Escalonamento Tolerante a Falhas para Clusters Multicores
Resumo
Um dos principais desafios da computação de alto desempenho é executar aplicações paralelas de grande porte com maior desempenho e livre de falhas. Entretanto, para obter melhores resultados é primordial o aproveitamento eficiente dos recursos disponíveis, explorando por exemplo o uso das diferentes memórias compartilhadas e distribuídas em novas arquiteturas multicores. Este trabalho propõe uma abordagem híbrida para o escalonamento tolerante a falhas sobre clusters baseados em processadores multicores. Um estudo de caso é proposto para uma aplicação paralela modelada por um Grafo Acíclico Direcionado (GAD) usando programação paralela híbrida OpenMP e MPI. O estudo proposto deve analisar as vantagens que este modelo pode trazer para estas arquiteturas, comparado com a abordagem anterior.
Referências
(May, 2008). OpenMP application program interface version 3.0 complete specifications. http://www.openmp.org/mp-documents/specs30.pdf.
(Nov, 2009). MPI: A message-passing interface standard version 2.1. https://www.mpi-forum.org/docs/mpi21-report.pdf.
Al-Omari, R., Somani, A. K., and Manimaran, G. (2005). An adaptive scheme for fault-tolerant scheduling of soft real-time tasks in multiprocessor systems. J. Parallel Distrib. Comput., 65(5):595–608.
Anne Benoit, M. H. and Robert, Y. (April 14-18, 2008). Fault tolerant scheduling of precedence task graphs on heterogeneous platforms. In Proceedings of the 22th ACM/IEEE International Parallel Distributed Processing Symposium IPDPS’08 - APDCM’08 IEEE Computer Society Press, Miami, Florida, USA.
Aversa, R., Di Martino, B., Rak, M., Venticinque, S., and Villano, U. (2005). Performance prediction through simulation of a hybrid MPI/OpenMP application. Parallel Comput.
Benoit, A., Hakem, M., and Robert, Y. (2008). Realistic models and efficient algorithms for fault tolerant scheduling on heterogeneous platforms. In Parallel Processing, 2008. ICPP ’08. 37th International Conference on, pages 8–12, Portland, Oregon, USA.
Boeres, C. and Rebello, V. E. F. (2004). Easygrid: towards a framework for the automatic grid enabling of legacy MPI applications: Research articles. Concurrency And Computation : Practice And Experience, 16(5):425–432.
Chorley, M. J., Walker, D. W., and Guest, M. F. (2009). Hybrid message-passing and shared-memory programming in a molecular dynamics application on multicore clusters. In Int. J. High Perform. Comput., pages 196–211.
da Silva, J. A. (2010). Tolerância a Falhas para Aplicações Autônomas em Grades Computacionais. PhD thesis, Instituto de Computação, Universidade Federal Fluminense, Niterói, RJ, Brasil.
de P. Nascimento, A., da C. Sena, A., da Silva, J. A., de C. Vianna, D. Q., Boeres, C., and Rebello, V. E. F. (2005). Managing the execution of large scale MPI applications on computational grids. 17th. International Symposium on Computer Architecture and High Performance Computing.
Qin, X. and Jiang, H. (2006). A novel fault-tolerant scheduling algorithm for precedence constrained tasks in real-time heterogeneous systems. Parallel Computing, 32(5):331–356.
Rabenseifner, R., Hager, G., and Jost, G. (2009a). Communication characteristics and hybrid MPI/OpenMP parallel programming on clusters of multi-core smp nodes. Proceedings of the Cray Users Group Conference 2009 (CUG 2009).
Rabenseifner, R., Hager, G., and Jost, G. (2009b). Hybrid MPI/OpenMP parallel programming on clusters of multi-core smp nodes.
Sardina, I. M. (2010). Escalonamento Estático de Tarefas Bi-objetivo e Tolerante a Falhas para Sistemas Distribuídos. PhD thesis, Instituto de Computação, Universidade Federal Fluminense, Niterói, RJ, Brasil.
Sardina, I. M., Boeres, C., and Drummond, L. M. A. (2011a). An efficient weighted bi-objective scheduling algorithm for heterogeneous systems. Parallel Computing, 37:349–364.
Sardina, I. M., Boeres, C., and Drummond, L. M. A. (2011b). Escalonamento estático bi-objetivo e tolerante a falhas para sistemas distribuídos. In Simposio Brasileiro de Redes de Computadores e Sistemas Distribuídos, Campo Grande.
Su, M. F., El-Kady, I., Bader, D. A., and Lin, S. (2004). A novel fdtd application featuring openmp-mpi hybrid parallelization. In Proceedings of the 2004 international Conference on Parallel Processing ICPP, pages 373–379, Washington, DC, USA. IEEE Computer Society.
Topcuouglu, H., Hariri, S., and Wu, M. (2002). Performance-effective and low-complexity task scheduling for heterogeneous computing. IEEE Transactions Parallel Distributed Systems, 13(3):260–274.
(Nov, 2009). MPI: A message-passing interface standard version 2.1. https://www.mpi-forum.org/docs/mpi21-report.pdf.
Al-Omari, R., Somani, A. K., and Manimaran, G. (2005). An adaptive scheme for fault-tolerant scheduling of soft real-time tasks in multiprocessor systems. J. Parallel Distrib. Comput., 65(5):595–608.
Anne Benoit, M. H. and Robert, Y. (April 14-18, 2008). Fault tolerant scheduling of precedence task graphs on heterogeneous platforms. In Proceedings of the 22th ACM/IEEE International Parallel Distributed Processing Symposium IPDPS’08 - APDCM’08 IEEE Computer Society Press, Miami, Florida, USA.
Aversa, R., Di Martino, B., Rak, M., Venticinque, S., and Villano, U. (2005). Performance prediction through simulation of a hybrid MPI/OpenMP application. Parallel Comput.
Benoit, A., Hakem, M., and Robert, Y. (2008). Realistic models and efficient algorithms for fault tolerant scheduling on heterogeneous platforms. In Parallel Processing, 2008. ICPP ’08. 37th International Conference on, pages 8–12, Portland, Oregon, USA.
Boeres, C. and Rebello, V. E. F. (2004). Easygrid: towards a framework for the automatic grid enabling of legacy MPI applications: Research articles. Concurrency And Computation : Practice And Experience, 16(5):425–432.
Chorley, M. J., Walker, D. W., and Guest, M. F. (2009). Hybrid message-passing and shared-memory programming in a molecular dynamics application on multicore clusters. In Int. J. High Perform. Comput., pages 196–211.
da Silva, J. A. (2010). Tolerância a Falhas para Aplicações Autônomas em Grades Computacionais. PhD thesis, Instituto de Computação, Universidade Federal Fluminense, Niterói, RJ, Brasil.
de P. Nascimento, A., da C. Sena, A., da Silva, J. A., de C. Vianna, D. Q., Boeres, C., and Rebello, V. E. F. (2005). Managing the execution of large scale MPI applications on computational grids. 17th. International Symposium on Computer Architecture and High Performance Computing.
Qin, X. and Jiang, H. (2006). A novel fault-tolerant scheduling algorithm for precedence constrained tasks in real-time heterogeneous systems. Parallel Computing, 32(5):331–356.
Rabenseifner, R., Hager, G., and Jost, G. (2009a). Communication characteristics and hybrid MPI/OpenMP parallel programming on clusters of multi-core smp nodes. Proceedings of the Cray Users Group Conference 2009 (CUG 2009).
Rabenseifner, R., Hager, G., and Jost, G. (2009b). Hybrid MPI/OpenMP parallel programming on clusters of multi-core smp nodes.
Sardina, I. M. (2010). Escalonamento Estático de Tarefas Bi-objetivo e Tolerante a Falhas para Sistemas Distribuídos. PhD thesis, Instituto de Computação, Universidade Federal Fluminense, Niterói, RJ, Brasil.
Sardina, I. M., Boeres, C., and Drummond, L. M. A. (2011a). An efficient weighted bi-objective scheduling algorithm for heterogeneous systems. Parallel Computing, 37:349–364.
Sardina, I. M., Boeres, C., and Drummond, L. M. A. (2011b). Escalonamento estático bi-objetivo e tolerante a falhas para sistemas distribuídos. In Simposio Brasileiro de Redes de Computadores e Sistemas Distribuídos, Campo Grande.
Su, M. F., El-Kady, I., Bader, D. A., and Lin, S. (2004). A novel fdtd application featuring openmp-mpi hybrid parallelization. In Proceedings of the 2004 international Conference on Parallel Processing ICPP, pages 373–379, Washington, DC, USA. IEEE Computer Society.
Topcuouglu, H., Hariri, S., and Wu, M. (2002). Performance-effective and low-complexity task scheduling for heterogeneous computing. IEEE Transactions Parallel Distributed Systems, 13(3):260–274.
Publicado
06/05/2013
Como Citar
SILVA, Brevik Ferreira da; SANTOS, Wellison Moura dos; SARDIÑA, Idalmis Milián; TEIXEIRA, Livia de Mesquita; ALBUQUERQUE, Felipe de.
Escalonamento Tolerante a Falhas para Clusters Multicores. In: WORKSHOP DE TESTES E TOLERÂNCIA A FALHAS (WTF), 14. , 2013, Brasília/DF.
Anais [...].
Porto Alegre: Sociedade Brasileira de Computação,
2013
.
p. 77-88.
ISSN 2595-2684.
DOI: https://doi.org/10.5753/wtf.2013.23017.
