this activity supports education activities which are focused on educating the next generation of computational scientists.

The applied mathematics research component develops theory, algorithms, and tools for enabling the solution of large scientific and engineering problems. This component supports leading-edge research at ten DOE laboratories and over 30 universities. Applied mathematicians and computational scientists supported by the basic research program are also active in Grand Challenge projects and other multidisciplinary research projects. Supported laboratory and university researchers meet every two years at a workshop designed to foster collaborations and idea exchanges.

The computer science and software tools component of this effort focuses on developing advanced software to facilitate the use of high performance systems to scientific problems. Efforts range from efficient operating systems and I/O software for MPP's, to frameworks for isolating application codes from the underlying hardware details, to tools for monitoring the performance of scientific applications. In addition, this component supports efforts to improve the management, visualization, and understanding of the results of high end computations.

The education programs focus on engaging and training students, teachers, and faculty from middle and high school through graduate school in computing, networking, and computational science, by leveraging the large DOE investment in these areas at universities and the national laboratories.

The Computational Science Graduate Fellowship (CSGF) Program supports over 50 doctoral students in computational science and engineering at selected universities. Participating fellows spend at least one summer working at a DOE laboratory in the area of their dissertation.

The Adventures in Supercomputing (AiS) Program trains middle and high school students and their teachers in the use of computers and networks and provides computers and curricular materials to participating schools. AiS encourages the participation of women and under represented minorities.

The Undergraduate Computational Engineering and Science (UCES) Program develops curricular materials and application-based modules for undergraduate computational science courses, tests this material at selected universities, and sponsors workshops and conferences for faculty and students. he UCES Program was expanded to involve the participation of selected HBCU's in teaching demonstrations and the development of curricula and instructional materials.

Model Nets, a networking technology assessment project to determine appropriate educational technology tools and curricula in a variety of school systems, was completed. A catalog of effective technologies for use by teachers, administrators, and local and state education officials is now available in electronic and printed formats.

Budget ($ M)

FY 95 Act

FY 96 Pres 35.81

FY 96 Est 35.81

FY 97 Rqst 36.40

Program Component Areas

FY 96 FY 97

HECC 31.81 32.90

LSN

HCS

HuCS

ETHR 4.00 3.50

Agency Ties

DARPA Partner

NSF Partner

DOE

NASA Partner

NIH

NSA

NIST

NOAA

EPA

ED

AHCPR

VA

Milestone Changes

FY 1995 Actual Milestones FY 1996 Estimated Milestones FY 1997 Agency Requested Milestones

Relocated the Lawrence Livermore National Laboratory (LLNL) applied mathematics group to Lawrence Berkeley National Laboratory (LBNL) as part of the NERSC relocation.

Computational scientists at Los Alamos National Laboratory (LANL) and LBNL, working with university partners and a commercial combustion firm, released a code based on adaptive mesh refinement techniques that simulates 3D combustion phenomena in complex geometries, with radiation effects included, accurately enough for the code to be used in the design of commercial low NOx burners.

Applied mathematicians at Argonne National Laboratory (ANL), in partnership with researchers at nearby Northwestern University, set up the Optimization Technology Center (OTC) at Northwestern, in order to enhance laboratory/university collaborations.

On-line services at the OTC enable researchers around the country to access software for solving large-scale optimization and programming problems.

Oak Ridge National Laboratory (ORNL) researchers developed and distributed improved codes for sparse matrix calculations that enabled progress on various Grand Challenge problems.

Advances in modeling and asymptotic techniques allowed LBNL applied mathematicians to correct long-standing errors in commonly-used turbulence formulas for viscous flows near solid boundaries.

LANL computational physicists, in partnership with SUNY Stony Brook applied mathematicians, extended the ability of front tracking codes to simulate interfacial phenomena of importance in fusion and weapons research.

Transfer improved operating systems and message passing models to U.S. computer vendors.

Investigate scalability of existing performance analysis tools and deliver beta versions of high performance scalable I/O libraries.

Develop plan to integrate scalable I/O with distributed and archival file and data systems.

Evaluated the quality of the first and second group of CSGF fellows as they enter the professional computational science career path.

Made general releases of updated graduate and undergraduate computational science texts available.

The AiS Program extended its computing and networking activities to middle school students at selected sites. Reduce support for the Ames Laboratory and Brookhaven National Laboratory applied mathematics programs as part of an overall redirection of funds into more computationally intensive areas.

LBNL and LANL computational combustion researchers, along with their university partners, will initiate simulations of internal combustion engines, in collaboration with industry partners.

Continue collaboration between ANL applied mathematicians and the DOE Office of Integrated Analysis and Forecasting on improving the predictive capabilities of the National Energy Modeling System.

Initiate research at LANL on the predictability of catastrophic events such as wildfires and earthquakes through the development of modeling and simulation tools.

Strengthen collaborative ties between laboratory and university researchers working on Grand Challenge problems and other multidisciplinary projects important to the DOE mission. Establish interoperability of MUSE and CAVE virtual reality technologies to explore domains of most advantageous use of these two technologies.

Deploy advanced scientific data management technologies to scientific users.

Evaluate strategies for producing large applications software packages that can be executed in environments ranging from networks of workstations to the highest performance massively parallel processors available.

Introduce advanced tools for parallel program diagnosis and tuning in preproduction versions.

Begin development of software tools which are designed around the Advanced Computational Testing and Simulation Toolkit interface definitions.

Prototype integrated, distributed multimedia scientific visualization environment for DOE researchers. Continue evaluation of the effectiveness of the CSGF Program.

Complete electronic computational science text projects through commercialization.

Replace Macintosh computers used in the AiS Program with PC's running LINUX software, in order to enhance local computing capabilities and train students in distributed computing.

Initiate follow-on program to the Model Nets Project that focuses on training in-service teachers and administrators to use and maintain local- and wide-area networks in their schools.