Data-intensive applications have become increasingly important in various domains such as bioinformatics and social network analysis. These applications usually involve large amount of data movement, data-driven computation and irregular communication patterns. To achieve high performance, programming environment should be able to provide balancing of communication and computation and low synchronization cost. Traditional data movement approaches for scientific applications are not well suited for such applications. Active Messages (AM) is an alternative model that is better suited for such applications by providing data-driven computation and asynchronous communication. Because MPI is the most widely used message passing model and many applications are already implemented by it, enabling MPI-interoperable active messages will allow existing applications to be incrementally modified to use active messages when necessary, reducing the effort of rewriting the entire applications. In this talk, Xin will introduce characteristics of data-intensive applications and discuss optimization opportunities, and will present our work on the framework of MPI-interoperable and generalized active messages.

Bio:

Xin Zhao is a third-year Ph.D. student from the Department of Computer Science at the University of Illinois at Urbana-Champaign (UIUC),advised by Prof. William Gropp. Her research interests focus on parallel programming models / runtime systems and data-intensive applications, with an emphasis on communication, resources management and dynamic execution. In this summer, she works with Dr. Pavan Balaji at Argonne National Laboratory on proposing MPI-interoperable and generalized Active Messages.