<p>This dissertation provides constructive proof, through the implementation of a middleware, that distribution transparency is practical, generic, and extensible. Fault tolerant distributed services can be developed by using the failure detection abilities of the middleware. By generic we mean that the middleware can be used for many different programming languages and paradigms. Distribution for each kind of language entity is done in terms of consistency protocols, which guarantee that the semantics of the entities are preserved in a distributed setting. The middleware allows new consistency protocols to be added easily. The efficiency of the middleware and the ease of integration are shown by coupling the middleware to a programming system, which encompasses the object oriented, the functional, and the concurrent-declarative programming paradigms. Our measurements show that the distribution middleware is competitive with the most popular distributed programming systems (JavaRMI, .NET, IBM CORBA).</p>

Clusters, i.e., several computers interconnected with acommunication network, provide a cost efficient way to achievehigh performance. Messages are the natural way of communicationin this kind of systems. However, it is widely argued thatusing a shared memory programming model reduces the programmingeffort. Hence it is interestingto investigate systems thatprovide shared memory on clusters. This thesis describes some performance aspects of providingsuch a shared memory using software. The systems that provide ashared memory in software are commonly called softwaredistributed shared memory systems, software DSM systems. Thethesis consists of seven papers that each describe differentaspects of software DSM systems. One of the main performance bottlenecks is the communicationnetwork and three papers in the thesis investigates thisbottleneck. One paper analyzes the traffic sent on the networkwhen applications are executed. In another paper a latencyhiding technique is described and evaluated that substantiallyincreases the performance of iterative applications, i.e.,applications with loops. The last paper investigates the use ofpriorities to reduce the latency of certain messages used bythe software DSM system. One paper is devoted to discussing how a portable softwareDSM system should be designed and this paper forms the basisfor the remaining three papers. These papers also present acompiler and run-time library for OpenMP which is a recentindustry shared memory programming model. The compiler and therun-time library were developed as part of the thesis. One ofthe three papers describes a prototype system for runningOpenMP applications on clusters. The second presents thecompiler and compares the performance of applications compiledwith the compiler with that of the applications compiled with acommercial compiler. The last paper describes the OpenMPrun-time library.

This dissertation provides constructive proof, through the implementation of a middleware, that distribution transparency is practical, generic, and extensible. Fault tolerant distributed services can be developed by using the failure detection abilities of the middleware. By generic we mean that the middleware can be used for many different programming languages and paradigms. Distribution for each kind of language entity is done in terms of consistency protocols, which guarantee that the semantics of the entities are preserved in a distributed setting. The middleware allows new consistency protocols to be added easily. The efficiency of the middleware and the ease of integration are shown by coupling the middleware to a programming system, which encompasses the object oriented, the functional, and the concurrent-declarative programming paradigms. Our measurements show that the distribution middleware is competitive with the most popular distributed programming systems (JavaRMI, .NET, IBM CORBA). / QC 20101029

<p>Clusters, i.e., several computers interconnected with acommunication network, provide a cost efficient way to achievehigh performance. Messages are the natural way of communicationin this kind of systems. However, it is widely argued thatusing a shared memory programming model reduces the programmingeffort. Hence it is interestingto investigate systems thatprovide shared memory on clusters.</p><p>This thesis describes some performance aspects of providingsuch a shared memory using software. The systems that provide ashared memory in software are commonly called softwaredistributed shared memory systems, software DSM systems. Thethesis consists of seven papers that each describe differentaspects of software DSM systems.</p><p>One of the main performance bottlenecks is the communicationnetwork and three papers in the thesis investigates thisbottleneck. One paper analyzes the traffic sent on the networkwhen applications are executed. In another paper a latencyhiding technique is described and evaluated that substantiallyincreases the performance of iterative applications, i.e.,applications with loops. The last paper investigates the use ofpriorities to reduce the latency of certain messages used bythe software DSM system.</p><p>One paper is devoted to discussing how a portable softwareDSM system should be designed and this paper forms the basisfor the remaining three papers. These papers also present acompiler and run-time library for OpenMP which is a recentindustry shared memory programming model. The compiler and therun-time library were developed as part of the thesis. One ofthe three papers describes a prototype system for runningOpenMP applications on clusters. The second presents thecompiler and compares the performance of applications compiledwith the compiler with that of the applications compiled with acommercial compiler. The last paper describes the OpenMPrun-time library.</p>

Made available in DSpace on 2014-06-11T19:32:51Z (GMT). No. of bitstreams: 0 Previous issue date: 2013-03-22Bitstream added on 2014-06-13T20:27:52Z : No. of bitstreams: 1 affonso_co_dr_guara.pdf: 1152405 bytes, checksum: 8f0081a68fda3c66a6f0ef4bcd0e9cac (MD5) / Para se obter produtos e serviços que atendam ao nível de produtividade exigida pelo mercado, deve-se otimizar vários fatores determinantes na usinagem da madeira. O atual objetivo da pesquisa em Inteligência Artificial dedica-se ao desenvolvimento de sistemas inteligentes flexíveis e auto ajustáveis, com vistas à diminuição da presença de operadores humanos, de forma que o controle destes processos seja realizado através de sistemas computacionais. A usinagem da madeira se caracteriza pela ação de vários agentes, que de forma geral, são muito complexos para serem representados de forma analítica, adicionalmente as respostas destes sistemas são não-lineares. Portanto, estas dificuldades na modelagem do processamento da madeira justificam a utilização de redes neurais como ferramenta para melhoria de processo, e consequente agregação de valor ao produto final. O objetivo deste trabalho foi utilizar a capacidade de aprendizagem e a generalização das redes neurais e outras técnicas de inteligência computacional no processamento de madeira. A metodologia utilizada consistiu em utilizar redes neurais do tipo Multilayer Perceptrons (MLP) associadas à Lógica Fuzzy para construção de controlador do processo de usinagem da madeira. Adicionalmente as redes neurais realizaram uma classificação de imagens com relação aos defeitos superficiais da madeira. Foi utilizadas bases de dados obtidas através dos processos reais de usinagem da madeira. Os resultados obtidos foram satisfatórios, o que confirma que as redes neurais foram uma... / In order to obtain products and services to exceed the level of productivity required by the market, many machining wood factors should be optimized. The current goal of research in Artificial Intelligence is dedicated to develop intelligent flexible systems, self-adjusting, to decrease the presence of human operators. The control of these processes is done through the help of computer systems composed from software and hardware. The modern industrial processes are characterized by the action of various agents that are generally too complex to be represented analytically, additionally answers these systems are non-linear. Therefore, these difficulties in modeling wood machining process justify the use of Neural Network as a tool for process improvement and to add value to the final product. Computational Intelligence techniques such as Neuro-Fuzzy Networks have been proved applicable to this problem, since they combine the ability to learn from examples and to generalize the information learned from the neural network with the ability of Fuzzy Logic to turn variables into linguistic rules. The objective of this work is to use the learning ability and generalization of neural networks and other techniques of Artificial Intelligence in machining materials, which have solid non-linear character. The results were satisfactory, thus confirming the neural... (Complete abstract click electronic access below)

Orientador: Marcos Tadeu Tiburcio Gonçalves / Banca: Manoel Cleber de Sampaio Alves / Banca: Maria Angelica Martins Costa / Banca: Raquel Gonçalves / Banca: Ivaldo de Domenico Valarelli / Resumo: Para se obter produtos e serviços que atendam ao nível de produtividade exigida pelo mercado, deve-se otimizar vários fatores determinantes na usinagem da madeira. O atual objetivo da pesquisa em Inteligência Artificial dedica-se ao desenvolvimento de sistemas inteligentes flexíveis e auto ajustáveis, com vistas à diminuição da presença de operadores humanos, de forma que o controle destes processos seja realizado através de sistemas computacionais. A usinagem da madeira se caracteriza pela ação de vários agentes, que de forma geral, são muito complexos para serem representados de forma analítica, adicionalmente as respostas destes sistemas são não-lineares. Portanto, estas dificuldades na modelagem do processamento da madeira justificam a utilização de redes neurais como ferramenta para melhoria de processo, e consequente agregação de valor ao produto final. O objetivo deste trabalho foi utilizar a capacidade de aprendizagem e a generalização das redes neurais e outras técnicas de inteligência computacional no processamento de madeira. A metodologia utilizada consistiu em utilizar redes neurais do tipo Multilayer Perceptrons (MLP) associadas à Lógica Fuzzy para construção de controlador do processo de usinagem da madeira. Adicionalmente as redes neurais realizaram uma classificação de imagens com relação aos defeitos superficiais da madeira. Foi utilizadas bases de dados obtidas através dos processos reais de usinagem da madeira. Os resultados obtidos foram satisfatórios, o que confirma que as redes neurais foram uma... (Resumo completo, clicar acesso eletrônico abaixo) / Abstract: In order to obtain products and services to exceed the level of productivity required by the market, many machining wood factors should be optimized. The current goal of research in Artificial Intelligence is dedicated to develop intelligent flexible systems, self-adjusting, to decrease the presence of human operators. The control of these processes is done through the help of computer systems composed from software and hardware. The modern industrial processes are characterized by the action of various agents that are generally too complex to be represented analytically, additionally answers these systems are non-linear. Therefore, these difficulties in modeling wood machining process justify the use of Neural Network as a tool for process improvement and to add value to the final product. Computational Intelligence techniques such as Neuro-Fuzzy Networks have been proved applicable to this problem, since they combine the ability to learn from examples and to generalize the information learned from the neural network with the ability of Fuzzy Logic to turn variables into linguistic rules. The objective of this work is to use the learning ability and generalization of neural networks and other techniques of Artificial Intelligence in machining materials, which have solid non-linear character. The results were satisfactory, thus confirming the neural... (Complete abstract click electronic access below) / Doutor