%0 Book Section %1 Eckhardt2013 %A Eckhardt, Wolfgang %A Heinecke, Alexander %A Bader, Reinhold %A Brehm, Matthias %A Hammer, Nicolay %A Huber, Herbert %A Kleinhenz, Hans-Georg %A Vrabec, Jadran %A Hasse, Hans %A Horsch, Martin %A Bernreuther, Martin %A Glass, Colin W. %A Niethammer, Christoph %A Bode, Arndt %A Bungartz, Hans-Joachim %B Supercomputing: 28th International Supercomputing Conference, ISC 2013, Leipzig, Germany, June 16-20, 2013. Proceedings %C Berlin, Heidelberg %D 2013 %E Kunkel, Julian Martin %E Ludwig, Thomas %E Meuer, Hans Werner %I Springer Berlin Heidelberg %K dynamics highly molecular parallel programming scalable simulation simulations vectorization %P 1--12 %R 10.1007/978-3-642-38750-0_1 %T 591 TFLOPS Multi-trillion Particles Simulation on SuperMUC %U https://doi.org/10.1007/978-3-642-38750-0_1 %X Anticipating large-scale molecular dynamics simulations (MD) in nano-fluidics, we conduct performance and scalability studies of an optimized version of the code ls1 mardyn. We present our implementation requiring only 32 Bytes per molecule, which allows us to run the, to our knowledge, largest MD simulation to date. Our optimizations tailored to the Intel Sandy Bridge processor are explained, including vectorization as well as shared-memory parallelization to make use of Hyperthreading. Finally we present results for weak and strong scaling experiments on up to 146016 Cores of SuperMUC at the Leibniz Supercomputing Centre, achieving a speed-up of 133k times which corresponds to an absolute performance of 591.2 TFLOPS. %@ 978-3-642-38750-0

@inbook{Eckhardt2013, abstract = {Anticipating large-scale molecular dynamics simulations (MD) in nano-fluidics, we conduct performance and scalability studies of an optimized version of the code ls1 mardyn. We present our implementation requiring only 32 Bytes per molecule, which allows us to run the, to our knowledge, largest MD simulation to date. Our optimizations tailored to the Intel Sandy Bridge processor are explained, including vectorization as well as shared-memory parallelization to make use of Hyperthreading. Finally we present results for weak and strong scaling experiments on up to 146016 Cores of SuperMUC at the Leibniz Supercomputing Centre, achieving a speed-up of 133k times which corresponds to an absolute performance of 591.2 TFLOPS.}, added-at = {2017-12-12T09:58:14.000+0100}, address = {Berlin, Heidelberg}, author = {Eckhardt, Wolfgang and Heinecke, Alexander and Bader, Reinhold and Brehm, Matthias and Hammer, Nicolay and Huber, Herbert and Kleinhenz, Hans-Georg and Vrabec, Jadran and Hasse, Hans and Horsch, Martin and Bernreuther, Martin and Glass, Colin W. and Niethammer, Christoph and Bode, Arndt and Bungartz, Hans-Joachim}, biburl = {https://puma.ub.uni-stuttgart.de/bibtex/2dffb2a37e29a0efe71208a49226516a6/c.niethammer}, booktitle = {Supercomputing: 28th International Supercomputing Conference, ISC 2013, Leipzig, Germany, June 16-20, 2013. Proceedings}, doi = {10.1007/978-3-642-38750-0_1}, editor = {Kunkel, Julian Martin and Ludwig, Thomas and Meuer, Hans Werner}, interhash = {a0dbaf1fbdff3c8d5401f74e8d6a7aad}, intrahash = {dffb2a37e29a0efe71208a49226516a6}, isbn = {978-3-642-38750-0}, keywords = {dynamics highly molecular parallel programming scalable simulation simulations vectorization}, pages = {1--12}, publisher = {Springer Berlin Heidelberg}, timestamp = {2017-12-12T08:58:14.000+0100}, title = {591 TFLOPS Multi-trillion Particles Simulation on SuperMUC}, url = {https://doi.org/10.1007/978-3-642-38750-0_1}, year = 2013 }