Seminars archive
Upcoming Seminars   No upcoming seminars currently scheduled.  Past Seminars   Seminar: Numerical Analysis and Scientific Computing Adaptive spline approximation: error analysis and construction of the partitions Yuliya Babenko, Kennesaw State University, GA   Seminar: Algebra The pretentious view of analytic number theory Robert Lemke Oliver, Emory University   Seminar: Algebra and Number Theory Visibility of Torsors of an Abelian Variety Saikat Biswas, Georgia Tech   Seminar: Algebra and Number Theory Crystalline cohomology of the Igusa tower and families of ordinary cuspforms Bryden Cais, University of Arizona   Seminar: Numerical Analysis and Scientific Computing Topological and Functional Properties of Proteins in ProteinProtein Interaction Networks Concettina Guerra, College of Computing, Georgia Institute of Technology, USA   Seminar: Computer Science Sensor Web: Research Challenges and Opportunities WenZhan Song, Georgia State University   Seminar: Algebra and Number Theory A characterization of cusp forms by means of the growth of their Fourier coefficients Winfried Kohnen, Heidelberg University   Seminar: Scientific Computing Seminar Gossipbased distributed matrix computations Hana Strakova, University of Vienna and GA Tech Venue: Mathematics and Science Center, Room W306 Show abstract Truly distributed matrix computations with randomized communication schedules, such as gossipbased algorithms, can offer many
attractive properties. Due to their randomized communication restricted only to direct neighbors they are
very flexible with respect to the underlying hardware infrastructure. They can operate on arbitrary topologies and they can be made
resilient against dynamic changes in the network, against message loss or node failures, and against asynchrony between compute nodes. Moreover, their overall cost can be reduced by accuracycommunication
tradeoffs. Such properties are attractive especially for looselycoupled distributed systems with unreliable communication
links, such as sensor or P2P networks. However, due to the growth in the number of nodes for future extremescale HPC systems and the anticipated decrease in reliability, some properties of gossipbased distributed algorithms may become important also for future HPC systems.
We are investigating distributed algorithms for various prototypical matrix computation problems which utilize gossipbased aggregation algorithms for performing reduction operations in a distributed manner. Questions addressed relate to the (communication) cost paid for the increased flexibility and robustness, to convergence properties, to numerical accuracy achieved, as
well as to the benefits of accuracycommunication cost tradeoffs.   Seminar SIAM Student Chapter: Numerical Analysis and Scientific Computing Introduction to GPU Computing: Basics of OpenCL Veronica Mejia Bustamante, Emory University   Seminar: Combinatorics Induced and noninduced Ramsey numbers of $k$partite, $k$uniform hypergraphs Steve La Fleur, Emory University 
