@conference {863, title = {Task-Based Programming for Seismic Imaging: Preliminary Results}, booktitle = {2014 IEEE International Conference on High Performance Computing and Communications (HPCC)}, year = {2014}, month = {2014-08}, publisher = {IEEE}, organization = {IEEE}, address = {Paris, France}, abstract = {The level of hardware complexity of current supercomputers is forcing the High Performance Computing (HPC) community to reconsider parallel programming paradigms and standards. The high-level of hardware abstraction provided by task-based paradigms make them excellent candidates for writing portable codes that can consistently deliver high performance across a wide range of platforms. While this paradigm has proved efficient for achieving such goals for dense and sparse linear solvers, it is yet to be demonstrated that industrial parallel codes{\textemdash}relying on the classical Message Passing Interface (MPI) standard and that accumulate dozens of years of expertise (and countless lines of code){\textemdash}may be revisited to turn them into efficient task-based programs. In this paper, we study the applicability of task-based programming in the case of a Reverse Time Migration (RTM) application for Seismic Imaging. The initial MPI-based application is turned into a task-based code executed on top of the PaRSEC runtime system. Preliminary results show that the approach is competitive with (and even potentially superior to) the original MPI code on a homogeneous multicore node, and can more efficiently exploit complex hardware such as a cache coherent Non Uniform Memory Access (ccNUMA) node or an Intel Xeon Phi accelerator.}, keywords = {plasma}, author = {Lionel Boillot and George Bosilca and Emmanuel Agullo and Henri Calandra} } @article {icl:730, title = {Matrices Over Runtime Systems at Exascale}, journal = {Supercomputing {\textquoteright}12 (poster)}, year = {2012}, month = {2012-11}, address = {Salt Lake City, Utah}, author = {Emmanuel Agullo and George Bosilca and Cedric Castagn{\`e}de and Jack Dongarra and Hatem Ltaeif and Stanimire Tomov} } @inproceedings {icl:605, title = {Algebraic Schwarz Preconditioning for the Schur Complement: Application to the Time-Harmonic Maxwell Equations Discretized by a Discontinuous Galerkin Method.}, journal = {The Twentieth International Conference on Domain Decomposition Methods}, year = {2011}, month = {2011-02}, address = {La Jolla, California}, url = {http://hal.inria.fr/inria-00577639}, author = {Emmanuel Agullo and Luc Giraud and Amina Guermouche and Azzam Haidar and Stephane Lanteri and Jean Roman} } @article {icl:653, title = {A Hybridization Methodology for High-Performance Linear Algebra Software for GPUs}, journal = {in GPU Computing Gems, Jade Edition}, volume = {2}, year = {2011}, month = {2011-00}, pages = {473-484}, publisher = {Elsevier}, keywords = {magma, morse}, author = {Emmanuel Agullo and Cedric Augonnet and Jack Dongarra and Hatem Ltaeif and Raymond Namyst and Samuel Thibault and Stanimire Tomov}, editor = {Wen-mei W. Hwu} } @article {icl:599, title = {LU Factorization for Accelerator-Based Systems}, journal = {IEEE/ACS AICCSA 2011}, year = {2011}, month = {2011-12}, address = {Sharm-El-Sheikh, Egypt}, keywords = {magma, morse}, author = {Emmanuel Agullo and Cedric Augonnet and Jack Dongarra and Mathieu Faverge and Julien Langou and Hatem Ltaeif and Stanimire Tomov} } @article {icl:606, title = {Parallel algebraic domain decomposition solver for the solution of augmented systems.}, journal = {Parallel, Distributed, Grid and Cloud Computing for Engineering, Ajaccio, Corsica, France, 12-15 April}, year = {2011}, month = {2011-00}, author = {Emmanuel Agullo and Luc Giraud and Amina Guermouche and Azzam Haidar and Jean Roman} } @article {icl:677, title = {QCG-OMPI: MPI Applications on Grids.}, journal = {Future Generation Computer Systems}, volume = {27}, number = {4}, year = {2011}, month = {2011-01}, pages = {435-369}, author = {Emmanuel Agullo and Camille Coti and Thomas Herault and Julien Langou and Sylvain Peyronnet and A. Rezmerita and Franck Cappello and Jack Dongarra} } @article {1364, title = {Autotuning Dense Linear Algebra Libraries on GPUs}, year = {2010}, month = {2010-06}, publisher = {Sixth International Workshop on Parallel Matrix Algorithms and Applications (PMAA 2010)}, address = {Basel, Switzerland}, author = {Rajib Nath and Stanimire Tomov and Emmanuel Agullo and Jack Dongarra} } @techreport {icl:585, title = {Faster, Cheaper, Better - A Hybridization Methodology to Develop Linear Algebra Software for GPUs}, journal = {LAPACK Working Note}, number = {230}, year = {2010}, month = {2010-00}, keywords = {magma, morse}, author = {Emmanuel Agullo and Cedric Augonnet and Jack Dongarra and Hatem Ltaeif and Raymond Namyst and Samuel Thibault and Stanimire Tomov} } @article {icl:553, title = {MaPHyS or the Development of a Parallel Algebraic Domain Decomposition Solver in the Course of the Solstice Project}, journal = {Sparse Days 2010 Meeting at CERFACS}, year = {2010}, month = {2010-06}, address = {Toulouse, France}, author = {Emmanuel Agullo and Luc Giraud and Amina Guermouche and Azzam Haidar and Jean Roman and Yohan Lee-Tin-Yien} } @article {icl:574, title = {QCG-OMPI: MPI Applications on Grids}, journal = {Future Generation Computer Systems}, volume = {27}, number = {4}, year = {2010}, month = {2010-03}, pages = {357-369}, author = {Emmanuel Agullo and Camille Coti and Thomas Herault and Julien Langou and Sylvain Peyronnet and A. Rezmerita and Franck Cappello and Jack Dongarra} } @inproceedings {icl:532, title = {QR Factorization of Tall and Skinny Matrices in a Grid Computing Environment}, journal = {24th IEEE International Parallel and Distributed Processing Symposium (also LAWN 224)}, year = {2010}, month = {2010-04}, address = {Atlanta, GA}, author = {Emmanuel Agullo and Camille Coti and Jack Dongarra and Thomas Herault and Julien Langou} } @inproceedings {icl:577, title = {QR Factorization on a Multicore Node Enhanced with Multiple GPU Accelerators}, journal = {Proceedings of IPDPS 2011}, number = {ICL-UT-10-04}, year = {2010}, month = {2010-10}, address = {Anchorage, AK}, keywords = {magma, morse, plasma}, author = {Emmanuel Agullo and Cedric Augonnet and Jack Dongarra and Mathieu Faverge and Hatem Ltaeif and Samuel Thibault and Stanimire Tomov} } @article {1362, title = {Scheduling Cholesky Factorization on Multicore Architectures with GPU Accelerators}, year = {2010}, month = {2010-07}, publisher = {2010 Symposium on Application Accelerators in High-Performance Computing (SAAHPC{\textquoteright}10), Poster}, address = {Knoxville, TN}, author = {Emmanuel Agullo and Cedric Augonnet and Jack Dongarra and Hatem Ltaeif and Raymond Namyst and Rajib Nath and Jean Roman and Samuel Thibault and Stanimire Tomov} } @inproceedings {icl:588, title = {Tile QR Factorization with Parallel Panel Processing for Multicore Architectures}, journal = {24th IEEE International Parallel and Distributed Processing Symposium (submitted)}, year = {2010}, month = {2010-00}, author = {Bilel Hadri and Emmanuel Agullo and Jack Dongarra} } @article {icl:552, title = {Towards a Complexity Analysis of Sparse Hybrid Linear Solvers}, journal = {PARA 2010}, year = {2010}, month = {2010-06}, address = {Reykjavik, Iceland}, author = {Emmanuel Agullo and Luc Giraud and Amina Guermouche and Azzam Haidar and Jean Roman} } @inproceedings {icl:488, title = {Comparative Study of One-Sided Factorizations with Multiple Software Packages on Multi-Core Hardware}, journal = {2009 International Conference for High Performance Computing, Networking, Storage, and Analysis (SC {\textquoteright}09) (to appear)}, year = {2009}, month = {2009-00}, author = {Emmanuel Agullo and Bilel Hadri and Hatem Ltaeif and Jack Dongarra} } @article {icl:524, title = {Enhancing Parallelism of Tile QR Factorization for Multicore Architectures}, journal = {Submitted to Transaction on Parallel and Distributed Systems}, year = {2009}, month = {2009-12}, keywords = {plasma}, author = {Bilel Hadri and Hatem Ltaeif and Emmanuel Agullo and Jack Dongarra} } @article {1352, title = {Numerical Linear Algebra on Emerging Architectures: The PLASMA and MAGMA Projects}, year = {2009}, month = {2009-11}, publisher = {The International Conference for High Performance Computing, Networking, Storage, and Analysis (SC09)}, address = {Portland, OR}, author = {Emmanuel Agullo and James Demmel and Jack Dongarra and Bilel Hadri and Jakub Kurzak and Julien Langou and Hatem Ltaeif and Piotr Luszczek and Rajib Nath and Stanimire Tomov and Asim YarKhan and Vasily Volkov} } @inproceedings {icl:486, title = {Numerical Linear Algebra on Emerging Architectures: The PLASMA and MAGMA Projects}, journal = {Journal of Physics: Conference Series}, volume = {180}, year = {2009}, month = {2009-00}, keywords = {magma, plasma}, author = {Emmanuel Agullo and James Demmel and Jack Dongarra and Bilel Hadri and Jakub Kurzak and Julien Langou and Hatem Ltaeif and Piotr Luszczek and Stanimire Tomov} } @techreport {icl:487, title = {Tall and Skinny QR Matrix Factorization Using Tile Algorithms on Multicore Architectures}, journal = {Innovative Computing Laboratory Technical Report (also LAPACK Working Note 222 and CS Tech Report UT-CS-09-645)}, number = {ICL-UT-09-03}, year = {2009}, month = {2009-09}, keywords = {plasma}, author = {Bilel Hadri and Hatem Ltaeif and Emmanuel Agullo and Jack Dongarra} } @inproceedings {icl:522, title = {Tile QR Factorization with Parallel Panel Processing for Multicore Architectures}, journal = {accepted in 24th IEEE International Parallel and Distributed Processing Symposium (IPDPS 2010)}, year = {2009}, month = {2009-12}, address = {Atlanta, GA}, keywords = {plasma}, author = {Bilel Hadri and Hatem Ltaeif and Emmanuel Agullo and Jack Dongarra} }