Thèse diffusée initialement dans le cadre d'un projet pilote des Presses de l'Université de Montréal/Centre d'édition numérique UdeM (1997-2008) avec l'autorisation de l'auteur.

Cette thèse porte premièrement sur la reconfiguration dynamique et la simulation hétérogène dans les Réseaux des Capteurs sans Fil. Ces réseaux sont constitués d'une multitude de systèmes électroniques communicants par radio-fréquence, très contraints en énergie. La partie de communication radio entre ces nœuds est la plus consommatrice. C'est pourquoi la minimisation du temps effectif est désirée. On a implémenté une solution qui consiste à envoyer au nœud un fichier de reconfiguration codé utilisant un langage de programmation haut niveau (MinTax). Le nœud sera capable de compiler ce fichier et générer le code object associé à son architecture, in-situ. Grâce au caractère abstrait du MinTax, plusieurs architectures matérielles et systèmes d'exploitation sont visés. Dans un deuxième temps, ce travail de thèse est lié au simulateur de réseaux de capteurs IDEA1TLM.IDEA1TLM permet de prédire quels circuits et configurations sont les plus adéquats à une application sans fil donnée. Ce simulateur a été amélioré pour permettre la simulation rapide des systèmes électroniques matériellement différents dans le même réseau ainsi que le logiciel présent sur les noeuds. Mots clés : Reconfiguration dynamique, Compilation in-situ, MinTax, Hétérogénéité, IDEA1TLM.

As transistors size and power limitations stroke computer industry, hardware parallelism arose as the solution, bringing old forgotten problems back into equation to solve the existing limitations of current parallel technologies. Compilers regain focus by being the most relevant puzzle piece in the quest for the expected computer performance improvements predicted by Moores law no longer possible without parallelism. Parallel research is mainly focused in either the language or architectural aspects, not really giving the needed attention to compiler problems, being the reason for the weak compiler support by many parallel languages or architectures, not allowing to exploit performance to the best. This thesis addresses these problems by presenting: Erbium, a low level streaming data-flow language supporting multiple producer and consumer task communication; a very efficient runtime implementation for x86 architectures also addressing other types of architectures; a compiler integration of the language as an intermediate representation in GCC; a study of the language primitives dependencies, allowing compilers to further optimise the Erbium code not only through specific parallel optimisations but also through traditional compiler optimisations, such as partial redundancy elimination and dead code elimination.

Th. doct.--Informatique--Paris 6, 1994.

Nowadays, most computer systems are equipped with powerful parallel devices such as Graphics Processing Units (GPUs). They are present in almost every computer system including mobile devices, tablets, desktop computers and servers. These parallel systems have unlocked the possibility for many scientists and companies to process significant amounts of data in shorter time. But the usage of these parallel systems is very challenging due to their programming complexity. The most common programming languages for GPUs, such as OpenCL and CUDA, are created for expert programmers, where developers are required to know hardware details to use GPUs. However, many users of heterogeneous and parallel hardware, such as economists, biologists, physicists or psychologists, are not necessarily expert GPU programmers. They have the need to speed up their applications, which are often written in high-level and dynamic programming languages, such as Java, R or Python. Little work has been done to generate GPU code automatically from these high-level interpreted and dynamic programming languages. This thesis presents a combination of a programming interface and a set of compiler techniques which enable an automatic translation of a subset of Java and R programs into OpenCL to execute on a GPU. The goal is to reduce the programmability and usability gaps between interpreted programming languages and GPUs. The first contribution is an Application Programming Interface (API) for programming heterogeneous and multi-core systems. This API combines ideas from functional programming and algorithmic skeletons to compose and reuse parallel operations. The second contribution is a new OpenCL Just-In-Time (JIT) compiler that automatically translates a subset of the Java bytecode to GPU code. This is combined with a new runtime system that optimises the data management and avoids data transformations between Java and OpenCL. This OpenCL framework and the runtime system achieve speedups of up to 645x compared to Java within 23% slowdown compared to the handwritten native OpenCL code. The third contribution is a new OpenCL JIT compiler for dynamic and interpreted programming languages. While the R language is used in this thesis, the developed techniques are generic for dynamic languages. This JIT compiler uniquely combines a set of existing compiler techniques, such as specialisation and partial evaluation, for OpenCL compilation together with an optimising runtime that compile and execute R code on GPUs. This JIT compiler for the R language achieves speedups of up to 1300x compared to GNU-R and 1.8x slowdown compared to native OpenCL.

A presente pesquisa compreende o estudo e a tradução para o português do prólogo e da retórica do Tesouro de Brunetto Latini. Trata-se de uma compilação escrita em francês, por volta de 1260, que reúne matérias úteis para a política, entendida como arte de governar uma cidade. O Tesouro está dividido em três livros, um sobre teórica, um sobre ética e outro sobre política, o qual inclui uma arte retórica e um manual do podestade. O estudo do prólogo mostrou que a política é o fio condutor deste texto no gênero humilde que, para ensinar, compila excertos de diversas autoridades de maneira ordenada. Por sua vez, a arte retórica, parte da política, segue de perto a matéria do tratado De inventione, de Cícero, e traz também preceitos para a escrita de cartas, preceitos de artes poéticas e exemplos históricos e bíblicos. Por seus preceitos e exemplos, que são tanto retóricos quanto éticos, o Tesouro ensina e faz o elogio da política, ao mesmo tempo que se apresenta como uma memória. / This study encompasses an analysis and a Portuguese translation of the prologue and the rhetoric of Brunetto Latini\'s Tresor. The text is a French compilation written around 1260, and it compiles materials useful to the science of politics, which is defined as the art of ruling a city. The Tresor is divided into three books, one about theorics, one about ethics, and the other about politics, which includes an art of rhetoric and a handbook for the podestà. The analysis of the prologue shows that politics are the guiding principle of the text, which is written in plain (attenuata) style, thus meant for instruction, and compiles and puts in a specific order excerpts taken from various authorities. As for the art of rhetoric, a part of politics, the text closely follows the subject matter of Cicero\'s De intentione, and also has precepts for letter writing, principles of poetry and historical and biblical examples. Through the precepts and examples, which are rhetoric as well as ethic, the Tresor teaches and praises politics, and at the same time operates as a memory.

This thesis concentrates on pumping and a usage of the EAFRD resources, on conditions that are the applicants asked to fulfil in order to use these dotation resources. A practical example is a compilation of a project for the agricultural cooperation Hluboké Mašůvky.

Uma das metas da Tabela Brasileira de Composição de Alimentos (TBCA-USP) é a disseminação de dados confiáveis de composição química e, por isso, a base de dados da TBCA-USP vem sendo continuamente atualizada com a inserção de novos alimentos e nutrientes, como os minerais. Embora nos últimos 20 anos tenha ocorrido uma intensificação dos estudos de micronutrientes, há pouca informação sobre o conteúdo de minerais nas tabelas de composição de alimentos, empregando metodologia adequada e, paralelamente, parte das informações existentes está dispersa em publicações. O objetivo deste trabalho consiste na compilação de dados de minerais de alimentos nacionais, visando sua introdução na TBCA-USP. Para a compilação de minerais, foi necessário atualizar o Manual e o Formulário de Compilação de Dados sobre Composição de Alimentos; no manual foram incluídas informações importantes a serem observadas durante a compilação de dados de minerais. Já no formulário, na planilha de avaliação da qualidade dos dados, incluíram-se campos relacionados ao tratamento das amostras e que são necessários ao conhecimento do compilador, uma vez que podem influenciar na quantidade de minerais do alimento, permitindo assim, uma melhor avaliação da confiabilidade da informação. O levantamento de dados foi realizado em periódicos nacionais e internacionais, dissertações, teses e dados internos de laboratório, totalizando 348 publicações com datas entre 1975 e 2011. Desse total, 186 foram descartadas e 162 foram selecionadas para compilação, pois continham informações detalhadas sobre o alimento estudado, com descrição e referência dos métodos analíticos utilizados e informações completas dos resultados analíticos. Das publicações compiladas, 26% foram referentes ao grupo das frutas e derivados; 25% ao grupo das hortaliças, algas, cogumelos, condimentos, espécies e derivados; 12,5% ao grupo de leguminosas, grãos e derivados; 10% ao grupo das carnes e derivados; 6,5% ao grupo dos peixes e frutos do mar; 6% ao grupo das bebidas; 4,5% ao grupo de leite e derivados; 4% ao grupo dos cereais e derivados; 2,5% ao grupo de alimentos manufaturados; 1% ao grupo dos ovos e derivados; 1% ao grupo de alimentos infantis; 0,5% ao grupo dos óleos e derivados e 0,5% ao grupo de produtos açucarados, totalizando 860 alimentos. Os minerais mais presentes nas publicações foram o ferro (76% das publicações), o cálcio (72%) e o zinco (66%). O Manual e o Formulário foram aprimorados visando facilitar a compilação de dados de minerais e o melhor detalhamento das informações disponibilizadas nas publicações. Os dados dos 860 alimentos compilados geraram o banco de dados de minerais da Tabela, e em breve serão disponibilizados na TBCA-USP (www.fcf.usp/tabela). / One of the goals of the Brazilian Food Composition Database - USP (TBCA-USP) is the dissemination of reliable data of chemical composition and, because of this, the TBCA-USP database is continually being updated with the insertion of new foods and nutrients, like minerals. Although in the last 20 years has occurred intensification in the micronutrients studies, there is little information about the content of minerals in food composition tables, employing adequate methodology and, in parallel, part of the existing information is dispersed in publications. The objective of this study is the compilation of mineral data in national foods, aiming your introduction in TBCA-USP. For the mineral compilation, it was necessary to update the manual and the form for compilation of food composition data; in the manual were included important informations to be observed during the compilation. In the case of form, the spreadsheet for analytical quality control were expanded to accommodate informations about the treatment of the samples as because they are necessary to the knowledge of the compiler, since this information can affect the minerals content of the food, resulting in a better avaliation of the confiability of information. The survey of mineral data was done in national and international publications, thesis and internal data of laboratories, resulted in 348 publications with date of publication between 1975 and 2011. Of this total, 186 were discarded and 162 were evaluated for compilation because they had detailed information about the food, with description and reference of the analytical methods used and complete information about the results. Of the compiled publications, 26% were related to fruits and fruit products group; 25% to vegetable and vegetable products; 12,5% to legumes, grains and their products; 10% to meat and meat products; 6,5% to fish and shellfish; 6% to beverages; 4,5% to milk and milk products; 4% to cereals and grain products; 2,5% to manufactured foods; 1% to eggs; 1% to infant foods; 0,5% to oils and fats and 0,5% to sugars and syrups, resulted in 860 foods items. The minerals that appear more times in the publications were iron (76% of the publications), calcium (72%) and zinc (66%). The Manual and the Form have been improved aiming to facilitate the compilation of mineral data and better detail of avaliable information in the publications. The data of 860 compiled foods resulted the mineral database of the TBCA-USP, and ready to soon be available in TBCA-USP (www.fcf.usp/tabela).

Advanced development environments are essential for efficient realization of complex industrial products. Powerful equation-based object-oriented (EOO) languages such as Modelica are successfully used for modeling and virtual prototyping complex physical systems and components. The Modelica language enables engineers to build large, sophisticated and complex models. Modelica environments should scale up and be able to handle these large models. This thesis addresses the scalability of Modelica tools by employing incremental compilation and dynamic loading. The design, implementation and evaluation of this approach is presented. OpenModelica is an open-source Modelica environment developed at PELAB in which we have implemented our strategy for incremental compilation and dynamic loading of functions. We have tested the performance of these strategies in a number of different scenarios in order to see how much of an impact they have on the compilation and execution time. Our solution contains an overhead of one or two hash calls during runtime as it uses dynamic hashes instead of static arrays.

In this report we evaluate different techniques for increasing the performance of an interpreted processing language in a telecommunication system, called Billing Gateway R8. We have implemented a prototype in which we first translate the language into C++ code, and then compile it using a C++ compiler. In our prototype we experienced a threefold increase in processing throughput, compared to the original system, when running on a Symmetric Multi Processor with four CPU:s that were under full load. The prototype also showed better scalability than Billing Gateway R8, due to less use of dynamic memory management.