Scheduling multi-task applications on heterogeneous platforms Anne - PowerPoint PPT Presentation

Framework Theoretical study Experiments Conclusion Scheduling multi-task applications on heterogeneous platforms Anne Benoit, Jean-Fran¸ cois Pineau, Yves Robert and Fr´ ed´ eric Vivien Laboratoire de l’Informatique du Parall´ elisme ´ Ecole Normale Sup´ erieure de Lyon, France Jean-Francois.Pineau@ens-lyon.fr http://graal.ens-lyon.fr/ ∼ jfpineau GDT GRAAL July 5, 2007 1/33 Jean-Fran¸ cois Pineau Bag of tasks

Framework Theoretical study Experiments Conclusion Outline Framework 1 Theoretical study 2 Steady state scheduling Off-line study Extension Experiments 3 Conclusion 4 2/33 Jean-Fran¸ cois Pineau Bag of tasks

Framework Theoretical study Experiments Conclusion Bag-of-tasks Applications Bag of tasks described by: the number of tasks the amount of computation of a task the amount of communication of a task their release date On-line scheduling. 3/33 Jean-Fran¸ cois Pineau Bag of tasks

Framework Theoretical study Experiments Conclusion Bag-of-tasks Applications Bag of tasks described by: the number of independent tasks the amount of computation of a task the amount of communication of a task their release date On-line scheduling. 3/33 Jean-Fran¸ cois Pineau Bag of tasks

Framework Theoretical study Experiments Conclusion Bag-of-tasks Applications Bag of tasks described by: the number of independent, identical tasks the amount of computation of a task the amount of communication of a task their release date On-line scheduling. 3/33 Jean-Fran¸ cois Pineau Bag of tasks

Framework Theoretical study Experiments Conclusion Bag-of-tasks Applications Bag of tasks described by: the number of independent, identical tasks the amount of computation of a task the amount of communication of a task their release date On-line scheduling. 3/33 Jean-Fran¸ cois Pineau Bag of tasks

Framework Theoretical study Experiments Conclusion Platform model Tasks Master Network Links � � � � � � � � � � � � � � � � ����� ����� ����� ����� � � � � � � � � ����� ����� ����� ����� � � � � � � � � � � � � � � � � ����� ����� ����� ����� � � � � � � � � ����� ����� ����� ����� � � � � � � � � ����� ����� � � � � ����� ����� � � � � ����� ����� � � � � ����� ����� � � � � ����� ����� � � � � ����� ����� � � � � � � � � � � � � � � � � � � � � � � � � � � � � ����� ����� � � � � ����� ����� � � � � ����� ����� � � � � ����� ����� � � � � ����� ����� � � � � ����� ����� � � � � ����� ����� � � � � ����� ����� � � � � ����� ����� � � � � ����� ����� � � � � ����� ����� � � � � ����� ����� � � � � ����� ����� � � � � ����� ����� � � � � ����� ����� ����� ����� � � � � � � � � Slaves 4/33 Jean-Fran¸ cois Pineau Bag of tasks

Framework Theoretical study Experiments Conclusion Master-slaves platform The master Receive the bags of tasks Send the tasks to the processors Bounded multi-port model The processors Parallels Identical Uniform Related 5/33 Jean-Fran¸ cois Pineau Bag of tasks

Framework Theoretical study Experiments Conclusion Master-slaves platform The master Receive the bags of tasks Send the tasks to the processors Bounded multi-port model The processors Parallels Identical Uniform Related 5/33 Jean-Fran¸ cois Pineau Bag of tasks

Framework Theoretical study Experiments Conclusion Notations Tasks n bags-of-tasks applications A k A i is composed of Π ( i ) tasks. w ( i ) : amount of computation of a task of A i δ ( i ) : amount of communication of a task of A i r ( i ) : release date of A i C ( i ) : completion time of A i 6/33 Jean-Fran¸ cois Pineau Bag of tasks

Framework Theoretical study Experiments Conclusion Notations Platform p processors, B : bound of the multi-port model. b u : bandwidth of the link between the master and P u , s u : computational speed of worker P u , 6/33 Jean-Fran¸ cois Pineau Bag of tasks

Framework Theoretical study Experiments Conclusion Notations Platform p processors, B : bound of the multi-port model. b u : bandwidth of the link between the master and P u , s ( k ) u : computational speed of related worker P u with tasks of A k , 6/33 Jean-Fran¸ cois Pineau Bag of tasks

Framework Theoretical study Experiments Conclusion Objective Scheduling the tasks to the processors in order to process this tasks according to the constraints, of the processors of the tasks optimizing an objective function 7/33 Jean-Fran¸ cois Pineau Bag of tasks

Framework Theoretical study Experiments Conclusion Objective function Objective function Makespan max C ( i ) or C ( max ) 8/33 Jean-Fran¸ cois Pineau Bag of tasks

Framework Theoretical study Experiments Conclusion Objective function Objective function Makespan max C ( i ) or C ( max ) Problem of satisfaction of the clients 8/33 Jean-Fran¸ cois Pineau Bag of tasks

Framework Theoretical study Experiments Conclusion Objective function Objective function Makespan Sum flow {C ( i ) − r ( i ) } � 8/33 Jean-Fran¸ cois Pineau Bag of tasks

Framework Theoretical study Experiments Conclusion Objective function Objective function Makespan Sum flow {C ( i ) − r ( i ) } � Problem of starvation 8/33 Jean-Fran¸ cois Pineau Bag of tasks

Framework Theoretical study Experiments Conclusion Objective function Objective function Makespan Sum flow Max flow max {C ( i ) − r ( i ) } 8/33 Jean-Fran¸ cois Pineau Bag of tasks

Framework Theoretical study Experiments Conclusion Objective function Objective function Makespan Sum flow Max flow max {C ( i ) − r ( i ) } Small applications can wait a long time 8/33 Jean-Fran¸ cois Pineau Bag of tasks

Framework Theoretical study Experiments Conclusion Objective function Objective function Makespan Sum flow Max flow Max Stretch max C ( i ) − r ( i ) Size of A i 8/33 Jean-Fran¸ cois Pineau Bag of tasks

Framework Theoretical study Experiments Conclusion Objective function Objective function Makespan Sum flow Max flow Max Stretch max C ( i ) − r ( i ) Size of A i Size of A i = Π ( i ) ? 8/33 Jean-Fran¸ cois Pineau Bag of tasks

Framework Theoretical study Experiments Conclusion Objective function Objective function Makespan Sum flow Max flow Max Stretch max C ( i ) − r ( i ) Size of A i Size of A i = w ( i ) ? 8/33 Jean-Fran¸ cois Pineau Bag of tasks

Framework Theoretical study Experiments Conclusion Objective function Objective function Makespan Sum flow Max flow Max Stretch max C ( i ) − r ( i ) Size of A i Size of A i = Π ( i ) ∗ w ( i ) ? 8/33 Jean-Fran¸ cois Pineau Bag of tasks

Framework Steady state scheduling Theoretical study Off-line study Experiments Extension Conclusion Outline Framework 1 Theoretical study 2 Steady state scheduling Off-line study Extension Experiments 3 Conclusion 4 9/33 Jean-Fran¸ cois Pineau Bag of tasks

Framework Steady state scheduling Theoretical study Off-line study Experiments Extension Conclusion Simple problem Problem Unique bag-of-tasks A 0 Large Π (0) 10/33 Jean-Fran¸ cois Pineau Bag of tasks

Framework Steady state scheduling Theoretical study Off-line study Experiments Extension Conclusion Simple problem Problem Unique bag-of-tasks A 0 Large Π (0) Objective Minimizing the makespan 10/33 Jean-Fran¸ cois Pineau Bag of tasks

Framework Steady state scheduling Theoretical study Off-line study Experiments Extension Conclusion Simple problem Problem Unique bag-of-tasks A 0 Large Π (0) Objective Minimizing the makespan Maximizing the throughput 10/33 Jean-Fran¸ cois Pineau Bag of tasks

Framework Steady state scheduling Theoretical study Off-line study Experiments Extension Conclusion Simple problem Problem Unique bag-of-tasks A 0 Large Π (0) Objective Minimizing the makespan Maximizing the throughput Throughput of worker P u : ρ ∗ (0) u p Total throughput ρ ∗ (0) = ρ ∗ (0) � u u =1 10/33 Jean-Fran¸ cois Pineau Bag of tasks