Avaliação de Desempenho de um Sistema de Programação Paralela Baseado em Tarefas Assíncronas
Resumo
Atualmente há uma disseminação pela exploração eficiente de recursos de processamento em computação paralela, incluindo coprocessadores e recursos remotos. A programação paralela ainda hoje é uma tarefa complexa, pois exige o entendimento das estratégias de paralelização. Esta dificuldade se torna ainda maior com a necessidade da exploração de recursos heterogêneos. Este artigo apresenta a avaliação de desempenho do sistema RTE, que oferece uma interface comum de programação paralela baseada em tarefas assíncronas a serem executadas em recursos de processamento heterogêneos. A avaliação foi realizada através de uma aplicação de Bioinformática para análise do genoma de diversos animais. Os resultados obtidos mostram que o RTE facilita a tarefa de desenvolvimento e a aplicação implementada com RTE apresentou um desempenho similar à aplicação usando a interface MPI.
Palavras-chave:
Computação de Alto Desempenho, Programação Paralela, Tarefas Assíncronas, Sistemas Heterogêneos
Referências
Broner, G. (2017). Supercomputing is the future of genomics research. Genetic Engineering & Biotechnology News, 37(3):18–19. https://www.genengnews.com/magazine/286/supercomputing-is-the-future-of-genomics-research/. Acessado em: 06/02/2020.
Dash, Y. (2019). An insight into parallel computing paradigm. In 2019 2nd International Conference on Intelligent Computing, Instrumentation and Control Technologies (ICI-CICT), volume 1, pages 804–808. IEEE. https://ieeexplore.ieee.org/abstract/document/8993136. Acessado em: 10/02/2020.
Devarakonda, R., Prakash, G., Guntupally, K., and Kumar, J. (2019). Big federal datacenters implementing fair data principles: Arm data center example. In 2019 IEEE International Conference on Big Data (Big Data), pages 6033–6036. IEEE. https://ieeexplore.ieee.org/abstract/document/9006051. Acessado em:19/02/1972.
Dolbeau, R., Bihan, S., and Bodin, F. (2007). HmppTM: A hybrid multi-core parallelprogramming environment. 28. https://www.researchgate.net/publication/240064180HMPPAhybridmulti-coreparallelprogrammingenvironment. Acessado em: 18/01/2020.Intel (2020).
Intel® oneAPI Programming Guide. Version Beta. https://software.intel.com/sites/default/files/oneAPIProgrammingGuide9.pdf. Acessado em: 27/03/2020.
Jeffers, J., Reinders, J., and Sodani, A. (2016).IntelR©Xeon PhiTMProcessor High Performance Programming: Knights Landing Edition. Morgan Kaufmann.
Keryell, R., Rovatsou, M., and Howes, L. (2019). SyclTMSpecification: SyclTMintegratesOpenCLTMdevices with modern C++.Khronos Working Group, Version 1.2.1. https://www.khronos.org/registry/SYCL/specs/sycl-1.2.1.pdf. Acessado em: 27/03/2020.
Khronos (2019). The OpenCLTMSpecification.Khronos Working Group, Version V2.2-11.https://www.khronos.org/registry/OpenCL/specs/2.2/pdf/OpenCLAPI.pdf. Acessado em: 20/01/2020.
Kirk, D. B. and Hwu, W.-m. W. (2017).Programming Massively Parallel Processors: A Hands-on Approach. Morgan kaufmann, 3rd edition.
LNCC (2019). Configuração do sdumont. https://sdumont.lncc.br/machine.php?pg=machine#. Acessado em: 15/12/2019.
MPI (2015). MPI: A Message-Passing Interface Standard.Message Passing Interface Forum, Version 3.1. https://www.mpi-forum.org/docs/mpi-3.1/mpi31-report.pdf. Acessado em: 11/01/2020.
NVIDIA (2019). CUDA C++ Programming Guide.NVIDIA Coporation, Version v10.2. https://docs.nvidia.com/cuda/pdf/CUDACProgrammingGuide.pdf. Acessado em: 17/01/2020.
OpenACC (2019). The OpenACCTMApplication Programming Interface. Retrieved March, Version 3.0. https://www.openacc.org/sites/default/files/inline-images/Specification/OpenACC.3.0.pdf. Acessado em:23/01/2020.
OpenMP (2018). Openmp application program interface. Version 5.0. https://www.openmp.org/wp-content/uploads/OpenMP-API-Specification-5.0.pdf. Acessado em: 06/01/20201.
Schmidt, B. and Hildebrandt, A. (2017). Next-generation sequencing: big data meets high performance computing. Drug discovery today, 22(4):712–717. https://www.sciencedirect.com/science/article/pii/S1359644617300582. Acessado em: 15/02/2020.Top500 (2019).
TOP500 List - November 2019. https://www.top500.org/list/2019/11/. Acessado em: 15/12/2019.
Yasuda-Masuoka, Y., Kwon, S., and Yoon, J. (2019). Foundry platform technology from high-performance to low-power for new high-performance computing (hpc) and iotera. pages 1–3. https://ieeexplore.ieee.org/abstract/document/8731116. Acessado em: 11/02/2020
Dash, Y. (2019). An insight into parallel computing paradigm. In 2019 2nd International Conference on Intelligent Computing, Instrumentation and Control Technologies (ICI-CICT), volume 1, pages 804–808. IEEE. https://ieeexplore.ieee.org/abstract/document/8993136. Acessado em: 10/02/2020.
Devarakonda, R., Prakash, G., Guntupally, K., and Kumar, J. (2019). Big federal datacenters implementing fair data principles: Arm data center example. In 2019 IEEE International Conference on Big Data (Big Data), pages 6033–6036. IEEE. https://ieeexplore.ieee.org/abstract/document/9006051. Acessado em:19/02/1972.
Dolbeau, R., Bihan, S., and Bodin, F. (2007). HmppTM: A hybrid multi-core parallelprogramming environment. 28. https://www.researchgate.net/publication/240064180HMPPAhybridmulti-coreparallelprogrammingenvironment. Acessado em: 18/01/2020.Intel (2020).
Intel® oneAPI Programming Guide. Version Beta. https://software.intel.com/sites/default/files/oneAPIProgrammingGuide9.pdf. Acessado em: 27/03/2020.
Jeffers, J., Reinders, J., and Sodani, A. (2016).IntelR©Xeon PhiTMProcessor High Performance Programming: Knights Landing Edition. Morgan Kaufmann.
Keryell, R., Rovatsou, M., and Howes, L. (2019). SyclTMSpecification: SyclTMintegratesOpenCLTMdevices with modern C++.Khronos Working Group, Version 1.2.1. https://www.khronos.org/registry/SYCL/specs/sycl-1.2.1.pdf. Acessado em: 27/03/2020.
Khronos (2019). The OpenCLTMSpecification.Khronos Working Group, Version V2.2-11.https://www.khronos.org/registry/OpenCL/specs/2.2/pdf/OpenCLAPI.pdf. Acessado em: 20/01/2020.
Kirk, D. B. and Hwu, W.-m. W. (2017).Programming Massively Parallel Processors: A Hands-on Approach. Morgan kaufmann, 3rd edition.
LNCC (2019). Configuração do sdumont. https://sdumont.lncc.br/machine.php?pg=machine#. Acessado em: 15/12/2019.
MPI (2015). MPI: A Message-Passing Interface Standard.Message Passing Interface Forum, Version 3.1. https://www.mpi-forum.org/docs/mpi-3.1/mpi31-report.pdf. Acessado em: 11/01/2020.
NVIDIA (2019). CUDA C++ Programming Guide.NVIDIA Coporation, Version v10.2. https://docs.nvidia.com/cuda/pdf/CUDACProgrammingGuide.pdf. Acessado em: 17/01/2020.
OpenACC (2019). The OpenACCTMApplication Programming Interface. Retrieved March, Version 3.0. https://www.openacc.org/sites/default/files/inline-images/Specification/OpenACC.3.0.pdf. Acessado em:23/01/2020.
OpenMP (2018). Openmp application program interface. Version 5.0. https://www.openmp.org/wp-content/uploads/OpenMP-API-Specification-5.0.pdf. Acessado em: 06/01/20201.
Schmidt, B. and Hildebrandt, A. (2017). Next-generation sequencing: big data meets high performance computing. Drug discovery today, 22(4):712–717. https://www.sciencedirect.com/science/article/pii/S1359644617300582. Acessado em: 15/02/2020.Top500 (2019).
TOP500 List - November 2019. https://www.top500.org/list/2019/11/. Acessado em: 15/12/2019.
Yasuda-Masuoka, Y., Kwon, S., and Yoon, J. (2019). Foundry platform technology from high-performance to low-power for new high-performance computing (hpc) and iotera. pages 1–3. https://ieeexplore.ieee.org/abstract/document/8731116. Acessado em: 11/02/2020
Publicado
30/06/2020
Como Citar
BÉLO, Denilson Souza; SATO, Liria; MIDORIKAWA, Edson.
Avaliação de Desempenho de um Sistema de Programação Paralela Baseado em Tarefas Assíncronas. In: WORKSHOP EM DESEMPENHO DE SISTEMAS COMPUTACIONAIS E DE COMUNICAÇÃO (WPERFORMANCE), 19. , 2020, Cuiabá.
Anais [...].
Porto Alegre: Sociedade Brasileira de Computação,
2020
.
p. 49-60.
ISSN 2595-6167.
DOI: https://doi.org/10.5753/wperformance.2020.11105.
