Task-parallel analysis of molecular dynamics trajectories. Paraskevakos, I., Chantzialexiou, G., Luckow, A., Cheatham, T., E., Khoshlessan, M., Beckstein, O., Fox, G., C., & Jha, S. In Proceedings of the 47th International Conference on Parallel Processing (ICPP 2018), 8, 2018. Association for Computing Machinery.
![Task-parallel analysis of molecular dynamics trajectories [pdf]](https://bibbase.org/img/filetypes/pdf.svg "pdf")
Paper doi abstract bibtex
Paper doi abstract bibtex Different parallel frameworks for implementing data analysis applications have been proposed by the HPC and Big Data communities. In this paper, we investigate three task-parallel frameworks: Spark, Dask and RADICAL-Pilot with respect to their ability to support data analytics on HPC resources and compare them with MPI. We investigate the data analysis requirements of Molecular Dynamics (MD) simulations which are significant consumers of supercomputing cycles, producing immense amounts of data. A typical large-scale MD simulation of a physical system of O(100k) atoms over \musecs can produce from O(10) GB to O(1000) GBs of data. We propose and evaluate different approaches for parallelization of a representative set of MD trajectory analysis algorithms, in particular the computation of path similarity and leaflet identification. We evaluate Spark, Dask and RADICAL-Pilot with respect to their abstractions and runtime engine capabilities to support these algorithms. We provide a conceptual basis for comparing and understanding different frameworks that enable users to select the optimal system for each application. We also provide a quantitative performance analysis of the different algorithms across the three frameworks.
@inproceedings{
title = {Task-parallel analysis of molecular dynamics trajectories},
type = {inproceedings},
year = {2018},
keywords = {Data analytics,MD analysis,MD simulations analysis,Task-parallel},
month = {8},
publisher = {Association for Computing Machinery},
day = {13},
id = {ec60cc2e-ca18-3e16-8fac-313503fbbe74},
created = {2019-10-01T17:21:02.886Z},
accessed = {2019-09-03},
file_attached = {true},
profile_id = {42d295c0-0737-38d6-8b43-508cab6ea85d},
last_modified = {2020-05-11T14:43:32.324Z},
read = {false},
starred = {false},
authored = {true},
confirmed = {false},
hidden = {false},
citation_key = {Paraskevakos2018},
private_publication = {false},
abstract = {Different parallel frameworks for implementing data analysis applications have been proposed by the HPC and Big Data communities. In this paper, we investigate three task-parallel frameworks: Spark, Dask and RADICAL-Pilot with respect to their ability to support data analytics on HPC resources and compare them with MPI. We investigate the data analysis requirements of Molecular Dynamics (MD) simulations which are significant consumers of supercomputing cycles, producing immense amounts of data. A typical large-scale MD simulation of a physical system of O(100k) atoms over \musecs can produce from O(10) GB to O(1000) GBs of data. We propose and evaluate different approaches for parallelization of a representative set of MD trajectory analysis algorithms, in particular the computation of path similarity and leaflet identification. We evaluate Spark, Dask and RADICAL-Pilot with respect to their abstractions and runtime engine capabilities to support these algorithms. We provide a conceptual basis for comparing and understanding different frameworks that enable users to select the optimal system for each application. We also provide a quantitative performance analysis of the different algorithms across the three frameworks.},
bibtype = {inproceedings},
author = {Paraskevakos, Ioannis and Chantzialexiou, George and Luckow, Andre and Cheatham, Thomas E. and Khoshlessan, Mahzad and Beckstein, Oliver and Fox, Geoffrey C. and Jha, Shantenu},
doi = {10.1145/3225058.3225128},
booktitle = {Proceedings of the 47th International Conference on Parallel Processing (ICPP 2018)}
}Downloads: 0
{"_id":"KARpYe67MpTQPByCT","bibbaseid":"paraskevakos-chantzialexiou-luckow-cheatham-khoshlessan-beckstein-fox-jha-taskparallelanalysisofmoleculardynamicstrajectories-2018","authorIDs":[],"author_short":["Paraskevakos, I.","Chantzialexiou, G.","Luckow, A.","Cheatham, T., E.","Khoshlessan, M.","Beckstein, O.","Fox, G., C.","Jha, S."],"bibdata":{"title":"Task-parallel analysis of molecular dynamics trajectories","type":"inproceedings","year":"2018","keywords":"Data analytics,MD analysis,MD simulations analysis,Task-parallel","month":"8","publisher":"Association for Computing Machinery","day":"13","id":"ec60cc2e-ca18-3e16-8fac-313503fbbe74","created":"2019-10-01T17:21:02.886Z","accessed":"2019-09-03","file_attached":"true","profile_id":"42d295c0-0737-38d6-8b43-508cab6ea85d","last_modified":"2020-05-11T14:43:32.324Z","read":false,"starred":false,"authored":"true","confirmed":false,"hidden":false,"citation_key":"Paraskevakos2018","private_publication":false,"abstract":"Different parallel frameworks for implementing data analysis applications have been proposed by the HPC and Big Data communities. In this paper, we investigate three task-parallel frameworks: Spark, Dask and RADICAL-Pilot with respect to their ability to support data analytics on HPC resources and compare them with MPI. We investigate the data analysis requirements of Molecular Dynamics (MD) simulations which are significant consumers of supercomputing cycles, producing immense amounts of data. A typical large-scale MD simulation of a physical system of O(100k) atoms over \\musecs can produce from O(10) GB to O(1000) GBs of data. We propose and evaluate different approaches for parallelization of a representative set of MD trajectory analysis algorithms, in particular the computation of path similarity and leaflet identification. We evaluate Spark, Dask and RADICAL-Pilot with respect to their abstractions and runtime engine capabilities to support these algorithms. We provide a conceptual basis for comparing and understanding different frameworks that enable users to select the optimal system for each application. We also provide a quantitative performance analysis of the different algorithms across the three frameworks.","bibtype":"inproceedings","author":"Paraskevakos, Ioannis and Chantzialexiou, George and Luckow, Andre and Cheatham, Thomas E. and Khoshlessan, Mahzad and Beckstein, Oliver and Fox, Geoffrey C. and Jha, Shantenu","doi":"10.1145/3225058.3225128","booktitle":"Proceedings of the 47th International Conference on Parallel Processing (ICPP 2018)","bibtex":"@inproceedings{\n title = {Task-parallel analysis of molecular dynamics trajectories},\n type = {inproceedings},\n year = {2018},\n keywords = {Data analytics,MD analysis,MD simulations analysis,Task-parallel},\n month = {8},\n publisher = {Association for Computing Machinery},\n day = {13},\n id = {ec60cc2e-ca18-3e16-8fac-313503fbbe74},\n created = {2019-10-01T17:21:02.886Z},\n accessed = {2019-09-03},\n file_attached = {true},\n profile_id = {42d295c0-0737-38d6-8b43-508cab6ea85d},\n last_modified = {2020-05-11T14:43:32.324Z},\n read = {false},\n starred = {false},\n authored = {true},\n confirmed = {false},\n hidden = {false},\n citation_key = {Paraskevakos2018},\n private_publication = {false},\n abstract = {Different parallel frameworks for implementing data analysis applications have been proposed by the HPC and Big Data communities. In this paper, we investigate three task-parallel frameworks: Spark, Dask and RADICAL-Pilot with respect to their ability to support data analytics on HPC resources and compare them with MPI. We investigate the data analysis requirements of Molecular Dynamics (MD) simulations which are significant consumers of supercomputing cycles, producing immense amounts of data. A typical large-scale MD simulation of a physical system of O(100k) atoms over \\musecs can produce from O(10) GB to O(1000) GBs of data. We propose and evaluate different approaches for parallelization of a representative set of MD trajectory analysis algorithms, in particular the computation of path similarity and leaflet identification. We evaluate Spark, Dask and RADICAL-Pilot with respect to their abstractions and runtime engine capabilities to support these algorithms. We provide a conceptual basis for comparing and understanding different frameworks that enable users to select the optimal system for each application. We also provide a quantitative performance analysis of the different algorithms across the three frameworks.},\n bibtype = {inproceedings},\n author = {Paraskevakos, Ioannis and Chantzialexiou, George and Luckow, Andre and Cheatham, Thomas E. and Khoshlessan, Mahzad and Beckstein, Oliver and Fox, Geoffrey C. and Jha, Shantenu},\n doi = {10.1145/3225058.3225128},\n booktitle = {Proceedings of the 47th International Conference on Parallel Processing (ICPP 2018)}\n}","author_short":["Paraskevakos, I.","Chantzialexiou, G.","Luckow, A.","Cheatham, T., E.","Khoshlessan, M.","Beckstein, O.","Fox, G., C.","Jha, S."],"urls":{"Paper":"https://bibbase.org/service/mendeley/42d295c0-0737-38d6-8b43-508cab6ea85d/file/2eca2404-2b6d-2ee9-1849-af87f3575e8a/Paraskevakos_et_al___2018___Task_parallel_analysis_of_molecular_dynamics_trajectories.pdf.pdf"},"biburl":"https://bibbase.org/service/mendeley/42d295c0-0737-38d6-8b43-508cab6ea85d","bibbaseid":"paraskevakos-chantzialexiou-luckow-cheatham-khoshlessan-beckstein-fox-jha-taskparallelanalysisofmoleculardynamicstrajectories-2018","role":"author","keyword":["Data analytics","MD analysis","MD simulations analysis","Task-parallel"],"metadata":{"authorlinks":{}},"downloads":0},"bibtype":"inproceedings","creationDate":"2019-09-12T13:19:08.415Z","downloads":0,"keywords":["data analytics","md analysis","md simulations analysis","task-parallel"],"search_terms":["task","parallel","analysis","molecular","dynamics","trajectories","paraskevakos","chantzialexiou","luckow","cheatham","khoshlessan","beckstein","fox","jha"],"title":"Task-parallel analysis of molecular dynamics trajectories","year":2018,"biburl":"https://bibbase.org/service/mendeley/42d295c0-0737-38d6-8b43-508cab6ea85d","dataSources":["zgahneP4uAjKbudrQ","ya2CyA73rpZseyrZ8","2252seNhipfTmjEBQ"]}