FLEX-MPI: an MPI extension for supporting dynamic load balancing on heterogeneous non-dedicated systems

Authors:
Gonzalo Martín;Maria-Cristina Marinescu;David E. Singh;Jesús Carretero
Affiliations:
Universidad Carlos III de Madrid, Leganés, Spain;Barcelona Supercomputing Center, Barcelona, Spain;Universidad Carlos III de Madrid, Leganés, Spain;Universidad Carlos III de Madrid, Leganés, Spain
Venue:
Euro-Par'13 Proceedings of the 19th international conference on Parallel Processing
Year:
2013

Citing 8
Cited 0

The NAS parallel benchmarks—summary and preliminary results

Proceedings of the 1991 ACM/IEEE conference on Supercomputing
Load Balancing in Parallel Computers: Theory and Practice

Load Balancing in Parallel Computers: Theory and Practice
HeteroMPI: Towards a message-passing library for heterogeneous networks of computers

Journal of Parallel and Distributed Computing
Dyn-MPI: Supporting MPI on medium-scale, non-dedicated clusters

Journal of Parallel and Distributed Computing
Dealing with Heterogeneity in Load Balancing Algorithms

ISPDC '06 Proceedings of the Proceedings of The Fifth International Symposium on Parallel and Distributed Computing
Dynamic Load Balancing on Dedicated Heterogeneous Systems

Proceedings of the 15th European PVM/MPI Users' Group Meeting on Recent Advances in Parallel Virtual Machine and Message Passing Interface
Performance instrumentation and measurement for terascale systems

ICCS'03 Proceedings of the 2003 international conference on Computational science
The university of Florida sparse matrix collection

ACM Transactions on Mathematical Software (TOMS)

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper introduces FLEX-MPI, a novel runtime approach for the dynamic load balancing of MPI-based SPMD applications running on heterogeneous platforms in the presence of dynamic external loads. To effectively balance the workload, FLEX-MPI monitors the actual performance of applications via hardware counters and the MPI profiling interface--with a negligible overhead and minimal code modifications. Our results show that by using this approach the execution time of an application may be significantly reduced.