Research Capabilities
The Computational & Mathematics (CSM) Division provides creative scientific and technological solutions to challenges in problems of national and global importance by pursuing a robust portfolio of fundamental and applied research at the frontiers of computational science, computer science, and mathematics.
- Computational Biology & Bioinformatics
Advances in the computational modeling and simulation of complex biological systems are transforming biological research from a qualitative, descriptive science to a quantitative, predictive science. PNNL's focus is on the design and efficient implementation of computational capabilities for the analysis of data from high-throughput experimental technologies, the abstraction of models from this data, and the predictive simulation of these models. Beyond the validation of experimental observations, these simulations enable the design and prediction of the outcome of new experiments. This is an essential part of the scientific discovery cycle aiming at the development of technical approaches to bioremediation, bioenergy production, and climate management. - High-Performance Computing
Employing high-performance computing to solve scientific problems by developing and implementing high-level programming abstractions and high-speed networks and communication tools. Our approach merges science and technology by 1) employing hardware that maximizes processor speed, memory and interconnect bandwidth, efficient use of secondary storage, and reliability; 2) developing algorithms that are scalable, resource-efficient, and load-balanced and that manage computational complexity and exploit space-time locality; and 3) creating programming models, numerical libraries, communication libraries, compilers, and debuggers that support data decomposition, low communication overhead, and portability. We are also leading providers of problem-solving environments that increase ease of use and availability of high-performance computing to nonspecialists. - Scientific Data Management
Innovating solutions that address challenges arising from exponential increases in the size and complexity of data. We deliver solutions that facilitate data collection, processing, and storage through engineered systems and applied research. In particular, our efforts focus on techniques to enable the overall management of tera-scale and larger scientific datasets exhibiting one or more of these characteristics:
1) distributed data, requiring timely aggregation of information from multiple locations; 2) heterogeneous data, requiring integration of information from multiple disparate, semantically inconsistent formats; and 3) temporal data, requiring analysis and processing of streaming data in near real-time.
POC: Terence Critchlow and Todd Halter - Computational Mathematics
Leveraging mathematical models to quantify and control scientific uncertainty to further scientific discovery. Scientific research and development is a process of gaining fundamental understanding of physical, chemical, and biological principles through computational modeling, experimentation, and data evaluation. As a leader in applied mathematics research, we develop novel data-analysis methods to extract hidden features, anomalies, and signatures from high-dimensional, large-volume, multimedia data in support of discovery and confident decision-making. We develop methods and tools to optimize data-gathering approaches through sampling and experimental design. - Modeling and Simulation
Developing and applying modeling and simulation of complex systems to facilitate predictive science. We develop and apply modeling and simulation tools to verify observed phenomena and to predict or explore the behavior of systems where experiment or observation is impractical. Our innovations have resulted in the development of tools for predicting climate change; understanding the molecular processes involved in cell and organism adjustment to changing conditions; predicting the structures, energetics, and reactions of molecules; and designing innovative materials for tomorrow's products. - Scientific Visualization
Scientific visualization is the representation of data graphically as a means of gaining understanding and insight into the data. In fields ranging from scientific research to national security, people must interpret huge volumes of information to identify patterns and trends, uncover surprises, and address emerging problems. Scientific visualization allows the researcher to gain insight into the system that is studied in ways previously impossible. As a leader in scientific visualization research, we develop innovative technologies and processes that enable the analysis of massive amounts of complex data from diverse sources to detect, predict, and visualize common threads and patterns in near real time. POC: Moe Khaleel - Problem Solving Environments
Problem Solving Environments (PSEs) are frameworks that integrate data and information with analytic tools that the analyst can interrogate within a cohesive environment. By providing a convenient, seamless environment for scientists and engineers to manipulate and analyze data as they perform modeling and simulation tasks, scientific discovery is improved.
POC: Eric Stephan