Title :
Non Linear Divisible Loads: There is No Free Lunch
Author :
Beaumont, Olivier ; Larcheveque, Hubert ; Marchal, Loris
Author_Institution :
LaBRI, Univ. of Bordeaux, Bordeaux, France
Abstract :
Divisible Load Theory (DLT) has received a lot of attention in the past decade. A divisible load is a perfect parallel task, that can be split arbitrarily and executed in parallel on a set of possibly heterogeneous resources. The success of DLT is strongly related to the existence of many optimal resource allocation and scheduling algorithms, what strongly differs from general scheduling theory. Moreover, recently, close relationships have been underlined between DLT, that provides a fruitful theoretical framework for scheduling jobs on heterogeneous platforms, and MapReduce, that provides a simple and efficient programming framework to deploy applications on large scale distributed platforms. The success of both have suggested to extend their framework to non-linear complexity tasks. In this paper, we show that both DLT and MapReduce are better suited to workloads with linear complexity. In particular, we prove that divisible load theory cannot directly be applied to quadratic workloads, such as it has been proposed recently. We precisely state the limits for classical DLT studies and we review and propose solutions based on a careful preparation of the dataset and clever data partitioning algorithms. In particular, through simulations, we show the possible impact of this approach on the volume of communications generated by MapReduce, in the context of Matrix Multiplication and Outer Product algorithms.
Keywords :
matrix multiplication; parallel programming; processor scheduling; resource allocation; DLT; MapReduce; clever data partitioning algorithms; divisible load theory; general scheduling theory; heterogeneous platforms; heterogeneous resources; large scale distributed platforms; matrix multiplication; nonlinear complexity tasks; nonlinear divisible loads; optimal resource allocation; outer product algorithms; parallel task; programming framework; quadratic workloads; scheduling algorithms; Complexity theory; Computational modeling; Load modeling; Processor scheduling; Resource management; Sorting; Streaming media; Divisible Load Theory; MapReduce; Matrix Multiplication; Resource Allocation; Scheduling; Sorting;
Conference_Titel :
Parallel & Distributed Processing (IPDPS), 2013 IEEE 27th International Symposium on
Conference_Location :
Boston, MA
Print_ISBN :
978-1-4673-6066-1
DOI :
10.1109/IPDPS.2013.94