To address our stakeholders' data understanding challenges, our Center's technology will be scalable to leverage parallel computing resources as well as to perform adequately in resource poor environments. Without scalable tools, it is simply not possible to analyze or visualize some of our stakeholders' data: they are producing data set sizes of unprecedented scale. The Center's visualization primary deployment vehicles, VisIt and SCIRun, have already made substantial investments in scalability. By leveraging those investments, our stakeholders will immediately benefit from their capabilities. Much of the Center's planned work on scalable I/O addresses the growing divide between data access times and compute power. This work includes hierarchical I/O representations that allow for views of multiple resolutions and inlined compression schemes that reduce the total amount of required I/O. In addition to I/O, we will implement techniques to optimize performance in a variety of compute environments. These techniques include out-of-core processing, level of detail processing, general parallel processing, and parallel rendering.
See Gallery: Software