O Impacto da Interconexão de Rede no Desempenho de Programas Paralelos
Resumo
O desempenho de aplicações paralelas depende de dois componentes principais do ambiente; o poder de processamento e a interconexão de rede. Neste trabalho, foi avaliado o impacto de uma interconexão de alto desempenho em programas paralelos em um cluster homogêneo de servidores interconectados por Gigabit Ethernet 1 Gbps e InfiniBand FDR 56 Gbps. Foi realizada uma caracterização do NAS Parallel Benchmarks em relação à computação, comunicação e custo de execução em instâncias da Microsoft Azure. Os resultados mostraram que, em aplicações altamente dependentes de rede, o desempenho pode ser significativamente melhorado ao utilizar InfiniBand a um custo de execução melhor, mesmo com o preço superior da instância.
Referências
[Alles et al. 2018] Alles, G. R., Carissimi, A., and Schnorr, L. M. (2018). Assessing the Computation and Communication Overhead of Linux Containers for HPC Applications. In Symposium on High Performance Computing Systems (WSCAD).
[Bailey et al. 1991] Bailey, D. H., Barszcz, E., Barton, J. T., Browning, D. S., Carter, R. L., Dagum, L., Fatoohi, R. A., Frederickson, P. O., Lasinski, T. A., Schreiber, R. S., Simon, H. D., Venkatakrishnan, V., and Weeratunga, S. K. (1991). The NAS Parallel Benchmarks
[Chakthranont et al. 2014] Chakthranont, N., Khunphet, P., Takano, R., and Ikegami, T. (2014). Exploring the Performance Impact of Virtualization on an HPC Cloud. In International Conference on Cloud Computing Technology and Science (CloudCom).
[Escudero-Sahuquillo et al. 2015] Escudero-Sahuquillo, J., Gran, E. G., Garcia, P. J., Flich, J., Skeie, T., Lysne, O., Quiles, F. J., and Duato, J. (2015). Efficient and CostEffective Hybrid Congestion Control for HPC Interconnection Networks. Transactions on Parallel and Distributed Systems (TPDS).
[Kamburugamuve et al. 2017] Kamburugamuve, S., Ramasamy, K., Swany, M., and Fox, G. (2017). Low Latency Stream Processing: Apache Heron with Infiniband/Intel Omni-Path. In International Conference on Utility and Cloud Computing (UCC).
[Liu et al. 2004] Liu, J., Wu, J., and Panda, D. K. (2004). High Performance RDMAbased MPI Implementation over InfiniBand. International Journal of Parallel Programming (IJPP).
[Maliszewski et al. 2019] Maliszewski, A. M., Vogel, A., Griebler, D., Roloff, E., Fernandes, L. G., and Navaux, P. O. A. (2019). Minimizing Communication Overheads in Container-based Clouds for HPC Applications. In Symposium on Computers and Communications (ISCC).
[Moura and Hutchison 2016] Moura, J. and Hutchison, D. (2016). Review and Analysis of Networking Challenges in Cloud Computing. Journal of Network and Computer Applications (JNCA).
[Roloff et al. 2017] Roloff, E., Diener, M., Gaspary, L. P., and Navaux, P. O. A. (2017) HPC Application Performance and Cost Efficiency in the Cloud. In Euromicro International Conference on Parallel, Distributed and Network-based Processing (PDP).
[Ruivo et al. 2014] Ruivo, T. P. P. D. L., Altayo, G. B., Garzoglio, G., Timm, S., Kim, H. W., Noh, S., and Raicu, I. (2014). Exploring Infiniband Hardware Virtualization in OpenNebula towards Efficient High-Performance Computing. In IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing (CCGRID).
[Vienne et al. 2012] Vienne, J., Chen, J., Wasi-Ur-Rahman, M., Islam, N. S., Subramoni, H., and Panda, D. K. (2012). Performance Analysis and Evaluation of InfiniBand FDR and 40GigE RoCE on HPC and Cloud Computing Systems. In Symposium on High-Performance Interconnects (HOTI).
[Zahid 2017] Zahid, F. (2017). Network Optimization for High Performance Cloud Computing. PhD thesis, Faculty of Mathematics and Natural Sciences, University of Oslo, Oslo, Norway.
[Zhang et al. 2016] Zhang, J., Lu, X., and Panda, D. K. (2016). Performance Characterization of Hypervisor-and Container-Based Virtualization for HPC on SR-IOV Enabled InfiniBand Clusters. In International Parallel and Distributed Processing Symposium Workshops (IPDPSW).
[Bailey et al. 1991] Bailey, D. H., Barszcz, E., Barton, J. T., Browning, D. S., Carter, R. L., Dagum, L., Fatoohi, R. A., Frederickson, P. O., Lasinski, T. A., Schreiber, R. S., Simon, H. D., Venkatakrishnan, V., and Weeratunga, S. K. (1991). The NAS Parallel Benchmarks
[Chakthranont et al. 2014] Chakthranont, N., Khunphet, P., Takano, R., and Ikegami, T. (2014). Exploring the Performance Impact of Virtualization on an HPC Cloud. In International Conference on Cloud Computing Technology and Science (CloudCom).
[Escudero-Sahuquillo et al. 2015] Escudero-Sahuquillo, J., Gran, E. G., Garcia, P. J., Flich, J., Skeie, T., Lysne, O., Quiles, F. J., and Duato, J. (2015). Efficient and CostEffective Hybrid Congestion Control for HPC Interconnection Networks. Transactions on Parallel and Distributed Systems (TPDS).
[Kamburugamuve et al. 2017] Kamburugamuve, S., Ramasamy, K., Swany, M., and Fox, G. (2017). Low Latency Stream Processing: Apache Heron with Infiniband/Intel Omni-Path. In International Conference on Utility and Cloud Computing (UCC).
[Liu et al. 2004] Liu, J., Wu, J., and Panda, D. K. (2004). High Performance RDMAbased MPI Implementation over InfiniBand. International Journal of Parallel Programming (IJPP).
[Maliszewski et al. 2019] Maliszewski, A. M., Vogel, A., Griebler, D., Roloff, E., Fernandes, L. G., and Navaux, P. O. A. (2019). Minimizing Communication Overheads in Container-based Clouds for HPC Applications. In Symposium on Computers and Communications (ISCC).
[Moura and Hutchison 2016] Moura, J. and Hutchison, D. (2016). Review and Analysis of Networking Challenges in Cloud Computing. Journal of Network and Computer Applications (JNCA).
[Roloff et al. 2017] Roloff, E., Diener, M., Gaspary, L. P., and Navaux, P. O. A. (2017) HPC Application Performance and Cost Efficiency in the Cloud. In Euromicro International Conference on Parallel, Distributed and Network-based Processing (PDP).
[Ruivo et al. 2014] Ruivo, T. P. P. D. L., Altayo, G. B., Garzoglio, G., Timm, S., Kim, H. W., Noh, S., and Raicu, I. (2014). Exploring Infiniband Hardware Virtualization in OpenNebula towards Efficient High-Performance Computing. In IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing (CCGRID).
[Vienne et al. 2012] Vienne, J., Chen, J., Wasi-Ur-Rahman, M., Islam, N. S., Subramoni, H., and Panda, D. K. (2012). Performance Analysis and Evaluation of InfiniBand FDR and 40GigE RoCE on HPC and Cloud Computing Systems. In Symposium on High-Performance Interconnects (HOTI).
[Zahid 2017] Zahid, F. (2017). Network Optimization for High Performance Cloud Computing. PhD thesis, Faculty of Mathematics and Natural Sciences, University of Oslo, Oslo, Norway.
[Zhang et al. 2016] Zhang, J., Lu, X., and Panda, D. K. (2016). Performance Characterization of Hypervisor-and Container-Based Virtualization for HPC on SR-IOV Enabled InfiniBand Clusters. In International Parallel and Distributed Processing Symposium Workshops (IPDPSW).
Publicado
08/11/2019
Como Citar
MALISZEWSKI, Anderson; ROLOFF, Eduardo; GRIEBLER, Dalvan; NAVAUX, Philippe Olivier.
O Impacto da Interconexão de Rede no Desempenho de Programas Paralelos. In: SIMPÓSIO EM SISTEMAS COMPUTACIONAIS DE ALTO DESEMPENHO (SSCAD), 20. , 2019, Campo Grande.
Anais [...].
Porto Alegre: Sociedade Brasileira de Computação,
2019
.
p. 73-84.
DOI: https://doi.org/10.5753/wscad.2019.8658.
