[Sandia National Laboratories]

[navigation panel]

Karen Devine

Center for Computing Research
Sandia National Laboratories
MS 1318, Org. 1426
P.O. Box 5800
Albuquerque, NM 87185-1318

My CCR webpage
(505) 845-7585

  • High performance computing (HPC): applications and algorithms
  • Parallel partitioning and load balancing
  • Task mapping and placement
  • Parallel graph algorithms
  • Combinatorial scientific computing
  • Parallel tensor computations
  • STEM Outreach
  • Current Projects:

  • Zoltan and Zoltan2 Parallel Data Management and Dynamic Load-Balancing Toolkit
    Toolkits of algorithms for combinatorial scientific computing, including dynamic partitioning/load-balancing algorithms, graph coloring, matrix ordering, and support services for dynamic, unstructured and adaptive applications.

  • FASTMath
    A DOE SciDAC Institute focusing on the development and analysis parallel enabling technology for scientific applications on leadership-class parallel computers. Topic areas include unstructured and structured mesh technology; linear, nonlinear and eigensolvers; architecture-aware tools.

  • Trilinos
    Libraries of linear, nonlinear and eigensolvers; discretatization tools; graph algorithms; partitioning and load balancing.

  • Past Projects:

  • MR-MPI: MPI-based MapReduce Library
  • CSCAPES SciDAC2 Institute: Combinatorial Scientific Computing and Petascale Simulations
  • ITAPS SciDAC2 Center: Interoperable Technologies for Advanced Petascale Simulations
  • MPSalsa: A Parallel Finite Element Program for Simulating Chemically Reacting Flows
  • Dynamic Load Balancing for Adaptive Numerical Methods
  • Selected Publications:

  • Scalable Matrix Computations on Scale-Free Graphs Using 2D Graph Partitioning
    E.G. Boman, K.D. Devine, S. Rajamanickam
    International Conference for High Performance Computing, Networking, Storage and Analysis (SC13), 2013.
    Preprint (PDF)

  • MapReduce in MPI for Large-Scale Graph Algorithms
    S. J. Plimpton, K. D. Devine
    Parallel Computing, 37, 610-632 (2011).
    Abstract (HTML)      Preprint (PDF)

  • Multi-Jagged: A Scalable Parallel Spatial Partitioning Algorithm
    M. Deveci, S. Rajamanickam, K.D. Devine, U.V. Catalyurek
    IEEE Trans. on Parallel and Distributed Systems, Article No. 10.1109/TPDS.2015.2412545, Accepted/Published March 2015

  • Exploiting Geometric Partitioning in Task Mapping for Parallel Computers
    M. Deveci, S. Rajamanickam, V.J. Leung, K.T. Pedretti, S.L. Olivier, D.P. Bunde, U. Catalyurek, K.D. Devine
    IEEE International Parallel and Distributed Processing Symposium IPDPS14, May 2014

  • Demonstrating Improved Application Performance Using Dynamic Monitoring and Task Mapping
    J.M. Brandt, K.D. Devine, A.C. Gentile, K. Pedretti
    Workshop on Monitoring and Analysis for High Performance Computing Systems Plus Applications (in conjunction with IEEE Cluster 2014), September 2014

  • Parallel Graph Coloring for Manycore Architectures
    M. Deveci, E.G. Boman, K.D. Devine, S. Rajamanickam
    IEEE International Parallel and Distributed Processing Symposium IPDPS16, May 2016

  • Hypergraph-based Dynamic Load Balancing for Adaptive Scientific Computations.
    Best Algorithms Paper Award
    U. Catalyurek, E. Boman, K. Devine, D. Bozdag, R. Heaphy, L.A. Riesen.
    IEEE International Parallel and Distributed Processing Symposium IPDPS07, March 2007.
    Abstract (HTML)      Preprint (PDF)      Citation (BIBTEX)

  • Software Design for Scientific Computing
    SIAM Annual Meeting Invited Plenary Talk, July 2008.
    Abstract      Presentation (PDF)

  • Recent publications

  • Older publications

  • Other Links:

  • My CCR webpage
  • Sandia's Center for Computing Research
  • Sandia's Computer Science Research Institute

  • Karen D. Devine
    November 2017
    © 2000 National Technology and Engineering Solutions of Sandia, LLC