Dynamic optimizers for object-oriented languages collect a variety of profile data to drive optimization decisions. In particular, the dynamic call graph (DCG) informs key structural optimizations such as which methods to optimize and how to optimize them. Unfortunately, current low-overhead call-stack hardware and software sampling methods are subject to sampling bias, which loses accuracy of 40 to 50% when compared with a perfect call graph. This paper introduces DCG correction, a novel approach that uses static and dynamic control-flow graphs (CFGs) to improve DCG accuracy. We introduce the static frequency dominator (FDOM) relation, which extends the dominator relation on the CFG to capture relative execution frequencies and expose static constraints on DCG edges, which we use to correct DCG edge frequencies. Using conservation of flow principles, we further show how to use dynamic CFG basic block profiles to correct DCG edge frequencies intraprocedurally and interprocedurally. We implement and evaluate DCG correction in Jikes RVM on the SPEC JVM98 and DaCapo benchmarks. Default DCG sampling attains an average accuracy of 52-59% compared with perfect, whereas FDOM correction improves average accuracy to 64-68%, while adding 0.2% average overhead. The dynamic correction raises accuracy to 85% on average, while adding 1.2% average overhead. We then provide dynamically corrected DCGs to the inliner with mixed results -1% average degradations and improvements across a variety of configurations. However, prior work shows that increased DCG accuracy in production VMs has benefits. We believe that high-accuracy DCGs will become more important in the future as the complexity and modularity of object-oriented programs increases.

Dos de las limitaciones de rendimiento más importantes en los procesadores de hoy en día provienen de las operaciones de memoria y de las dependencias de control. Para resolver estos problemas, las memorias cache y los predictores de salto son dos alternativas hardware bien conocidas que explotan, entre otros factores, el reuso temporal de memoria y la correlación de saltos. En otras palabras, estas estructuras tratan de explotar la redundancia dinámica existente en los programas. Esta redundancia proviene parcialmente de la forma en que los programadores escriben código, pero también de limitaciones existentes en el modelo de compilación tradicional, lo cual introduce instrucciones de memoria y de salto innecesarias. Pensamos que los compiladores deberían ser muy agresivos optimizando programas, y por tanto ser capaces de eliminar una parte importante de esta redundancia. Por otro lado, las optimizaciones aplicadas en tiempo de enlace o directamente al programa ejecutable final han recibido una atención creciente en los últimos años, debido a limitaciones existentes en el modelo de compilación tradicional. Incluso aplicando sofisticados análisis y transformaciones interprocedurales, un compilador tradicional no es capaz de optimizar un programa como una entidad completa. Un problema similar aparece aplicando técnicas de compilación dirigidas por profiling: grandes proyectos se ven forzados a recompilar todos y cada uno de sus módulos para aprovechar dicha información. Por el contrario, seria más conveniente construir la aplicación completa, instrumentarla para obtener información de profiling y optimizar entonces el binario final sin recompilar ni un solo fichero fuente.En esta tesis presentamos nuevas técnicas de compilación dirigidas por profiling para eliminar la redundancia encontrada en programas ejecutables a nivel binario (esto es, redundancia binaria), incluso aunque estos programas hayan sido compilados agresivamente con un novísimo compilador comercial. Nuestras técnicas de eliminación de redundancia están diseñadas para eliminar operaciones de memoria y de salto redundantes, que son las más importantes para mitigar los problemas de rendimiento que hemos mencionado. Estas propuestas están basadas en técnicas de eliminación de redundancia parcial sensibles al camino de ejecución. Los resultados muestran que aplicando nuestras optimizaciones, somos capaces de alcanzar una reducción del 14% en el tiempo de ejecución de nuestro conjunto de programas.En este trabajo también revisamos el problemas del análisis de alias en programas ejecutables, identificando el por qué la desambiguación de memoria es uno de los puntos débiles en la modificación de código objeto. Proponemos varios análisis para ser aplicados en el contexto de optimizadores binarios. Primero un análisis de alias estricto para descubrir dependencias de memoria sensibles al camino de ejecución, el cual es usado en nuestras optimizaciones para la eliminación de redundancias de memoria. Seguidamente, dos análisis especulativos de posibles alias para detección de independencias de memoria. Estos análisis están basados en introducir información especulativa en tiempo de análisis, lo que incrementa la precisión en partes importantes de código manteniendo el análisis eficiente. Los resultados muestran que nuestras propuestas son altamente útiles para incrementar la desambiguación de memoria de código binario, lo que se traduce en oportunidades para aplicar optimizaciones. Todos nuestros algoritmos, tanto de análisis como de optimización, han sido implementados en un optimizador binario, enfatizando los problemas más relevantes en la aplicaciones de nuestros algoritmos en código ejecutable, sin la ayuda de gran parte de la información de alto nivel presente en compiladores tradicionales. / Two of the most important performance limiters in today's processor families comes from solving the memory wall and handling control dependencies. In order to address these issues, cache memories and branch predictors are well-known hardware proposals that take advantage of, among other things, exploiting both temporal memory reuse and branch correlation. In other words, they try to exploit the dynamic redundancy existing in programs. This redundancy comes partly from the way that programmers write source code, but also from limitations in the compilation model of traditional compilers, which introduces unnecessary memory and conditional branch instructions. We believe that today's optimizing compilers should be very aggressive in optimizing programs, and then they should be expected to optimize a significant part of this redundancy away.On the other hand, optimizations performed at link-time or directly applied to final program executables have received increased attention in recent years, due to limitations in the traditional compilation model. First, even though performing sophisticated interprocedural analyses and transformations, traditional compilers do not have the opportunity to optimize the program as a whole. A similar problem arises when applying profile-directe compilation techniques: large projects will be forced to re-build every source file to take advantage of profile information. By contrast, it would be more convenient to build the full application, instrument it to obtain profile data and then re-optimize the final binary without recompiling a single source file.In this thesis we present new profile-guided compiler optimizations for eliminating the redundancy encountered on executable programs at binary level (i.e.: binary redundancy), even though these programs have been compiled with full optimizations using a state-ofthe- art commercial compiler. In particular, our Binary Redundancy Elimination (BRE) techniques are targeted at eliminating both redundant memory operations and redundant conditional branches, which are the most important ones for addressing the performance issues that we mentioned above in today's microprocessors. These new proposals are mainly based on Partial Redundancy Elimination (PRE) techniques for eliminating partial redundancies in a path-sensitive fashion. Our results show that, by applying our optimizations, we are able to achieve a 14% execution time reduction in our benchmark suite.In this work we also review the problem of alias analysis at the executable program level, identifying why memory disambiguation is one of the weak points of object code modification. We then propose several alias analyses to be applied in the context of linktime or executable code optimizers. First, we present a must-alias analysis to recognize memory dependencies in a path- sensitive fashion, which is used in our optimization for eliminating redundant memory operations. Next, we propose two speculative may-alias data-flow algorithms to recognize memory independencies. These may-alias analyses are based on introducing unsafe speculation at analysis time, which increases alias precision on important portions of code while keeping the analysis reasonably cost-efficient. Our results show that our analyses prove to be very useful for increasing memory disambiguation accuracy of binary code, which turns out into opportunities for applying optimizations.All our algorithms, both for the analyses and the optimizations, have been implemented within a binary optimizer, which overcomes most of the existing limitations of traditional source-code compilers. Therefore, our work also points out the most relevant issues of applying our algorithms at the executable code level, since most of the high-level information available in traditional compilers is lost.

The solar PV modules are influenced by a variety of loss mechanisms by which the energy yield is affected. A PV system is the sum of individual PV modules which should ideally operate similarly, however, inhomogeneous soiling, mismatching, and degradation, which are the main focus in this study, lead to dissimilarities in PV modules operating behavior and thus, lead to losses which will be assessed intensively in terms of energy yield. The dissimilarities in PV modules are referred to the ambient conditions or the PV modules characteristics which result in different modules’ maximum power point (MPP) and thus, different currents generated by each PV modules which cause the mismatching. However, the weakest PV module current governs the string current, and the weakest string voltage governs the voltage. Power optimizers are electronic devices connected to the PV modules which adjust the voltages of the PV modules in order to obtain the same current as the weakest module and thus, extract the modules’ MPP. Hence, the overall performance of the PV plant is enhanced. On the other hand, the power optimizers add additional cost to the plant’s investment cost and thus, the extra energy yield achieved by utilizing the power optimizers must be sufficient to compensate the additional cost of the power optimizers. This is assessed by designing three systems, a reference system with SMA inverters, a system utilizes Tigo power optimizers and SMA inverters, and a system utilizes SolarEdge power optimizers and inverters. The study considers four different locations which are Borlänge, Madrid, Abu Dhabi, and New Delhi. An Excel model is created and validated to emulate the inhomogeneous soiling and to evaluate the economic feasibility of the power optimiz ers. The model’s inputs are obtained from PVsyst and the precipitation data is obtained from Meteoblue and SMHI database. The economic model is based on the relation between Levelized Cost of Electricity (LCOE) which will be used to derive the discount rate. Graphs representing the discounted payback period as a function of the feed-in tariff for different discount rates is created in order to obtain the discounted payback period. The amount of extra energy yielded by the Tigo and the SolarEdge systems is dependent on the soiling accumulated on the PV modules. Relative to the reference system, 6.5 % annual energy gain by the systems utilizing the power optimizers in soiling conditions, up to 2.1 % in the degradation conditions, and up to 9.7 % annual energy gain at 10 % mismatching rate. The extra energy yield is dependent on the location, however, the Tigo and the SolarEdge systems have yielded more energy than the reference system in all cases except one case when the mismatch losses is set to zero. The precipitation pattern is very influential, and a scare precipitation leads to a reduction in the energy yield, in this case, the Tigo and the SolarEdge systems overall performance is enhanced and the extra energy gain becomes greater. The Tigo system yield slightly more energy than the SolarEdge system in most cases, however, during the plant’s lifetime, the SolarEdge system could become more efficient than the Tigo system which is referred to the system’s sizing ratio. The degradation of the system or the soiling accumulation decreases the irradiation and thus, a slightly oversized PV array become suitable and deliver an optimal power to the inverters. The SolarEdge system is feasible in all scenarios in terms of LCOE and discounted payback period, although its slightly lower performance relative to the Tigo system, this is referred to its low initial cost in comparison to the other systems. The Tigo system is mostly infeasible although it yields more energy than the reference and the SolarEdge systems, this is referred iii to its relatively high initial cost. However, feed- in tariffs higher than 20 € cent / kWh make all systems payback within less than 10 years. The results have overall uncertainty within ± 6.5 % including PVsyst, Excel model, and the precipitation uncertainties. The uncertainty in the degradation and the mismatching calculations is limited to PVsyst uncertainty which is ± 5 %. The uncertainties in LCOE in the location of New Delhi, since it is the worst-case scenario, are 5.1 % and 4 % for the reference and the systems utilizing power optimizers, respectively. Consequently, accommodating the uncertainties to the benefits gained by utilizing power optimizers indicates that the energy gain would oscillate in the range of 6 % - 6.9 % for the soiling calculations, 2 % - 2.2 % for the degradation simulations, and 9.2 % - 10.2 % for the mismatching simulations at 10 % mismatchrate.

Modern database systems use a query optimizer to identify the most efficient strategy, called "plan", to execute declarative SQL queries. Optimization is a mandatory exercise since the difference between the cost of best plan and a random choice could be in orders of magnitude. The role of query optimization is especially critical for the decision support queries featured in data warehousing and data mining applications. For a query on a given database and system configuration, the optimizer's plan choice is primarily a function of the selectivities of the base relations participating in the query. A pictorial enumeration of the execution plan choices of a database query optimizer over this relational selectivity space is called a "plan diagram". It has been shown recently that these diagrams are often remarkably complex and dense, with a large number of plans covering the space. An interesting research problem that immediately arises is whether complex plan diagrams can be reduced to a significantly smaller number of plans, without materially compromising the query processing quality. The motivation is that reduced plan diagrams provide several benefits, including quantifying the redundancy in the plan search space, enhancing the applicability of parametric query optimization, identifying error-resistant and least-expected-cost plans, and minimizing the overhead of multi-plan approaches. In this thesis, we investigate the plan diagram reduction issue from theoretical, statistical and empirical perspectives. Our analysis shows that optimal plan diagram reduction, w.r.t. minimizing the number of plans in the reduced diagram, is an NP-hard problem, and remains so even for a storage-constrained variation. We then present CostGreedy, a greedy reduction algorithm that has tight and optimal performance guarantees, and whose complexity scales linearly with the number of plans in the diagram. Next, we construct an extremely fast estimator, AmmEst, for identifying the location of the best tradeoff between the reduction in plan cardinality and the impact on query processing quality. Both CostGreedy and AmmEst have been incorporated in the publicly-available Picasso optimizer visualization tool. Through extensive experimentation with benchmark query templates on industrial-strength database optimizers, we demonstrate that with only a marginal increase in query processing costs, CostGreedy reduces even complex plan diagrams running to hundreds of plans to "anorexic" levels (small absolute number of plans). While these results are produced using a highly conservative upper-bounding of plan costs based on a cost monotonicity constraint, when the costing is done on "actuals" using remote plan costing, the reduction obtained is even greater - in fact, often resulting in a single plan in the reduced diagram. We also highlight how anorexic reduction provides enhanced resistance to selectivity estimate errors, a long-standing bane of good plan selection. In summary, this thesis demonstrates that complex plan diagrams can be efficiently converted to anorexic reduced diagrams, a result with useful implications for the design and use of next-generation database query optimizers.

2D to 3D reconstruction is an ill-posed problem in the field of Autonomous Robot Navigation. Many practitioners are tend to utilize the enormous success of Deep Learning techniques like CNN, ANN etc to solve tasks related to this 2D to 3D reconstruction. Generally, every deep learning model involves implementation of different optimizers related to the tasks to lower the possible negativity in its results and selection of hyper parameter values for these optimizers during the process of training the model with required dataset.Selection of this optimizer hyper-parameters requires in-depth knowledge and trials and errors. So proposing optimal hyper parameters for optimizers results in no waste in computational resources and time.Hence solution for the selected task cab found easily. The main objective of this research is to propose optimal hyper parameter values of various deep learning optimizers related to 2D to 3D reconstruction and proposing best optimizer among them in terms of computational time and resources To achieve the goal of this study two research methods are used in our work. The first one is a Systematic Literature Review; whose main goal is to reveal the widely selected and used optimizers for 2D to 3D reconstruction model using 3D Deep Learning techniques.The second, an experimental methodology is deployed, whose main goal is to propose the optimal hyper parameter values for respective optimizers like Adam, SGD+Momentum, Adagrad, Adadelta and Adamax which are used in 3D reconstruction models. In case of the computational time, Adamax optimizer outperformed all other optimizers used with training time (1970min), testing time (3360 min), evaluation-1 (16 min) and evaluation-2 (14 min).In case of Average Point cloud points, Adamax outperformed all other optimizers used with Mean value of 28451.04.In case of pred->GT and GT->pred values , Adamax optimizer outperformed all other optimizers with mean values of 4.742 and 4.600 respectively. Point Cloud Images with respective dense cloud points are obtained as results of our experiment.From the above results,Adamax optimizer is proved to be best in terms of visualization of Point Cloud images with optimal hyper parameter values as below:Epochs : 1000    Learning Rate : 1e-2    Chunk size : 32    Batch size : 32.  In this study,'Adamax' optimizer with optimal hyper para meter values and better Point Cloud Image is proven to be the best optimizer that can be used in a 2D to 3D reconstruction related task that deals with Point Cloud images

Orientador: Marcelo Carvalho Minhoto Teixeira / Resumo: Nesta tese, utiliza-se um otimizador analógico não derivativo proposto por Teixeira & Żak em 1999 como principal ferramenta para os sistemas de controle dos projetos desenvolvidos. Tal otimizador é composto por blocos não lineares e pode ser classificado como um sistema neural artificial. Sistemas chaveados têm grande aplicação prática na otimização de sistemas e são caracterizados por possuírem subsistemas e uma lei de chaveamento que seleciona cada subsistema a cada momento. Deve-se definir condições para que seja possível projetar uma lei de chaveamento que atenda requisitos de projeto. O estudo de técnicas de controle extremal na solução de problemas de busca pelo rastreamento do máximo ponto de potência (do inglês: Maximum Power Point Tracking - MPPT), vem apresentando resultados interessantes na literatura e um tipo de sistema à qual essa técnica pode ser aplicada, é na geração fotovoltaica. Aplica-se o otimizador analógico citado na busca do MPPT de uma célula fotovoltaica, com o objetivo de observar o controle extremal atuando em um processo de otimização, estendendo o controle para quando existem variações de irradiação solar (cenário de uma possível passagem de nuvens). Também observa-se o comportamento do sistema quanto a manter seu correto funcionamento e estabilidade ultimate bounded. A contribuição principal desta tese foi uma nova proposta de utilização conjunta do otimizador de Teixeira & Żak no projeto de controladores ˙ chaveados baseados na minimização da d... (Resumo completo, clicar acesso eletrônico abaixo) / Abstract: On this thesis, a non-derivative analog optimizer, proposed by Teixeira & Żak in 1999, was used as the main tool for the proposed control system. Such optimizer is structured by nonlinear blocks and can be classified as an artificial neural system. Switched systems have great theoretical and practical application in systems optimization and are characterized by having subsystems, and a switching law that selects each subsystem at each moment. It is necessary to define conditions so that it is possible to design a switching law for the desired performance of the controlled system. The study of Extremum Seeking Control techniques in the solution of problems of Maximum Power Point Tracking has presented interesting results, and one type of system which this technique can be applied is in the photovoltaic generation. The analog optimizer is applied in the Maximum Power Point Tracking of a photovoltaic cell, with the objective of observing the actuation of the extremal seeking control in an optimization process, extending the control when there are solar irradiation variations (a possible clouds passage scenario). And also observe the behavior of the system and how to maintain its correct functioning and ultimate bounded stability. The main contribution of this thesis was a new procedure for using the mentioned analog optimizer in the design of switched controllers based on the minimization of the derivative of a Lyapunov function. This method allows the relaxed design of controll... (Complete abstract click electronic access below) / Doutor