In 1991, the University of Tennessee (UT) and Oak Ridge National Laboratory (ORNL) established the Joint Institute for Computational Sciences (JICS) to advance scientific discovery and the knowledge of computational modeling and simulation, and to educate the next generation of researchers in using computational techniques to address scientific problems.
JICS provides a framework for ORNL to work with UT and other university partners to realize three strategic goals:
- Create new modeling and simulation capabilities for computers operating at the petascale and beyond and use these capabilities to solve the most pressing contemporary science and engineering challenges.
- Train scientists and engineers to model and simulate natural and engineered systems on supercomputers and educate a new generation of researchers expert in applying computational simulation in research and education.
- Create a leading cyber infrastructure for science and engineering in the southeastern United States.
JICS is housed in a 52,000-square-foot building on the ORNL campus across the street from the Oak Ridge Leadership Computing Facility. The JICS team includes UT-ORNL Joint Faculty appointees, JICS Research Affiliates, postdoctoral fellows, graduate students, and administrative staff. JICS Joint Faculty members hold a dual position as faculty members within a University of Tennessee department and as staff scientists within an ORNL research group. JICS also has scientific support staff to aid JICS affiliates and students in their use of leading edge computational resources managed by NICS.
JICS is also home to the National Institute for Computational Sciences (NICS), a National Science Foundation-funded center for high-performance computing managed by UT. The mission of NICS is to enable the scientific discoveries of researchers nationwide by providing cutting-edge computational resources and education, outreach and training for under-represented groups. NICS currently manages some of the main NSF TeraGrid computational resources, including the Cray XT5 called Kraken. With over 98 thousand computational cores, Kraken is the world’s largest computer managed by academia.
For additional information, see http://www.jics.utk.edu.
JICS places emphasis on five areas of research:
Applied Computational Mathematics
Research in Applied Computational Mathematics includes working on important national priorities with advanced computing systems, working cooperatively with U.S. Industry to enable efficient, cost-competitive design, and working with universities to enhance science education and scientific awareness. Research in this domain includes numerical analysis; computational methods for solving partial differential equation; multi-resolution analysis and numerical algorithms; numerical methods for high-dimensional problems; multi-scale computations and algorithms; applied stochastic analysis; Monte Carlo methods and sampling algorithms; coarse-graining methods; statistical mechanics; computational methods in materials science; complex fluids and polymers; analysis of large data sets; and parallel computing.
Computational Biology and Bioinformatics
Research in Computational Biology and Bioinformatics derives new knowledge about biology and biological processes through computational genomic approaches. Computational analyses are applied to best-characterized biological systems in order to reveal novel functional features that cannot be obtained by experimental techniques alone. In this work, an array of bioinformatics tools are used—from gene finding and similarity searches to phylogenetics and structure prediction—and several tiers of hardware—from workstations and stand-alone servers to Linux clusters and supercomputers sifting through trillions of letters of DNA and protein sequences in search of answers to basic biological questions. Research focuses on fundamental biological processes, such as signal transduction, gene regulation and protein-protein interactions studied through the prism of molecular evolution. Most JICS Computational Biology and Bioinformatics studies generate testable hypotheses that are often taken directly into experiment in “wet” laboratories. Computational biology research leads to better understanding of biological systems and has direct applications to medicine, environment, bioenergy, and agriculture.
The overall mission of the Visualization research area is to help researchers gain a better understanding of their data through visualization techniques. Visualization specialists seek out and engage with projects at ORNL and with collaborators who might benefit from applying visual data techniques to scientific data and to find ways of doing visualization that are different, more effective and better integrated with other research activities at ORNL. Work in Visualization at JICS focuses on the display of biomedical and scientific data, interactive simulation, and 3D rendering. Questions of interest include: What does the brain look like as it responds to injury? How can computational simulations of physical processes be visualized? What is the essence of visualizing abstract data?
Computational chemists at JICS develop and apply new capabilities for the understanding, prediction and control of chemical processes ranging from the molecular to the nanoscale, and beyond, using a multidisplinary approach that integrates chemistry, physics, materials science, mathematics, and biology.
Material Science research focuses on theoretical and computational mechanics of crystal plasticity at small length scales, microstructural evolution, and the constitutive behavior of amorphous alloys. Understanding the constitutive response and failure behavior of advanced materials is an issue of fundamental importance for applications not only in the traditional aerospace and automotive industries, but also in modern electronic and energy applications. ORNL has a wide range of user facilities and research programs that offer multi-length-scale and multi-time-scale experimental capabilities, which are commensurate with the microstructural length scales in many advanced materials of interest to the DOE. Theoretical and computational modeling of these experiments will enable the development of mechanism-based material theory with which materials scientists and engineers can have direct design and predictive capabilities. The access to the unique computational facilities in JICS enables the development of petascale and beyond computer simulation codes that can be projected to the future machines. On the other hand, the computational capabilities in these leadership-computing facilities allow us to explore and extrapolate the predictive capabilities of molecular and other first-principle simulations to the modern experimental resolutions. Along these lines, working in the intimate collaboration between domain scientists, computational material scientists, computer scientists and architects, as provided in JICS, will help achieve goals that would be very difficult to reach separately.
Tony Mezzacappa, Director
Joint Institute for Computational Sciences
Oak Ridge National Laboratory
P.O. Box 2008
Oak Ridge, TN 37831-6173