Architectural Challenges and Solutions for Petascale Visualization and Analysis Hank Childs, LLNL

With petascale applications comes petascale data; gleaning knowledge and insight from large-scale data is widely accepted to be a limiting factor in many fields of scientific endeavor. With larger data and platforms of increasing parallelism come significant software architectural challenges in the regime of visual data analysis. Solutions responsive to these challenges require "smarter" approaches, such as in situ and level of detail processing, which are well established by a decade's worth of research. The challenge now is to deploy these approaches, at production quality, to a large and diverse scientific audience. This challenge is complicated by the varied activities that fall under the postprocessing umbrella: data exploration, moviemaking, quantitative analysis, comparative analysis, and visual debugging. In this talk, we will prescribe an approach that uses a combination of "smart" methods to deploy all of these postprocessing techniques at the petascale, both directly inside user code and also via high performing, standalone applications.