Elisabeth Larsson's talk: "Using data versioning for dependency-aware task-based parallelisation"
HPC research group invites you to attend the following talk:
Title: Using data versioning for dependency-aware task-based parallelisation
Speaker: Elisabeth Larsson (Uppsala University)
Date: Fri, 20/Jul/2012, 11:00
In computational science, high performance parallel computing, making efficient use of modern multicore based computer hardware, is necessary in order to deal with complex real-life application problems. However, with increased hardware complexity, the cost in man hours of writing and re-writing software to adapt to evolving computer systems is becoming prohibitive. Task based parallel programming models aim to allow the application programmers to focus on the algorithms and applications, while the performance is handled by a run-time system that schedules the tasks onto nodes, cores, and accelerators. In this talk we describe a family of task parallel programming models where data dependencies are represented through data versioning. Benefits of using this type of model are that it is easy to represent different types of dependencies and that scheduling decisions can be made locally. We show experimental results indicating that a thread parallel shared memory implementation as well as a hybrid thread/MPI distributed memory implementation scale well.
Elisabeth Larsson got her PhD n Numerical Analysis from Uppsala University, Sweden in 2000. She then spent one year at the University of Colorado at Boulder, USA as a postdoc. After the postdoc, she got position at Uppsala University, where she is now Associate Professor of Scientific Computing. Her main research interests are Radial Basis Function Approximation, High Performance Computing, and Computational Finance. She is Director of the Centre for Interdisciplinary Mathematics (CIM) at Uppsala University and she is also affiliated with the Linnaeus center of excellence UPMARC, Uppsala Progamming for Multicore Architectures Research Center.