Valerio Pascucci

Multi-scale Morse Theory and Data Streaming for Science Discovery

Advanced techniques for understanding large scale scientific data are a crucial ingredient for the success of any supercomputing center and therefore for the advancement of most modern science. Developing such techniques involves a number of major challenges including the real-time management of massive data, the quantitative analysis of scientific features of unprecedented complexity, and the effective communication to scientists, decision makers and the general public. Addressing these challenges requires interdisciplinary research in diverse topics including the mathematical foundations of data representations, the design of robust, efficient algorithms, the integration with relevant applications in physics, biology, or medicine, and the construction of the proper level of abstractions enabling succinct and trustworthy exchange of information.

In this talk, I will present a discrete topological framework for the representation and analysis of large scale scientific data. Due to the combinatorial nature of this framework, we can implement the core constructs of Morse theory without the approximations and instabilities of classical numerical techniques. The robustness of this framework allows providing the user with formal guarantees on the quality of the results generated and the implied conclusions. Topological cancellations are used to build multi-scale representations that capture local and global trends present in the data and allow presenting the output models at the proper level of abstraction.

To address the data management problem, we introduced a novel cache oblivious data layout that enables high performance selective queries on multiple terabytes of raw data. The combination of this data access methods with progressive streaming techniques allows to manage massive amounts of data at interactive rates on simple workstations or laptop computers.

The effectiveness of the proposed system has enabled the first successful quantitative analysis of several massively parallel simulations including results that help understanding the causes of global climate change, the design of high efficiency/low emission combustion processes, the structural properties of porous media collecting micrometeoroid in space missions, and the turbulent mixing layer of hydrodynamic instabilities arising in supernova explosions and inertial confinement fusion.

During the talk, I will provide a live demonstration of some software tools developed in this effort and conclude with a discussion of the organizational aspects that enabled the development of new techniques and their successful deployment to the hands of the users for science discovery.

Short Bio

Valerio Pascucci is an Associate Professor of Computer Science at the Scientific Computing and Imaging Institute and School of Computing of the University of Utah. Previously, Valerio was a Group Leader and Project Leader in the Center for Applied Scientific Computing at the Lawrence Livermore National Laboratory, and Adjunct Professor of Computer Science at the University of California Davis. Prior to his CASC tenure, he was a senior research associate at the CS and TICAM Departments of the University of Texas at Austin. Valerio’s research interests include scientific data analysis and visualization, progressive multi-resolution techniques, discrete topology, computer graphics, computational geometry, geometric programming, and solid modeling. Valerio is coauthor of more than one hundred refereed journal and conference papers and is an Associate Editor of the IEEE Transactions on Visualization and Computer Graphics. For more information see