Optimizing database systems to achieve the maximum attainable throughput of the underlying hardware is one of the many difficult tasks that face Database Administrators. With the increased use of database systems in many environments, this task has even become more difficult. One of the parameters that needs to be configured is the number of worker tasks that the database server uses (the multiprogramming level). This thesis will focus on how to automatically adjust the number of database server worker tasks to achieve maximum throughput under varying workload characteristics. The underlying intuition is that every workload has an optimal multiprogramming level that can achieve the best throughput given the workload characteristic.

Optimizing database systems to achieve the maximum attainable throughput of the underlying hardware is one of the many difficult tasks that face Database Administrators. With the increased use of database systems in many environments, this task has even become more difficult. One of the parameters that needs to be configured is the number of worker tasks that the database server uses (the multiprogramming level). This thesis will focus on how to automatically adjust the number of database server worker tasks to achieve maximum throughput under varying workload characteristics. The underlying intuition is that every workload has an optimal multiprogramming level that can achieve the best throughput given the workload characteristic.

O desenvolvimento de aplicações de forma a atingir níveis de desempenho próximos aos níveis teóricos de uma determinada plataforma é uma tarefa que exige conhecimento técnico do ambiente de hardware, uma vez que o software deve explorar detalhes específicos da plataforma em questão. Pelo fato do software ser específico à plataforma, caso ela evolua ou se altere, as otimizações realizadas podem não explorar a nova arquitetura de forma eficiente. Auto-tuners são sistemas que surgiram como um meio automatizado de adaptar um determinado software a uma arquitetura alvo. Essa adaptação ocorre através de uma busca empírica de valores ótimos para parâmetros específicos de uma aplicação, a fim de ajustá-los às características do hardware, ou ainda através da geração de códigofonte otimizado para a plataforma. Este trabalho propõe um módulo auto-tuner orientado à adaptação parametrizada de uma aplicação paralela, que trabalha variando os fatores da dimensão do domínio bidimensional, o número de processos e a extensão das regiões de sobreposição. Para cada variação dos fatores, o auto-tuner testa a aplicação na arquitetura paralela de forma a buscar a combinação de parâmetros com melhor desempenho. Para possibilitar o auto-tuning, foi desenvolvida uma classe em linguagem C++ denominada Mesh, baseada no padrão MPI. A classe busca abstrair a decomposição de domínios de uma aplicação paralela por meio do uso de Orientação a Objetos, e facilita a variação da extensão das regiões de sobreposição entre os subdomínios. Os resultados experimentais demonstraram que o auto-tuner explora o ganho de desempenho pela variação do número de processos da aplicação, que também é tratado pelo módulo auto-tuner. A arquitetura paralela utilizada na validação não se mostrou ideal para uma otimização através do aumento da extensão das regiões sobrepostas entre subdomínios. / Achieving the peak performance level of a particular platform requires technical knowledge of the hardware environment involved, since the software must explore specific details inherent to the hardware. Once the software is optimized for a target platform, if the hardware evolves or is changed, the software probably would not be as efficient in the new environment. This performance portability problem is addressed by software auto-tuning, which emerged in the past decade as an automated technique to adapt a particular software to an underlying hardware. The software adaptation is performed by an auto-tuner. The auto-tuner is an entity that empirically adjusts specific application parameters in order to improve the overall application performance, or even generates source-code optimized for the target platform. This dissertation proposes an auto-tuner to optimize the domain decomposition of a parallel application that performs stencil computations. The proposed auto-tuner works in a parameterized adaptation fashion, and varies the dimensions of a 2D domain, the number of parallel processes and the extension of the overlapping zones between subdomains. For each combination of parameter values, the auto-tuner probes the application in the parallel architecture in order to seek the best combination of values. In order to make auto-tuning possible, it is proposed a C++ class called Mesh, based on the Message Passing Interface (MPI) standard. The role of this class is to abstract the domain decomposition from the application using the Object Orientation facilities provided by C++, and also to enable the extension of the overlapping zones between subdomain. The experimental results showed that the performance gains were mainly due to the variation of the number of processes, which was one of the application factors dealt by the auto-tuner. The parallel architecture used in the experiments showed itself as not adequate for optimizing the domain decomposition by increasing the overlapping zones extension.

ABSTRACT In this dissertation, several adaptive control design schemes for a class of nonlinear systems are proposed. The first topic of the research is concerned with self-tuning PID controller design. The main problem of designing PID controller is how to determine the values of three control gains, i.e., proportional gain , integral gain , and derivative gain . We attempt to use the technique of adaptive control based on the Lyapunov approach to design the PID controller for some class of partially known nonlinear systems. Three PID control gains are adjusted on-line such that better output performance can be achieved. The stability of the closed-loop PID control systems is analyzed and guaranteed by introducing a supervisory control and a modified adaptation law with projection. Second, two kinds of adaptive neural control systems including the direct and indirect neural controls are considered by using simple single auto-tuning neuron. We will first propose a novel neuron called auto-tuning neuron and use it to take place of the roles of the traditional neural networks used in the direct and indirect adaptive neural control systems. This can greatly reduce the computational time and network complexities due to the simple configuration of the auto-tuning neuron. It is also easy for hardware implementation. Third, based on the idea borrowed from natural evolution, genetic algorithm can search for optimal or near-optimal solutions for an optimization problem over the search domain. An optimization technique of real-coded genetic algorithm is used to design the PID controller by minimizing the performance index of integrated absolute error. The improvements of our results over that using other methods are also illustrated. In the last part of each section, some computer simulation results will also be provided to illustrate our proposed methods.

O desenvolvimento de aplicações de forma a atingir níveis de desempenho próximos aos níveis teóricos de uma determinada plataforma é uma tarefa que exige conhecimento técnico do ambiente de hardware, uma vez que o software deve explorar detalhes específicos da plataforma em questão. Pelo fato do software ser específico à plataforma, caso ela evolua ou se altere, as otimizações realizadas podem não explorar a nova arquitetura de forma eficiente. Auto-tuners são sistemas que surgiram como um meio automatizado de adaptar um determinado software a uma arquitetura alvo. Essa adaptação ocorre através de uma busca empírica de valores ótimos para parâmetros específicos de uma aplicação, a fim de ajustá-los às características do hardware, ou ainda através da geração de códigofonte otimizado para a plataforma. Este trabalho propõe um módulo auto-tuner orientado à adaptação parametrizada de uma aplicação paralela, que trabalha variando os fatores da dimensão do domínio bidimensional, o número de processos e a extensão das regiões de sobreposição. Para cada variação dos fatores, o auto-tuner testa a aplicação na arquitetura paralela de forma a buscar a combinação de parâmetros com melhor desempenho. Para possibilitar o auto-tuning, foi desenvolvida uma classe em linguagem C++ denominada Mesh, baseada no padrão MPI. A classe busca abstrair a decomposição de domínios de uma aplicação paralela por meio do uso de Orientação a Objetos, e facilita a variação da extensão das regiões de sobreposição entre os subdomínios. Os resultados experimentais demonstraram que o auto-tuner explora o ganho de desempenho pela variação do número de processos da aplicação, que também é tratado pelo módulo auto-tuner. A arquitetura paralela utilizada na validação não se mostrou ideal para uma otimização através do aumento da extensão das regiões sobrepostas entre subdomínios. / Achieving the peak performance level of a particular platform requires technical knowledge of the hardware environment involved, since the software must explore specific details inherent to the hardware. Once the software is optimized for a target platform, if the hardware evolves or is changed, the software probably would not be as efficient in the new environment. This performance portability problem is addressed by software auto-tuning, which emerged in the past decade as an automated technique to adapt a particular software to an underlying hardware. The software adaptation is performed by an auto-tuner. The auto-tuner is an entity that empirically adjusts specific application parameters in order to improve the overall application performance, or even generates source-code optimized for the target platform. This dissertation proposes an auto-tuner to optimize the domain decomposition of a parallel application that performs stencil computations. The proposed auto-tuner works in a parameterized adaptation fashion, and varies the dimensions of a 2D domain, the number of parallel processes and the extension of the overlapping zones between subdomains. For each combination of parameter values, the auto-tuner probes the application in the parallel architecture in order to seek the best combination of values. In order to make auto-tuning possible, it is proposed a C++ class called Mesh, based on the Message Passing Interface (MPI) standard. The role of this class is to abstract the domain decomposition from the application using the Object Orientation facilities provided by C++, and also to enable the extension of the overlapping zones between subdomain. The experimental results showed that the performance gains were mainly due to the variation of the number of processes, which was one of the application factors dealt by the auto-tuner. The parallel architecture used in the experiments showed itself as not adequate for optimizing the domain decomposition by increasing the overlapping zones extension.

O desenvolvimento de aplicações de forma a atingir níveis de desempenho próximos aos níveis teóricos de uma determinada plataforma é uma tarefa que exige conhecimento técnico do ambiente de hardware, uma vez que o software deve explorar detalhes específicos da plataforma em questão. Pelo fato do software ser específico à plataforma, caso ela evolua ou se altere, as otimizações realizadas podem não explorar a nova arquitetura de forma eficiente. Auto-tuners são sistemas que surgiram como um meio automatizado de adaptar um determinado software a uma arquitetura alvo. Essa adaptação ocorre através de uma busca empírica de valores ótimos para parâmetros específicos de uma aplicação, a fim de ajustá-los às características do hardware, ou ainda através da geração de códigofonte otimizado para a plataforma. Este trabalho propõe um módulo auto-tuner orientado à adaptação parametrizada de uma aplicação paralela, que trabalha variando os fatores da dimensão do domínio bidimensional, o número de processos e a extensão das regiões de sobreposição. Para cada variação dos fatores, o auto-tuner testa a aplicação na arquitetura paralela de forma a buscar a combinação de parâmetros com melhor desempenho. Para possibilitar o auto-tuning, foi desenvolvida uma classe em linguagem C++ denominada Mesh, baseada no padrão MPI. A classe busca abstrair a decomposição de domínios de uma aplicação paralela por meio do uso de Orientação a Objetos, e facilita a variação da extensão das regiões de sobreposição entre os subdomínios. Os resultados experimentais demonstraram que o auto-tuner explora o ganho de desempenho pela variação do número de processos da aplicação, que também é tratado pelo módulo auto-tuner. A arquitetura paralela utilizada na validação não se mostrou ideal para uma otimização através do aumento da extensão das regiões sobrepostas entre subdomínios. / Achieving the peak performance level of a particular platform requires technical knowledge of the hardware environment involved, since the software must explore specific details inherent to the hardware. Once the software is optimized for a target platform, if the hardware evolves or is changed, the software probably would not be as efficient in the new environment. This performance portability problem is addressed by software auto-tuning, which emerged in the past decade as an automated technique to adapt a particular software to an underlying hardware. The software adaptation is performed by an auto-tuner. The auto-tuner is an entity that empirically adjusts specific application parameters in order to improve the overall application performance, or even generates source-code optimized for the target platform. This dissertation proposes an auto-tuner to optimize the domain decomposition of a parallel application that performs stencil computations. The proposed auto-tuner works in a parameterized adaptation fashion, and varies the dimensions of a 2D domain, the number of parallel processes and the extension of the overlapping zones between subdomains. For each combination of parameter values, the auto-tuner probes the application in the parallel architecture in order to seek the best combination of values. In order to make auto-tuning possible, it is proposed a C++ class called Mesh, based on the Message Passing Interface (MPI) standard. The role of this class is to abstract the domain decomposition from the application using the Object Orientation facilities provided by C++, and also to enable the extension of the overlapping zones between subdomain. The experimental results showed that the performance gains were mainly due to the variation of the number of processes, which was one of the application factors dealt by the auto-tuner. The parallel architecture used in the experiments showed itself as not adequate for optimizing the domain decomposition by increasing the overlapping zones extension.

We optimize parameters in OpenACC clauses for a stencil evaluation kernel executed on Graphical Processing Units (GPUs) using a variety of machine learning and optimization search algorithms, individually and in hybrid combinations, and compare execution time performance to the best possible obtained from brute force search. Several auto-tuning techniques – historic learning, random walk, simulated annealing, Nelder-Mead, and genetic algorithms – are evaluated over a large two-dimensional parameter space not satisfactorily addressed to date by OpenACC compilers, consisting of gang size and vector length. A hybrid of historic learning and Nelder-Mead delivers the best balance of high performance and low tuning effort. GPUs are employed over an increasing range of applications due to the performance available from their large number of cores, as well as their energy efficiency. However, writing code that takes advantage of their massive fine-grained parallelism requires deep knowledge of the hardware, and is generally a complex task involving program transformation and the selection of many parameters. To improve programmer productivity, the directive-based programming model OpenACC was announced as an industry standard in 2011. Various compilers have been developed to support this model, the most notable being those by Cray, CAPS, and PGI. While the architecture and number of cores have evolved rapidly, the compilers have failed to keep up at configuring the parallel program to run most e ciently on the hardware. Following successful approaches to obtain high performance in kernels for cache-based processors using auto-tuning, we approach this compiler-hardware gap in GPUs by employing auto-tuning for the key parameters “gang” and “vector” in OpenACC clauses. We demonstrate results for a stencil evaluation kernel typical of seismic imaging over a variety of realistically sized three-dimensional grid configurations, with different truncation error orders in the spatial dimensions. Apart from random walk and historic learning based on nearest neighbor in grid size, most of our heuristics, including the one that proves best, appear to be applied in this context for the first time. This work is a stepping-stone towards an OpenACC auto-tuning framework for more general high-performance numerical kernels optimized for GPU computations.

Systematic investigation of the Friction Stir Welding (FSW) process shows that a fixed rotational velocity and feed rate may not yield uniform mechanical properties along the length of a weldment. Nevertheless, correlations between process parameters and post-weld material properties have successfully demonstrated that peak temperature and cooling rate drive post-weld properties. There have been many reported methodologies for controlling friction stir welding, with varying degrees of cost to implement and effectiveness. However, comparing data from uncontrolled FSW of AA 6111-T4 sheet with controlled FSW at temperatures ranging from 375 °C to 450 °C demonstrates that a simplified methodology of a single-loop PID controlling with spindle speed may be used to effectively control temperature. This methodology can be simply used with any machine that already has the ability to actively control spindle speed, and has been previously shown to be able to be auto-tuned with a single weld. Additionally, implementation of this method compared to uncontrolled FSW in AA6111 at linear weld speeds of 1-2 meters per minute showed improved mechanical properties and greater consistency in properties along the length of the weld under temperature control. Further results indicate that a minimum spindle rpm may exist above which tensile specimens did not fracture within the weld centerline, regardless of temperature. This work demonstrates that a straight-forward, PID-based implementation of temperature control at high weld rates can produce high quality welds with auto-tuned gains. This method also shows promise in application to other processes in the Friction Stir family, and preliminary results in an application to the Additive Friction Stir Deposition (AFSD) process are also presented.

Les progrès en termes de puissance de calcul ont entraîné de nombreuses évolutions dans le domaine de la science et de ses applications. La résolution de systèmes linéaires survient fréquemment dans le calcul scientifique, comme par exemple lors de la résolution d'équations aux dérivées partielles par la méthode des éléments finis. Le temps de résolution découle alors directement des performances des opérations algébriques mises en jeu.Cette thèse a pour but de développer des algorithmes parallèles innovants pour la résolution de systèmes linéaires creux de grandes tailles. Nous étudions et proposons comment calculer efficacement les opérations d'algèbre linéaire sur plateformes de calcul multi-coeur hétérogènes-GPU afin d'optimiser et de rendre robuste la résolution de ces systèmes. Nous proposons de nouvelles techniques d'accélération basées sur la distribution automatique (auto-tuning) des threads sur la grille GPU suivant les caractéristiques du problème et le niveau d'équipement de la carte graphique, ainsi que les ressources disponibles. Les expérimentations numériques effectuées sur un large spectre de matrices issues de divers problèmes scientifiques, ont clairement montré l'intérêt de l'utilisation de la technologie GPU, et sa robustesse comparée aux bibliothèques existantes comme Cusp.L'objectif principal de l'utilisation du GPU est d'accélérer la résolution d'un problème dans un environnement parallèle multi-coeur, c'est-à-dire "Combien de temps faut-il pour résoudre le problème?". Dans cette thèse, nous nous sommes également intéressés à une autre question concernant la consommation énergétique, c'est-à-dire "Quelle quantité d'énergie est consommée par l'application?". Pour répondre à cette seconde question, un protocole expérimental est établi pour mesurer la consommation d'énergie d'un GPU avec précision pour les opérations fondamentales d'algèbre linéaire. Cette méthodologie favorise une "nouvelle vision du calcul haute performance" et apporte des réponses à certaines questions rencontrées dans l'informatique verte ("green computing") lorsque l'on s'intéresse à l'utilisation de processeurs graphiques.Le reste de cette thèse est consacré aux algorithmes itératifs synchrones et asynchrones pour résoudre ces problèmes dans un contexte de calcul hétérogène multi-coeur-GPU. Nous avons mis en application et analysé ces algorithmes à l'aide des méthodes itératives basées sur les techniques de sous-structurations. Dans notre étude, nous présentons les modèles mathématiques et les résultats de convergence des algorithmes synchrones et asynchrones. La démonstration de la convergence asynchrone des méthodes de sous-structurations est présentée. Ensuite, nous analysons ces méthodes dans un contexte hybride multi-coeur-GPU, qui devrait ouvrir la voie vers les méthodes hybrides exaflopiques.Enfin, nous modifions la méthode de Schwarz sans recouvrement pour l'accélérer à l'aide des processeurs graphiques. La mise en oeuvre repose sur l'accélération par les GPUs de la résolution locale des sous-systèmes linéaires associés à chaque sous-domaine. Pour améliorer les performances de la méthode de Schwarz, nous avons utilisé des conditions d'interfaces optimisées obtenues par une technique stochastique basée sur la stratégie CMA-ES (Covariance Matrix Adaptation Evolution Strategy). Les résultats numériques attestent des bonnes performances, de la robustesse et de la précision des algorithmes synchrones et asynchrones pour résoudre de grands systèmes linéaires creux dans un environnement de calcul hétérogène multi-coeur-GPU. / Advances in computational power have led to many developments in science and its applications. Solving linear systems occurs frequently in scientific computing, as in the finite element discretization of partial differential equations. The running time of the overall resolution is a direct result of the performance of the involved algebraic operations.In this dissertation, different ways of efficiently solving large and sparse linear systems are put forward. We present the best way to effectively compute linear algebra operations in an heterogeneous multi-core-GPU environment in order to make solvers such as iterative methods more robust and therefore reduce the computing time of these systems. We propose new techniques to speed algorithms up the auto-tuning of the threading design, according to the problem characteristics and the equipment level in the hardware and available resources. Numerical experiments performed on a set of large-size sparse matrices arising from diverse engineering and scientific problems, have clearly shown the benefit of the use of GPU technology to solve large sparse systems of linear equations, and its robustness and accuracy compared to existing libraries such as Cusp.The main priority of the GPU program is computational time to obtain the solution in a parallel environment, i.e, "How much time is needed to solve the problem?". In this thesis, we also address another question regarding energy issues, i.e., "How much energy is consumed by the application?". To answer this question, an experimental protocol is established to measure the energy consumption of a GPU for fundamental linear algebra operations accurately. This methodology fosters a "new vision of high-performance computing" and answers some of the questions outlined in green computing when using GPUs.The remainder of this thesis is devoted to synchronous and asynchronous iterative algorithms for solving linear systems in the context of a multi-core-GPU system. We have implemented and analyzed these algorithms using iterative methods based on sub-structuring techniques. Mathematical models and convergence results of synchronous and asynchronous algorithms are presented here, as are the convergence results of the asynchronous sub-structuring methods. We then analyze these methods in the context of a hybrid multi-core-GPU, which should pave the way for exascale hybrid methods.Lastly, we modify the non-overlapping Schwarz method to accelerate it, using GPUs. The implementation is based on the acceleration of the local solutions of the linear sub-systems associated with each sub-domain using GPUs. To ensure good performance, optimized conditions obtained by a stochastic technique based on the Covariance Matrix Adaptation Evolution Strategy (CMA-ES) are used. Numerical results illustrate the good performance, robustness and accuracy of synchronous and asynchronous algorithms to solve large sparse linear systems in the context of an heterogeneous multi-core-GPU system.

De nombreuses applications en calcul scientifique traitent des matrices creuses de grande taille ayant des structures régulières ou irrégulières. Pour réduire à la fois la complexité spatiale et la complexité temporelle du traitement, ces matrices nécessitent l'utilisation d'une structure particulière de stockage (ou de compression) des données ainsi que l’utilisation d’architectures cibles parallèles/distribuées. Le choix du format de compression le plus adéquat dépend généralement de plusieurs facteurs dont, en particulier, la structure de la matrice creuse, l'architecture de la plateforme cible et la méthode numérique utilisée. Etant donné la diversité de ces facteurs, un choix optimisé pour un ensemble de données d'entrée peut induire de mauvaises performances pour un autre. D’où l’intérêt d’utiliser un système permettant la sélection automatique du meilleur format de compression (MFC) en prenant en compte ces différents facteurs. C’est dans ce cadre précis que s’inscrit cette thèse. Nous détaillons notre approche en présentant la modélisation d'un système automatique qui, étant donnée une matrice creuse, une méthode numérique, un modèle de programmation parallèle et une architecture, permet de sélectionner automatiquement le MFC. Dans une première étape, nous validons notre modélisation par une étude de cas impliquant (i) Horner, et par la suite le produit matrice-vecteur creux (PMVC), comme méthodes numériques, (ii) CSC, CSR, ELL, et COO comme formats de compression, (iii) le data parallèle comme modèle de programmation et (iv) une plateforme multicœurs comme architecture cible. Cette étude nous permet d’extraire un ensemble de métriques et paramètres qui influent sur la sélection du MFC. Nous démontrons que les métriques extraites de l'analyse du modèle data parallèle ne suffisent pas pour prendre une décision (sélection du MFC). Par conséquent, nous définissons de nouvelles métriques impliquant le nombre d'opérations effectuées par la méthode numérique et le nombre d’accès à la mémoire. Ainsi, nous proposons un processus de décision prenant en compte à la fois l'analyse du modèle data parallèle et l'analyse de l’algorithme. Dans une deuxième étape, et en se basant sur les données que nous avons extrait précédemment, nous utilisons les algorithmes du Machine Learning pour prédire le MFC d’une matrice creuse donnée. Une étude expérimentale ciblant une plateforme parallèle multicœurs et traitant des matrices creuses aléatoires et/ou provenant de problèmes réels permet de valider notre approche et d’évaluer ses performances. Comme travail futur, nous visons la validation de notre approche en utilisant d'autres plateformes parallèles telles que les GPUs. / Several applications in scientific computing deals with large sparse matrices having regular or irregular structures. In order to reduce required memory space and computing time, these matrices require the use of a particular data storage structure as well as the use of parallel/distributed target architectures. The choice of the most appropriate compression format generally depends on several factors, such as matrix structure, numerical method and target architecture. Given the diversity of these factors, an optimized choice for one input data set will likely have poor performances on another. Hence the interest of using a system allowing the automatic selection of the Optimal Compression Format (OCF) by taking into account these different factors. This thesis is written in this context. We detail our approach by presenting a design of an auto-tuner system for OCF selection. Given a sparse matrix, a numerical method, a parallel programming model and an architecture, our system can automatically select the OCF. In a first step, we validate our modeling by a case study that concerns (i) Horner scheme, and then the sparse matrix vector product (SMVP), as numerical methods, (ii) CSC, CSR, ELL, and COO as compression formats; (iii) data parallel as a programming model; and (iv) a multicore platform as target architecture. This study allows us to extract a set of metrics and parameters that affect the OCF selection. We note that data parallel metrics are not sufficient to accurately choose the most suitable format. Therefore, we define new metrics involving the number of operations and the number of indirect data access. Thus, we proposed a new decision process taking into account data parallel model analysis and algorithm analysis.In the second step, we propose to use machine learning algorithm to predict the OCF for a given sparse matrix. An experimental study using a multicore parallel platform and dealing with random and/or real-world random matrices validates our approach and evaluates its performances. As future work, we aim to validate our approach using other parallel platforms such as GPUs.