Data Analytics & Visualization
The Visualization Group's mission is to assist researchers in achieving their scientific knowledge discovery goals through visualization and analysis, while simultaneously advancing the state of the art through our own research. Their objective is to develop new capabilities in high performance visualization, analysis, and related data-intensive technologies. The development of these capabilities is driven by the needs of contemporary computational and experimental science projects central to the mission of the DOE Office of Science.
Our collaborators are diverse, ranging from theoretical astrophysicists to computational and experimental bioscientists. All have a theme in common: the need to understand complex systems through visual inspection of simulation results.
Our portfolio includes projects in basic and applied research, advanced software development, and deployment to the scientific community. Examples from our portfolio:
- Algorithm development and performance optimization for modern, extreme-scale architectures. One objective is to study how fundamental algorithmic architecture and designs can make effective use of modern computational architectures, through techniques like hybrid parallelism. Recent results show promise for scaling fundamental visualization methods like volume rendering and streamline calculation to very high levels of concurrency, a requirement for transitioning into the exascale regime. Related, our studies have shown that up to a 30x performance gain is possible on some codes and platforms depending upon settings for tunable algorithmic parameters, algorithmic optimizations, and use of device-specific features.
- Visualization and analysis software tool development and deployment. Another group objective is to impact science by enabling scientific knowledge discovery through the use of visualization and analysis software tools. Our portfolio includes productization of many of our research ideas, and we have a long history of doing just that through individual software tools (see the Visualization Group's software page), as well as taking a leadership role in large, multi-institution efforts that serve large computational and experimental science communities like the SciDAC-2 Visualization and Analytics Center for Enabling Technologies (VACET) , which brought production-quality, petascale visualization and analysis capabilities and tools to DOE's supercomputer centers, the SciDAC-3 Scalable Data Management, Analysis, and Visualization Institute (SDAV), which continues this mission through 2017, and the Advanced Simulation Capabilities for Environmental Management (ASCEM) effort, which is developing and deploying an integrated approach for efficiently and cost-effectively minimizing environmental, safety, and health risks associated with the movement and impact of underground contaminants. Many of these efforts involve deployment in VisIt, the large-scale visualization and analysis tool, and our team is active developers of this project.
- Extreme-scale climate data analysis. We are designing, developing, and deploying next-generation tools, which are distributed to the world-wide climate science community, for analyzing and visualizing massive climate data sets to aid in the understanding of how our climate is changing. More information.
- Carbon sequestration. We have designed, developed, and deployed leading-edge computer vision, image processing, and analysis technology to aid science researchers in understanding how to best store carbon dioxide in porous media. More information.
- End-to-end data management, analysis, and visualization of extreme-scale simulation data. Contemporary plasma physics simulations have recently entered the regime of simulating trillions of particles, which produce an extremely large amount of data: a trillion particles of output, or 30 terabytes of data per-timestep. Our team has developed new methods and tools to enable the capture, storage, analysis, and visualization of such data, and demonstrated these methods on modern DOE supercomputing platforms. More information.
- Diverse research and development, highlights of which are located on the Visualization Group's Vignettes pages, on topics ranging from algorithmic development in visualization and analysis, to applying these techniques and technologies to solve specific science knowledge discovery problems.
Group Leader: Wes Bethel
» Visit the Data Analytics & Visualization Group site.
The Visualization Group's mission is to assist researchers in achieving their scientific knowledge discovery goals through visualization and analysis, while simultaneously advancing the state of the art through our own research. Their objective is to develop new capabilities in high performance visualization, analysis, and related data-intensive technologies. The development of these capabilities is driven by the needs of contemporary computational and experimental science projects central to… Read More »
Visualization plays an integral role in the scientific process--allowing a way to see the unseen by creating images of experimental data or theoretical simulation results. The projects listed on this page, which date from 1993 to the present, contain highlights recent or current LBNL Visualization Group projects, many of which were performed in collaboration with science stakeholders. The full set of Vignettes is located on the LBNL Visualization Group… Read More »