Towards the Scalability of Dynamic Loop Scheduling Techniques via Discrete Event Simulation. Balasubramanian, M., Sukhija, N., Ciorba, F. M., Banicescu, I., & Srivastava, S. In Proceedings of the 26th IEEE International Parallel & Distributed Processing Symposium Workshops (IPDPSW), 21st International Heterogeneity in Computing Workshop (HCW12) 13th IEEE International Workshop on Parallel and Distributed Scientific and Engineering Computing (PDSEC12), pages 1343–1351, May, 2012. abstract bibtex To improve their performance, scientific applications often use loop scheduling algorithms as techniques for load balancing data parallel computations. Over the years, a number of dynamic loop scheduling (DLS) techniques have been developed. These techniques are based on probabilistic analyses, and are effective in addressing unpredictable load imbalances in the system arising from various sources, such as, variations in application, algorithmic, and systemic characteristics. Modern, high-end computing facilities can now offer petascale performance (10^15 flops), and several initiatives have already begun with the goal of achieving exascale performance (10^18 flops) towards the end of the current decade. Efficient and scalable algorithms are therefore required to utilize the petascale and exascale resources. In this paper, a study of the scalability of DLS techniques via discrete event simulation is presented, both in terms of number of processors, and problem size. To facilitate the scalability study, a dynamic loop scheduler was designed and was implemented using the SimGrid [1] simulation framework. The results of the study demonstrate the scalability of the DLS techniques and their effectiveness in addressing load imbalance in large scale computing systems.
@inproceedings{balasubramanian:2012:a,
Abstract = {To improve their performance, scientific applications often use loop scheduling algorithms as techniques for load balancing data parallel computations. Over the years, a number of dynamic loop scheduling (DLS) techniques have been developed. These techniques are based on probabilistic analyses, and are effective in addressing unpredictable load imbalances in the system arising from various sources, such as, variations in application, algorithmic, and systemic characteristics. Modern, high-end computing facilities can now offer petascale performance (10^15 flops), and several initiatives have already begun with the goal of achieving exascale performance (10^18 flops) towards the end of the current decade. Efficient and scalable algorithms are therefore required to utilize the petascale and exascale resources. In this paper, a study of the scalability of DLS techniques via discrete event simulation is presented, both in terms of number of processors, and problem size. To facilitate the scalability study, a dynamic loop scheduler was designed and was implemented using the SimGrid [1] simulation framework. The results of the study demonstrate the scalability of the DLS techniques and their effectiveness in addressing load imbalance in large scale computing systems.},
Author = {Balasubramanian, Mahadevan and Sukhija, Nitin and Ciorba, Florina M. and Banicescu, Ioana and Srivastava, Srishti},
Booktitle = {Proceedings of the 26th IEEE International Parallel & Distributed Processing Symposium Workshops (IPDPSW), 21st International Heterogeneity in Computing Workshop (HCW12) 13th IEEE International Workshop on Parallel and Distributed Scientific and Engineering Computing (PDSEC12)},
Date-Added = {2016-01-05 15:25:56 +0000},
Date-Modified = {2016-01-06 11:58:17 +0000},
Keywords = {2012; dynamic loop scheduling, scalability, simgrid},
Location = {Shanghai, China},
Month = may,
Pages = {1343--1351},
Title = {{Towards the Scalability of Dynamic Loop Scheduling Techniques via Discrete Event Simulation}},
Year = {2012}}
%% This BibTeX bibliography file was created using BibDesk.
%% http://bibdesk.sourceforge.net/
%% Created for Florina M. Ciorba at 2016-01-06 14:14:52 +0100
%% Saved with string encoding Unicode (UTF-8)
Downloads: 0
{"_id":"dJoxCLa29MwDgJ3G5","bibbaseid":"balasubramanian-sukhija-ciorba-banicescu-srivastava-towardsthescalabilityofdynamicloopschedulingtechniquesviadiscreteeventsimulation-2012","author_short":["Balasubramanian, M.","Sukhija, N.","Ciorba, F. M.","Banicescu, I.","Srivastava, S."],"bibdata":{"bibtype":"inproceedings","type":"inproceedings","abstract":"To improve their performance, scientific applications often use loop scheduling algorithms as techniques for load balancing data parallel computations. Over the years, a number of dynamic loop scheduling (DLS) techniques have been developed. These techniques are based on probabilistic analyses, and are effective in addressing unpredictable load imbalances in the system arising from various sources, such as, variations in application, algorithmic, and systemic characteristics. Modern, high-end computing facilities can now offer petascale performance (10^15 flops), and several initiatives have already begun with the goal of achieving exascale performance (10^18 flops) towards the end of the current decade. Efficient and scalable algorithms are therefore required to utilize the petascale and exascale resources. In this paper, a study of the scalability of DLS techniques via discrete event simulation is presented, both in terms of number of processors, and problem size. To facilitate the scalability study, a dynamic loop scheduler was designed and was implemented using the SimGrid [1] simulation framework. The results of the study demonstrate the scalability of the DLS techniques and their effectiveness in addressing load imbalance in large scale computing systems.","author":[{"propositions":[],"lastnames":["Balasubramanian"],"firstnames":["Mahadevan"],"suffixes":[]},{"propositions":[],"lastnames":["Sukhija"],"firstnames":["Nitin"],"suffixes":[]},{"propositions":[],"lastnames":["Ciorba"],"firstnames":["Florina","M."],"suffixes":[]},{"propositions":[],"lastnames":["Banicescu"],"firstnames":["Ioana"],"suffixes":[]},{"propositions":[],"lastnames":["Srivastava"],"firstnames":["Srishti"],"suffixes":[]}],"booktitle":"Proceedings of the 26th IEEE International Parallel & Distributed Processing Symposium Workshops (IPDPSW), 21st International Heterogeneity in Computing Workshop (HCW12) 13th IEEE International Workshop on Parallel and Distributed Scientific and Engineering Computing (PDSEC12)","date-added":"2016-01-05 15:25:56 +0000","date-modified":"2016-01-06 11:58:17 +0000","keywords":"2012; dynamic loop scheduling, scalability, simgrid","location":"Shanghai, China","month":"May","pages":"1343–1351","title":"Towards the Scalability of Dynamic Loop Scheduling Techniques via Discrete Event Simulation","year":"2012","bibtex":"@inproceedings{balasubramanian:2012:a,\n\tAbstract = {To improve their performance, scientific applications often use loop scheduling algorithms as techniques for load balancing data parallel computations. Over the years, a number of dynamic loop scheduling (DLS) techniques have been developed. These techniques are based on probabilistic analyses, and are effective in addressing unpredictable load imbalances in the system arising from various sources, such as, variations in application, algorithmic, and systemic characteristics. Modern, high-end computing facilities can now offer petascale performance (10^15 flops), and several initiatives have already begun with the goal of achieving exascale performance (10^18 flops) towards the end of the current decade. Efficient and scalable algorithms are therefore required to utilize the petascale and exascale resources. In this paper, a study of the scalability of DLS techniques via discrete event simulation is presented, both in terms of number of processors, and problem size. To facilitate the scalability study, a dynamic loop scheduler was designed and was implemented using the SimGrid [1] simulation framework. The results of the study demonstrate the scalability of the DLS techniques and their effectiveness in addressing load imbalance in large scale computing systems.},\n\tAuthor = {Balasubramanian, Mahadevan and Sukhija, Nitin and Ciorba, Florina M. and Banicescu, Ioana and Srivastava, Srishti},\n\tBooktitle = {Proceedings of the 26th IEEE International Parallel & Distributed Processing Symposium Workshops (IPDPSW), 21st International Heterogeneity in Computing Workshop (HCW12) 13th IEEE International Workshop on Parallel and Distributed Scientific and Engineering Computing (PDSEC12)},\n\tDate-Added = {2016-01-05 15:25:56 +0000},\n\tDate-Modified = {2016-01-06 11:58:17 +0000},\n\tKeywords = {2012; dynamic loop scheduling, scalability, simgrid},\n\tLocation = {Shanghai, China},\n\tMonth = may,\n\tPages = {1343--1351},\n\tTitle = {{Towards the Scalability of Dynamic Loop Scheduling Techniques via Discrete Event Simulation}},\n\tYear = {2012}}\n\n%% This BibTeX bibliography file was created using BibDesk.\n%% http://bibdesk.sourceforge.net/\n\n%% Created for Florina M. Ciorba at 2016-01-06 14:14:52 +0100 \n\n\n%% Saved with string encoding Unicode (UTF-8) \n\n\n\n","author_short":["Balasubramanian, M.","Sukhija, N.","Ciorba, F. M.","Banicescu, I.","Srivastava, S."],"bibbaseid":"balasubramanian-sukhija-ciorba-banicescu-srivastava-towardsthescalabilityofdynamicloopschedulingtechniquesviadiscreteeventsimulation-2012","role":"author","urls":{},"keyword":["2012; dynamic loop scheduling","scalability","simgrid"],"metadata":{"authorlinks":{}}},"bibtype":"inproceedings","biburl":"https://bibbase.org/f/5soafvAi5dxziNqBd/conferences.bib","dataSources":["yquS8QakKCWDYDAk9","aMnnnpEwMvMZwRFdC","we4PjHBkvY3ihPaQy","yezXfWaY6nTQG2nAv","pWKBSwppQZXZa3qnh"],"keywords":["2012; dynamic loop scheduling","scalability","simgrid"],"search_terms":["towards","scalability","dynamic","loop","scheduling","techniques","via","discrete","event","simulation","balasubramanian","sukhija","ciorba","banicescu","srivastava"],"title":"Towards the Scalability of Dynamic Loop Scheduling Techniques via Discrete Event Simulation","year":2012}