Paper Info
Title
Dynamic load balancing of MPI+OpenMP applications
Abstract
The hybrid programming model MPI+OpenMP are useful to solve the problems of load balancing of parallel applications independently of the architecture. Typical approaches to balance parallel applications using two levels of parallelism or only MPI consist of including complex codes that dynamically detect which data domains are more computational intensive and either manually redistribute the allocated processors or manually redistribute data. This approach has two drawbacks: it is time consuming and it requires an expert in application analysis. In this paper we present an automatic and dynamic approach for load balancing MPI+OpenMP applications. The system will calculate the percentage of load imbalance and will decide a processor distribution for the MPI processes that eliminates the computational load imbalance. Results show that this method can balance effectively applications without analyzing nor modifying them and that in the cases that the application was well balanced does not incur in a great overhead for the dynamic instrumentation and analysis realized.
Year
DOI
Venue
2004
10.1109/ICPP.2004.1327921
ICPP
Keywords
Field
DocType
application program interfaces,message passing,open systems,resource allocation,shared memory systems,MPI processes,OpenMP,autonomic computing,computational load imbalance,hybrid programming model,load balancing,parallel application,processor distribution,resource management
Resource management,Autonomic computing,Data domain,Load balancing (computing),Computer science,Parallel computing,Resource allocation,Open system (systems theory),Dynamic load balancing,Message passing,Distributed computing
Conference
ISSN
ISBN
Citations 
0190-3918
0-7695-2197-5
24
PageRank 
References 
Authors
1.42
7
3
Name
Order
Citations
PageRank
Julita Corbalán133923.32
Alejandro Duran294361.43
Jesús Labarta31862165.09