2016
|
Maron, Carlos A. F.; Griebler, Dalvan; Schepke, Claudio; Fernandes, Luiz Gustavo Desempenho de OpenStack e OpenNebula em Estações de Trabalho: Uma Avaliação com Microbenchmarks e NPB Journal Article doi In: Revista Eletrônica Argentina-Brasil de Tecnologias da Informação e da Comunicação (REABTIC), vol. 1, no. 6, pp. 15, 2016. @article{larcc:nas_workstations:REABTIC:16,
title = {Desempenho de OpenStack e OpenNebula em Estações de Trabalho: Uma Avaliação com Microbenchmarks e NPB},
author = {Carlos A. F. Maron and Dalvan Griebler and Claudio Schepke and Luiz Gustavo Fernandes},
url = {http://larcc.setrem.com.br/wp-content/uploads/2017/03/MARON_REABTIC_2016.pdf},
doi = {10.5281/zenodo.345597},
year = {2016},
date = {2016-12-01},
journal = {Revista Eletrônica Argentina-Brasil de Tecnologias da Informação e da Comunicação (REABTIC)},
volume = {1},
number = {6},
pages = {15},
publisher = {SETREM},
address = {Três de Maio, Brazil},
abstract = {IaaS (Infrastructure as a Service) clouds provide on-demand computing resources (i.e, memory, networking, storage and processing unit) for running applications. Studies that evaluate the IaaS cloud performance are limited to the virtualization layer and ignore the impact of management tools analysis. In contrast, our research investigates the impact of them in order to identify if there are influences or differences between OpenStack and OpenNebula. We used intensive workloads (microbenchmarks) and scientific parallel applications. Statistically, the results demonstrated that OpenNebula was 11.07% better using microbenchmarks and 8.41% with scientific parallel applications.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
IaaS (Infrastructure as a Service) clouds provide on-demand computing resources (i.e, memory, networking, storage and processing unit) for running applications. Studies that evaluate the IaaS cloud performance are limited to the virtualization layer and ignore the impact of management tools analysis. In contrast, our research investigates the impact of them in order to identify if there are influences or differences between OpenStack and OpenNebula. We used intensive workloads (microbenchmarks) and scientific parallel applications. Statistically, the results demonstrated that OpenNebula was 11.07% better using microbenchmarks and 8.41% with scientific parallel applications. |
Maron, Carlos A. F.; Vogel, Adriano; Benedetti, Vera L. L.; Shubeita, Fauzi; Schepke, Claudio; Griebler, Dalvan Panorama Geral e Resultados do Projeto HiPerfCloud Inproceedings In: 15th Jornada de Pesquisa SETREM, pp. 4, SETREM, Três de Maio, Brazil, 2016. @inproceedings{larcc:hiperfcloud:JP:16,
title = {Panorama Geral e Resultados do Projeto HiPerfCloud},
author = {Carlos A. F. Maron and Adriano Vogel and Vera L. L. Benedetti and Fauzi Shubeita and Claudio Schepke and Dalvan Griebler},
url = {http://larcc.setrem.com.br/wp-content/uploads/2017/03/HiPerfCloud_JP_SETREM_2016.pdf},
year = {2016},
date = {2016-10-01},
booktitle = {15th Jornada de Pesquisa SETREM},
pages = {4},
publisher = {SETREM},
address = {Três de Maio, Brazil},
abstract = {O projeto HiPerfCloud, em andamento no LARCC da Faculdade SETREM, desenvolve pesquisas em nível de infraestrutura em nuvens computacionais. O objetivo do projeto é analisar o impacto que aplicações científicas de alto desempenho sofrem quando executadas em nuvens privadas e avaliar as tecnologias de implantação envolvidas. As publicações de artigos em eventos nacionais e internacionais do projeto tem colaborado com o estado da arte da área. As descobertas recentes apontaram que aspectos de infraestrutura, rede e virtualização, exercem influência no desempenho de aplicações executadas em nuvem, enquanto as ferramentas de IaaS possuem contrastes em relação ao gerenciamento (escalonamento, disponibilidade, segurança).},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
O projeto HiPerfCloud, em andamento no LARCC da Faculdade SETREM, desenvolve pesquisas em nível de infraestrutura em nuvens computacionais. O objetivo do projeto é analisar o impacto que aplicações científicas de alto desempenho sofrem quando executadas em nuvens privadas e avaliar as tecnologias de implantação envolvidas. As publicações de artigos em eventos nacionais e internacionais do projeto tem colaborado com o estado da arte da área. As descobertas recentes apontaram que aspectos de infraestrutura, rede e virtualização, exercem influência no desempenho de aplicações executadas em nuvem, enquanto as ferramentas de IaaS possuem contrastes em relação ao gerenciamento (escalonamento, disponibilidade, segurança). |
Pieper, Ricardo; Griebler, Dalvan; Lovato, Adalberto Towards a Software as a Service for Biodigestor Analytics Journal Article doi In: Revista Eletrônica Argentina-Brasil de Tecnologias da Informação e da Comunicação (REABTIC), vol. 1, no. 5, pp. 15, 2016. @article{larcc:saas_analytics:REABTIC:16,
title = {Towards a Software as a Service for Biodigestor Analytics},
author = {Ricardo Pieper and Dalvan Griebler and Adalberto Lovato},
url = {http://larcc.setrem.com.br/wp-content/uploads/2017/04/PIEPER_REABTIC_2016.pdf},
doi = {10.5281/zenodo.345587},
year = {2016},
date = {2016-08-01},
journal = {Revista Eletrônica Argentina-Brasil de Tecnologias da Informação e da Comunicação (REABTIC)},
volume = {1},
number = {5},
pages = {15},
publisher = {SETREM},
address = {Três de Maio, Brazil},
abstract = {The field of machine learning is becoming even more important in the last years. The ever-increasing amount of data and complexity of computational problems challenges the currently available technology. Meanwhile, anaerobic digesters represent a good alternative for renewable energy production in Brazil. However, performing efficient and accurate predictions/analytics while completely abstracting machine learning details from end-users might not be a simple task to achieve. Usually, such tools are made for a specific scenario and may not fit with particular and general needs. Our goal was to create a SaaS for biogas data analytics by using a neural network. Therefore, an open source, cloud-enabled SaaS (Software as a Service) was developed and deployed in LARCC (Laboratory of Advanced Researches on Cloud Computing) at SETREM. The results have shown the SaaS application is able to perform predictions. The neural network's accuracy is not significantly worse than a state-of-the-art implementation, and its training speed is faster. The user interface demonstrates to be intuitive, and the predictions were accurate when providing the training algorithm with sufficient data. In addition, the file processing and network training time were good enough under traditional workload conditions.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
The field of machine learning is becoming even more important in the last years. The ever-increasing amount of data and complexity of computational problems challenges the currently available technology. Meanwhile, anaerobic digesters represent a good alternative for renewable energy production in Brazil. However, performing efficient and accurate predictions/analytics while completely abstracting machine learning details from end-users might not be a simple task to achieve. Usually, such tools are made for a specific scenario and may not fit with particular and general needs. Our goal was to create a SaaS for biogas data analytics by using a neural network. Therefore, an open source, cloud-enabled SaaS (Software as a Service) was developed and deployed in LARCC (Laboratory of Advanced Researches on Cloud Computing) at SETREM. The results have shown the SaaS application is able to perform predictions. The neural network's accuracy is not significantly worse than a state-of-the-art implementation, and its training speed is faster. The user interface demonstrates to be intuitive, and the predictions were accurate when providing the training algorithm with sufficient data. In addition, the file processing and network training time were good enough under traditional workload conditions. |
Barth, Andréia; Wolfer, Camila; Lovato, Adalberto; Griebler, Dalvan Avaliação da Irradiação Solar como Fonte de Energia Renovável no Noroeste do Estado do Rio Grande do Sul Através de Uma Rede Neural Journal Article doi In: Revista Eletrônica Argentina-Brasil de Tecnologias da Informação e da Comunicação (REABTIC), vol. 1, no. 5, pp. 15, 2016. @article{larcc:neural_networks:REABTIC:16,
title = {Avaliação da Irradiação Solar como Fonte de Energia Renovável no Noroeste do Estado do Rio Grande do Sul Através de Uma Rede Neural},
author = {Andréia Barth and Camila Wolfer and Adalberto Lovato and Dalvan Griebler},
url = {http://larcc.setrem.com.br/wp-content/uploads/2018/02/ANDREIA_CAMILA_REABTIC_2016.pdf},
doi = {10.5281/zenodo.345585},
year = {2016},
date = {2016-08-01},
journal = {Revista Eletrônica Argentina-Brasil de Tecnologias da Informação e da Comunicação (REABTIC)},
volume = {1},
number = {5},
pages = {15},
publisher = {SETREM},
address = {Três de Maio, RS, Brazil},
abstract = {Solar irradiation is one of the cleanest renewable energy sources of nowadays. In this work, the goal was to implement a neural network capable of evaluating the solar irradiation in the Northwest region of Rio Grande do Sul. In case, this assessment targets meteorological data, from January to April 2015. The network Perceptron was implemented and trained using MATLAB software. The results have indicated that the system obtained a highly accurate and that the region is a good enough place for stemmed energy production of solar irradiation.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Solar irradiation is one of the cleanest renewable energy sources of nowadays. In this work, the goal was to implement a neural network capable of evaluating the solar irradiation in the Northwest region of Rio Grande do Sul. In case, this assessment targets meteorological data, from January to April 2015. The network Perceptron was implemented and trained using MATLAB software. The results have indicated that the system obtained a highly accurate and that the region is a good enough place for stemmed energy production of solar irradiation. |
Griebler, Dalvan Domain-Specific Language & Support Tool for High-Level Stream Parallelism PhD Thesis Faculdade de Informática - PPGCC - PUCRS, 2016. @phdthesis{GRIEBLER:PHD:16,
title = {Domain-Specific Language & Support Tool for High-Level Stream Parallelism},
author = {Dalvan Griebler},
url = {http://tede2.pucrs.br/tede2/handle/tede/6776},
year = {2016},
date = {2016-06-01},
address = {Porto Alegre, Brazil},
school = {Faculdade de Informática - PPGCC - PUCRS},
abstract = {Stream-based systems are representative of several application domains including video, audio, networking, graphic processing, etc. Stream programs may run on different kinds of parallel architectures (desktop, servers, cell phones, and supercomputers) and represent significant workloads on our current computing systems. Nevertheless, most of them are still not parallelized. Moreover, when new software has to be developed, programmers often face a trade-off between coding productivity, code portability, and performance. To solve this problem, we provide a new Domain-Specific Language (DSL) that naturally/on-the-fly captures and represents parallelism for stream-based applications. The aim is to offer a set of attributes (through annotations) that preserves the program's source code and is not architecture-dependent for annotating parallelism. We used the C++ attribute mechanism to design a ``textitde-facto'' standard C++ embedded DSL named SPar. However, the implementation of DSLs using compiler-based tools is difficult, complicated, and usually requires a significant learning curve. This is even harder for those who are not familiar with compiler technology. Therefore, our motivation is to simplify this path for other researchers (experts in their domain) with support tools (our tool is CINCLE) to create high-level and productive DSLs through powerful and aggressive source-to-source transformations. In fact, parallel programmers can use their expertise without having to design and implement low-level code. The main goal of this thesis was to create a DSL and support tools for high-level stream parallelism in the context of a programming framework that is compiler-based and domain-oriented. Thus, we implemented SPar using CINCLE. SPar supports the software developer with productivity, performance, and code portability while CINCLE provides sufficient support to generate new DSLs. Also, SPar targets source-to-source transformation producing parallel pattern code built on top of FastFlow and MPI. Finally, we provide a full set of experiments showing that SPar provides better coding productivity without significant performance degradation in multi-core systems as well as transformation rules that are able to achieve code portability (for cluster architectures) through its generalized attributes.},
keywords = {},
pubstate = {published},
tppubtype = {phdthesis}
}
Stream-based systems are representative of several application domains including video, audio, networking, graphic processing, etc. Stream programs may run on different kinds of parallel architectures (desktop, servers, cell phones, and supercomputers) and represent significant workloads on our current computing systems. Nevertheless, most of them are still not parallelized. Moreover, when new software has to be developed, programmers often face a trade-off between coding productivity, code portability, and performance. To solve this problem, we provide a new Domain-Specific Language (DSL) that naturally/on-the-fly captures and represents parallelism for stream-based applications. The aim is to offer a set of attributes (through annotations) that preserves the program's source code and is not architecture-dependent for annotating parallelism. We used the C++ attribute mechanism to design a ``textitde-facto'' standard C++ embedded DSL named SPar. However, the implementation of DSLs using compiler-based tools is difficult, complicated, and usually requires a significant learning curve. This is even harder for those who are not familiar with compiler technology. Therefore, our motivation is to simplify this path for other researchers (experts in their domain) with support tools (our tool is CINCLE) to create high-level and productive DSLs through powerful and aggressive source-to-source transformations. In fact, parallel programmers can use their expertise without having to design and implement low-level code. The main goal of this thesis was to create a DSL and support tools for high-level stream parallelism in the context of a programming framework that is compiler-based and domain-oriented. Thus, we implemented SPar using CINCLE. SPar supports the software developer with productivity, performance, and code portability while CINCLE provides sufficient support to generate new DSLs. Also, SPar targets source-to-source transformation producing parallel pattern code built on top of FastFlow and MPI. Finally, we provide a full set of experiments showing that SPar provides better coding productivity without significant performance degradation in multi-core systems as well as transformation rules that are able to achieve code portability (for cluster architectures) through its generalized attributes. |
