In Situ Processing and Visualization for Peta- and Exa-Scale Simulations

点击翻译按钮获取中文摘要

"This award is funded under the American Recovery and Reinvestment Act of 2009 (Public Law 111-5)."The proposed research will realize in-situ visualization at extreme scale, and deliver a feasible data-understanding solution to scientists whose work will rely on extreme-scale supercomputing, such as those in the NSF PetaApps program. To make the in-situ approach feasible, project will have to either transform existing visualization algorithms or develop a set of new algorithms. These algorithms must be not only scalable but also low cost. The proposed work will build on the PI?s prior success and on-going effort in developing visualization algorithms and systems, as well as his long track record of successful collaboration with application scientists in a variety of areas. The PI directs the Institute for Ultrascale Visualization, a 5-year project funded by the DOE SciDAC program to address the data visualization challenges of petascale computing. The proposed project will be able to leverage, in particular, the Institute?s research in scalable parallel visualization and I/O. This leverage is crucial since it allows this project to focus on application requirements, hardware capabilities/limitations, integration, evaluation, and coupling with the simulation. Broader Impact: The key impact anticipates is to achieve is the demonstration of a working and attractive solution for scientists to possibly comprehend the vast amounts of data generated by petacale computing.PI's won?t have to store petascale data for validating, verifying, and analyzing petascale simulations. This project will provide training for graduate students in the area of computational science and high-performance computing. The new algorithms and software tools developed in the project will be disseminated through the annual visualization workshops and tutorials that the PI has been offering to both the supercomputing and visualization communities.