{"id":100,"date":"2020-06-01T19:09:20","date_gmt":"2020-06-01T19:09:20","guid":{"rendered":"http:\/\/mlhpc2.peter-zaspel.de\/?page_id=100"},"modified":"2020-12-28T12:48:09","modified_gmt":"2020-12-28T12:48:09","slug":"high-performance-computing","status":"publish","type":"page","link":"https:\/\/www.peter-zaspel.de\/?page_id=100","title":{"rendered":"High Performance Computing"},"content":{"rendered":"\n
\n
\n
\"\"
Weak scaling results for a fluid solver.<\/figcaption><\/figure><\/div>\n<\/div>\n\n\n\n
\n
\"\"
Performance per watt measurements in a parallelization.<\/figcaption><\/figure><\/div>\n<\/div>\n<\/div>\n\n\n\n

In the area of High Performance Computing, a parallel fluid dynamics solver for multi-GPU hardware has been ported to run on up to 36 GPUs with over 70% parallel weak scaling efficiency [Zas15; GZ10; ZG11; ZG13]. Furthermore, multi-GPU parallel uncertainty quantification methods [Zas15; GRZ19], a GPU parallel algebraic multigrid solver [Zas15; Zas], and GPU and multi-GPU parallel matrix approximation methods running on up to 1024 GPUs of the Titan cluster (the former top 1 system) at Oak Ridge National Lab [Zas15; Zas19a; HZ19a] have been developed. Most of the algorithms in machine learning and big data of this group have been also implemented in a (distributed-memory) GPU-\/CPU-parallel fashion [Zas16; Zas19b; HZ19b].<\/p>\n\n\n\n

\n\n\n\n

Related work<\/h4>\n\n\n\n
  • [GRZ19] M. Griebel, C. Rieger, and Peter Zaspel. \u201cKernel-based stochastics collocation for the random two-phase Navier-Stokes equations\u201d. In: International Journal for Uncertainty Quantification, 9(5), 2019, pp. 471\u2013492.<\/li>
  • [HZ19a] H. Harbrecht and P. Zaspel. A scalable H-matrix approach for the solution of boundary integral equations on multi-GPU clusters. Submitted to Computers & Mathematics with Applications, February 2019; available as arXiv Preprint. 2019.<\/li>
  • [HZ19b] H. Harbrecht and P. Zaspel. \u201cOn the Algebraic Construction of Sparse Multilevel Approximations of Elliptic Tensor Product Problems\u201d. In: Journal of Scientific Computing, 78(2), 2019, pp. 1272\u20131290.<\/li>
  • [Zas19a] P. Zaspel. \u201cAlgorithmic Patterns for H-Matrices on Many-Core Processors\u201d. In: Journal of Scientific Computing, 78(2), 2019, pp. 1174\u20131206.<\/li>
  • [Zas19b] P. Zaspel. \u201cEnsemble Kalman filters for reliability estimation in perfusion inference\u201d. In: International Journal for Uncertainty Quantification, 9(1), 2019, pp. 15\u201332.<\/li>
  • [Zas16] P. Zaspel. \u201cSubspace correction methods in algebraic multi-level frames\u201d. In: Linear Algebra and its Applications, 488, 2016, pp. 505\u2013521.<\/li>
  • [Zas15] P. Zaspel. \u201cParallel RBF Kernel-Based Stochastic Collocation for Large-Scale Random PDEs\u201d. Dissertation. Institut f\u00fcr Numerische Simulation, Universit\u00e4t Bonn, 2015.<\/li>
  • [Pfl+14] D. Pfl\u00fcger et al. \u201cEXAHD: An Exa-scalable Two-Level Sparse Grid Approach for Higher-Dimensional Problems in Plasma Physics and Beyond\u201d. In: Euro-Par 2014: Parallel Processing Workshops. Vol. 8806. Lecture Notes in Computer Science. Springer International Publishing, 2014, pp. 565\u2013576.<\/li>
  • [ZG13] P. Zaspel and M. Griebel. \u201cSolving incompressible two-phase flows on multi-GPU clusters\u201d. In: Computers & Fluids, 80(0), 2013, pp. 356\u2013364.<\/li>
  • [ZG11] P. Zaspel and M. Griebel. \u201cMassively Parallel Fluid Simulations on Amazon\u2019s HPC Cloud\u201d. In: First International Symposium on Network Cloud Computing and Applications (NCCA), 2011. 2011, pp. 73\u201378.<\/li>
  • [GZ10] M. Griebel and P. Zaspel. \u201cA multi-GPU accelerated solver for the three-dimensional two-phase incompressible Navier-Stokes equations\u201d. In: Computer Science – Research and Development, 25(1\u20132), 2010, pp. 65\u201373.<\/li>
  • [Zas] P. Zaspel. Analysis and parallelization strategies for Ruge-St\u00fcben AMG on many-core processors. Preprint series of the Department of Mathematics and Computer Science, University of Basel, June 2017.<\/li><\/ul>\n","protected":false},"excerpt":{"rendered":"

    In the area of High Performance Computing, a parallel fluid dynamics solver for multi-GPU hardware has been ported to run on up to 36 GPUs with over 70% parallel weak scaling efficiency [Zas15; GZ10; ZG11; ZG13]. Furthermore, multi-GPU parallel uncertainty<\/p>\n","protected":false},"author":1,"featured_media":0,"parent":0,"menu_order":0,"comment_status":"closed","ping_status":"closed","template":"","meta":{"_import_markdown_pro_load_document_selector":0,"_import_markdown_pro_submit_text_textarea":"","_mc_calendar":[],"footnotes":""},"class_list":["post-100","page","type-page","status-publish","hentry"],"_links":{"self":[{"href":"https:\/\/www.peter-zaspel.de\/index.php?rest_route=\/wp\/v2\/pages\/100","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.peter-zaspel.de\/index.php?rest_route=\/wp\/v2\/pages"}],"about":[{"href":"https:\/\/www.peter-zaspel.de\/index.php?rest_route=\/wp\/v2\/types\/page"}],"author":[{"embeddable":true,"href":"https:\/\/www.peter-zaspel.de\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.peter-zaspel.de\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=100"}],"version-history":[{"count":9,"href":"https:\/\/www.peter-zaspel.de\/index.php?rest_route=\/wp\/v2\/pages\/100\/revisions"}],"predecessor-version":[{"id":265,"href":"https:\/\/www.peter-zaspel.de\/index.php?rest_route=\/wp\/v2\/pages\/100\/revisions\/265"}],"wp:attachment":[{"href":"https:\/\/www.peter-zaspel.de\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=100"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}