This thesis presents how performance data from hardware accelerators can be included in event logs. It extends the capabilities of trace-based performance analysis to also monitor and record data from this novel parallelization layer. The increasing awareness to power consumption of computing devices has led to an interest in hybrid computing architectures as well. High-end computers, workstations, and mobile devices start to employ hardware accelerators to offload computationally intense and parallel tasks, while at the same time retaining a highly efficient scalar compute unit for non-parallel tasks. This execution pattern is typically asynchronous so that the scalar unit can resume other work while the hardware accelerator is busy. Performance analysis tools provided by the hardware accelerator vendors cover the situation of one host using one device very well. Yet, they do not address the needs of the high performance computing community. This thesis investigates ways to extend existing methods for recording events from highly parallel applications to also cover scenarios in which hardware accelerators aid these applications. After introducing a generic approach that is suitable for any API based acceleration paradigm, the thesis derives a suggestion for a generic performance API for hardware accelerators and its implementation with NVIDIA CUPTI. In a next step the visualization of event logs containing data from execution streams on different levels of parallelism is discussed. In order to overcome the limitations of classic performance profiles and timeline displays, a graph-based visualization using Parallel Performance Flow Graphs (PPFGs) is introduced. This novel technical approach is using program states in order to display similarities and differences between the potentially very large number of event streams and, thus, enables a fast way to spot load imbalances. The thesis concludes with the in-depth analysis of a case-study of PIConGPU---a highly parallel, multi-hybrid plasma physics simulation---that benefited greatly from the developed performance analysis methods. / Diese Dissertation zeigt, wie der Ablauf von Anwendungsteilen, die auf Hardwarebeschleuniger ausgelagert wurden, als Programmspur mit aufgezeichnet werden kann. Damit wird die bekannte Technik der Leistungsanalyse von Anwendungen mittels Programmspuren so erweitert, dass auch diese neue Parallelitätsebene mit erfasst wird. Die Beschränkungen von Computersystemen bezüglich der elektrischen Leistungsaufnahme hat zu einer steigenden Anzahl von hybriden Computerarchitekturen geführt. Sowohl Hochleistungsrechner, aber auch Arbeitsplatzcomputer und mobile Endgeräte nutzen heute Hardwarebeschleuniger um rechenintensive, parallele Programmteile auszulagern und so den skalaren Hauptprozessor zu entlasten und nur für nicht parallele Programmteile zu verwenden. Dieses Ausführungsschema ist typischerweise asynchron: der Skalarprozessor kann, während der Hardwarebeschleuniger rechnet, selbst weiterarbeiten. Die Leistungsanalyse-Werkzeuge der Hersteller von Hardwarebeschleunigern decken den Standardfall (ein Host-System mit einem Hardwarebeschleuniger) sehr gut ab, scheitern aber an einer Unterstützung von hochparallelen Rechnersystemen. Die vorliegende Dissertation untersucht, in wie weit auch multi-hybride Anwendungen die Aktivität von Hardwarebeschleunigern aufzeichnen können. Dazu wird die vorhandene Methode zur Erzeugung von Programmspuren für hochparallele Anwendungen entsprechend erweitert. In dieser Untersuchung wird zuerst eine allgemeine Methodik entwickelt, mit der sich für jede API-gestützte Hardwarebeschleunigung eine Programmspur erstellen lässt. Darauf aufbauend wird eine eigene Programmierschnittstelle entwickelt, die es ermöglicht weitere leistungsrelevante Daten aufzuzeichnen. Die Umsetzung dieser Schnittstelle wird am Beispiel von NVIDIA CUPTI darstellt. Ein weiterer Teil der Arbeit beschäftigt sich mit der Darstellung von Programmspuren, welche Aufzeichnungen von den unterschiedlichen Parallelitätsebenen enthalten. Um die Einschränkungen klassischer Leistungsprofile oder Zeitachsendarstellungen zu überwinden, wird mit den parallelen Programmablaufgraphen (PPFGs) eine neue graphenbasisierte Darstellungsform eingeführt. Dieser neuartige Ansatz zeigt eine Programmspur als eine Folge von Programmzuständen mit gemeinsamen und unterchiedlichen Abläufen. So können divergierendes Programmverhalten und Lastimbalancen deutlich einfacher lokalisiert werden. Die Arbeit schließt mit der detaillierten Analyse von PIConGPU -- einer multi-hybriden Simulation aus der Plasmaphysik --, die in großem Maße von den in dieser Arbeit entwickelten Analysemöglichkeiten profiert hat.

This thesis presents an efficient self-consistent Hybrid Spectral Ray Tracing (HSRT) technique for analysis and design of multi-scale sub-millimeter wave problems, where sub-wavelength features are modeled using rigorous methods, and complex structures with dimensions in the order of tens or even hundreds of wavelengths are modeled by asymptotic methods. Quasi-optical devices are used in imaging arrays for sub-millimeter and terahertz applications, THz time-domain spectroscopy (THz-TDS), high-speed wireless communications, and space applications to couple terahertz radiation from space to a hot electron bolometer. These devices and structures, as physically small they have become, are very large in terms of the wavelength of the driving quasi-optical sources and may have dimension in the tens or even hundreds of wavelengths. Simulation and design optimization of these devices and structures is an extremely challenging electromagnetic problem. The analysis of complex electrically large unbounded wave structures using rigorous methods such as method of moments (MoM), finite element method (FEM), and finite difference time domain (FDTD) method can become almost impossible due to the need for large computational resources. Asymptotic high-frequency techniques are used for analysis of electrically large quasi-optical systems and hybrid methods for solving multi-scale problems. Spectral Ray Tracing (SRT) has a number of unique advantages as a candidate for hybridization. The SRT method has the advantages of Spectral Theory of Diffraction (STD). STD can model reflection, refraction and diffraction of an arbitrary wave incident on the complex structure, which is not the case for diffraction theories such as Geometrical Theory of Diffraction (GTD), Uniform theory of Diffraction (UTD) and Uniform Asymptotic Theory (UAT). By including complex rays, SRT can effectively analyze both near-fields and far-fields accurately with minimal approximations. In this thesis, a novel matrix representation of SRT is presented that uses only one spectral integration per observation point and applied to modeling a hemispherical and hyper-hemispherical lens. The hybridization of SRT with commercially available FEM and MoM software is proposed in this work to solve the complexity of multi-scale analysis. This yields a computationally efficient self-consistent HSRT algorithm. Various arrangements of the Hybrid SRT method such as FEM-SRT, and MoM-SRT, are investigated and validated through comparison of radiation patterns with Ansoft HFSS for the FEM method, FEKO for MoM, Multi-level Fast Multipole Method (MLFMM) and physical optics. For that a bow-tie terahertz antenna backed by hyper-hemispherical silicon lens, an on-chip planar dipole fabricated in SiGe:C BiCMOS technology and attached to a hyper-hemispherical silicon lens and a double-slot antenna backed by silica lens will be used as sample structures to be analyzed using the HSRT. Computational performance (memory requirement, CPU/GPU time) of developed algorithm is compared to other methods in commercially available software. It is shown that the MoM-SRT, in its present implementation, is more accurate than MoM-PO but comparable in speed. However, as shown in this thesis, MoM-SRT can take advantage of parallel processing and GPU. The HSRT algorithm is applied to simulation of on-chip dipole antenna backed by Silicon lens and integrated with a 180-GHz VCO and radiation pattern compared with measurements. The radiation pattern is measured in a quasi-optical configuration using a power detector. In addition, it is shown that the matrix formulation of SRT and HSRT are promising approaches for solving complex electrically large problems with high accuracy. This thesis also expounds on new measurement setup specifically developed for measuring integrated antennas, radiation pattern and gain of the embedded on-chip antenna in the mmW/ terahertz range. In this method, the radiation pattern is first measured in a quasi-optical configuration using a power detector. Subsequently, the radiated power is estimated form the integration over the radiation pattern. Finally, the antenna gain is obtained from the measurement of a two-antenna system.

The primary objective of this study is to evaluate the industry impact of a hypothetical Foot and Mouth Disease (FMD) outbreak in the Texas High Plains using alternative animal tracing levels. To accomplish this objective, an epidemiological disease spread model, AUSSPREAD, is used to simulate the FMD outbreak and an economic model is used to examine the impacts of different animal identification levels in cattle. The different levels of animal identification relate to the model?s ability to trace back the subsequent infected and/or dangerous contacts with which the initial outbreak herd has been in contact. The study examines direct disease management costs (slaughter, euthanasia, disposal, surveillance, and cleaning disinfection), forgone income, and other indirect costs (indemnity payments and welfare slaughter) for outbreaks originating from a large beef operation, a feedlot, and a saleyard across subsequent tracing periods from 1 to 10 days. Welfare slaughter and quarantine costs were estimated for the best and worst outbreaks from the feedlot operation. It is noteworthy that total direct costs of a FMD outbreak would be more extensive than the current study's calculations, which only analyzed the direct disease management costs. The increased days to trace dangerous contacts presented overall increases in outbreak losses over each outbreak scenario. Although outcome averages appear insensitive at times under the assumptions applied, the epidemiological model presented the possibility that traceability could reduce the risk of extreme outcomes in respect to the overall distribution of losses. For each cattle operation, the outbreaks stayed consistent or marginally increased with their respective average costs, but their maximum losses rose steadily, across the trace periods examined. The impact of increased traceability and decreased outbreak length can be justified in affecting FMD outbreak costs in a positive manner. The results provide the industry with estimations of different outbreak scenarios which can be used to inform the decision on the NAIS system. Longer tracing periods, larger simulations (by iteration), and further study of the model is necessary in order to more accurately imitate FMD outbreaks within the Texas High Plains and its detrimental effects.

There is a variety of tools to measure the performance of Linux systems and the applications running on them. However, the resulting performance data is often presented in plain text format or only with a very basic user interface. For large systems with many cores and concurrent threads, it is increasingly difficult to present the data in a clear way for analysis. Moreover, certain performance analysis and debugging tasks require the use of a high-resolution time-line based approach, again entailing data visualization challenges. Tools in the area of High Performance Computing (HPC) have long been able to scale to hundreds or thousands of parallel threads and help finding performance anomalies. We therefore present a solution to gather performance data using Linux performance monitoring interfaces. A combination of sampling and careful instrumentation allows us to obtain detailed performance traces with manageable overhead. We then convert the resulting output to the Open Trace Format (OTF) to bridge the gap between the recording infrastructure and HPC analysis tools. We explore ways to visualize the data by using the graphical tool Vampir. The combination of established Linux and HPC tools allows us to create an interface for easy navigation through time-ordered performance data grouped by thread or CPU and to help users find opportunities for performance optimizations.

This thesis presents an efficient self-consistent Hybrid Spectral Ray Tracing (HSRT) technique for analysis and design of multi-scale sub-millimeter wave problems, where sub-wavelength features are modeled using rigorous methods, and complex structures with dimensions in the order of tens or even hundreds of wavelengths are modeled by asymptotic methods. Quasi-optical devices are used in imaging arrays for sub-millimeter and terahertz applications, THz time-domain spectroscopy (THz-TDS), high-speed wireless communications, and space applications to couple terahertz radiation from space to a hot electron bolometer. These devices and structures, as physically small they have become, are very large in terms of the wavelength of the driving quasi-optical sources and may have dimension in the tens or even hundreds of wavelengths. Simulation and design optimization of these devices and structures is an extremely challenging electromagnetic problem. The analysis of complex electrically large unbounded wave structures using rigorous methods such as method of moments (MoM), finite element method (FEM), and finite difference time domain (FDTD) method can become almost impossible due to the need for large computational resources. Asymptotic high-frequency techniques are used for analysis of electrically large quasi-optical systems and hybrid methods for solving multi-scale problems. Spectral Ray Tracing (SRT) has a number of unique advantages as a candidate for hybridization. The SRT method has the advantages of Spectral Theory of Diffraction (STD). STD can model reflection, refraction and diffraction of an arbitrary wave incident on the complex structure, which is not the case for diffraction theories such as Geometrical Theory of Diffraction (GTD), Uniform theory of Diffraction (UTD) and Uniform Asymptotic Theory (UAT). By including complex rays, SRT can effectively analyze both near-fields and far-fields accurately with minimal approximations. In this thesis, a novel matrix representation of SRT is presented that uses only one spectral integration per observation point and applied to modeling a hemispherical and hyper-hemispherical lens. The hybridization of SRT with commercially available FEM and MoM software is proposed in this work to solve the complexity of multi-scale analysis. This yields a computationally efficient self-consistent HSRT algorithm. Various arrangements of the Hybrid SRT method such as FEM-SRT, and MoM-SRT, are investigated and validated through comparison of radiation patterns with Ansoft HFSS for the FEM method, FEKO for MoM, Multi-level Fast Multipole Method (MLFMM) and physical optics. For that a bow-tie terahertz antenna backed by hyper-hemispherical silicon lens, an on-chip planar dipole fabricated in SiGe:C BiCMOS technology and attached to a hyper-hemispherical silicon lens and a double-slot antenna backed by silica lens will be used as sample structures to be analyzed using the HSRT. Computational performance (memory requirement, CPU/GPU time) of developed algorithm is compared to other methods in commercially available software. It is shown that the MoM-SRT, in its present implementation, is more accurate than MoM-PO but comparable in speed. However, as shown in this thesis, MoM-SRT can take advantage of parallel processing and GPU. The HSRT algorithm is applied to simulation of on-chip dipole antenna backed by Silicon lens and integrated with a 180-GHz VCO and radiation pattern compared with measurements. The radiation pattern is measured in a quasi-optical configuration using a power detector. In addition, it is shown that the matrix formulation of SRT and HSRT are promising approaches for solving complex electrically large problems with high accuracy. This thesis also expounds on new measurement setup specifically developed for measuring integrated antennas, radiation pattern and gain of the embedded on-chip antenna in the mmW/ terahertz range. In this method, the radiation pattern is first measured in a quasi-optical configuration using a power detector. Subsequently, the radiated power is estimated form the integration over the radiation pattern. Finally, the antenna gain is obtained from the measurement of a two-antenna system.

Text in English / Botswana has witnessed highest TB rates in the southern African countries, ranking the fourth after South Africa, Swaziland and Zimbabwe. In 2012, the TB rate was on average 531/100 000 population. About 2 380 contacts out of a possible 8 110 (amounting to 29.30%) were traced nationally (Botswana 2011:8), indicating a possible gap of 5 730 which was yet to be traced in 2011. The TBCT strategies might be inadequate leading to absence of screening and treating TB contacts and reducing PTB related deaths. The purpose of this study was to describe utilisation of current TBCT and develop a strategy for a more effective TBCT in Botswana. Data was collected through a quantitative cross-sectional research design. The study further described the association between TBCT strategies and practices and determined the gaps, challenges and needs in the TBCT. Results revealed under-tracing of contacts in the number of registered and enumerated TB contacts. The results further established the risk of mixing TB contacts and the general patients. The differences in the perceptions and knowledge of the cause of TB as well as poor utilisation of the current programmes by the PTB patients denotes the need for aggressive awareness raising and health promotion strategies. The results were used to develop an alternative strategy, the IC-TBCT, which has a potential to trace all TB contacts. The strategy encourages participation, effective accountability and involvement of the beneficiaries in all efforts aiming at early contact identification and reducing the incidence of PTB. / Health Studies / D. Litt. et Phil. (Health Studies)

Este projeto de pesquisa, que desperta a vivência e suscita questionamentos pessoais sobre os aparentes resultados da formação continuada de professores com o uso das Tecnologias Digitais de Informação e Comunicação TDIC , inicia-se com a formação introdutória para uso de software de programação para professores de informática educativa do município de São Paulo. E, em razão do rápido avanço nas relações entre as novas tecnologias e a rede, em todas as áreas socioeconômico-culturais, constata ser imprescindível expandir o conceito de fluência digital, para fortalecer o currículo Trivium contemporâneo e melhorar a situação real da educação brasileira. E questiona: qual leitura segundo a abordagem antiessencialista da tecnologia é realizada, pelo professor em formação, do software de programação, neste caso o Scratch, para utilizá-lo com seus alunos em nossas atuais escolas? Para tal investigação, utiliza-se o estudo de caso com a técnica do process tracing, adaptada à educação, de maneira a transferir a análise das causas e dos resultados para o processo em si, que é desconhecido, na criação de uma teoria. Por meio das manifestações e das atividades reais, desde a formação até a aplicação do Scratch em classe, preveem-se e testam-se evidências em mecanismo causal hipotético. Tais evidências, confirmadas ou não pelas observações em classe, por entrevistas, questionários, depoimentos e rotinas de aula, nos casos em estudo, contribuem para que se proponha a Formação Insubordinada de Professores para o uso das TDIC FoIP-TDIC. O modelo traz como premissas serem as TDIC dependentes da cultura; por meio delas, o professor pode estimular a criatividade, a consciência cidadã, a investigação e a resolução de problemas reais e relevantes, de forma a elaborar criticamente seu cotidiano. A FoIP-TDIC relativiza as evidências tidas por formadores e formandos como necessárias e/ou imprescindíveis em uma formação. Portanto, pretende ampliar ideias para a estrutura e a organização de formação inicial ou continuada do professor da educação básica, seja ela presencial ou semipresencial. Aprofundar pesquisas da FoIP-TDIC poderá contribuir para a formação rumo à Educação 4.0 em uma sociedade que se deseja solidária. / This research project, which awakens the experience and raises personal questions about the apparent results of the ongoing training of teachers with the use of Digital Information and Communication Technologies (TDIC), begins with the introductory training for the use of programming software for teachers of educational computing in the city of São Paulo. Because of the rapid progress in the relations between the new technologies and the network in all socioeconomic and cultural areas, it is essential to expand the concept of digital fluency, to strengthen the contemporary Trivium curriculum and to improve the real situation of Brazilian education. He questions: what reading - according to the antiessentialist approach of technology - is carried out by the teacher in training of programming software, in this case Scratch, to use it with his students in our current schools? For this investigation, the case study is used with the process tracing technique, adapted to education, in order to transfer the analysis of causes and results to the process itself, which is unknown, in the creation of a theory. Through the manifestations and real activities, from the formation to the application of the Scratch in class, evidence and tests in hypothetical causal mechanism are foreseen and they are tested. These evidences, confirmed or not by classroom observations, through interviews, questionnaires, statements and classroom routines, in the cases under study, contribute to the proposed Insubordinate Teacher Training for the use of TDIC - FoIP-TDIC. The model assumes that the culture-dependent TDICs are; through them, the teacher can stimulate creativity, citizen awareness, research and resolution of real and relevant problems, in order to, critically elaborate their daily lives. The FoIP-TDIC relativizes the evidences taken by trainers and trainees as necessary and / or essential in a formation. Therefore, it intends to broaden ideas for the structure and organization of initial or continuing training of the elementary school teacher, whether it is classroom or semi-presential. Deepening FoIP-TDIC research may contribute to the formation towards Education 4.0 in a society that wishes to be in solidarity.

O objetivo desta dissertação é construir uma hipótese de mecanismo causal do esquerdismo nos governos de esquerda renovadora da América Latina a partir do estudo de caso do racha à esquerda do Frente Amplio (FA) do Uruguai, e testar esta hipótese com o caso do racha à esquerda do Partido dos Trabalhadores (PT) do Brasil. Utilizando theory building process-tracing, a partir da literatura sobre partidos políticos foram elaboradas, e testadas no caso uruguaio, três hipóteses de partes de mecanismo causal, as quais estabelecem: (i) o movimento ao centro feito pelo FA para atingir objetivos eleitorais é parte do mecanismo causal do racha da fração 26M; (ii) as situações de crise com o governo do Presidente Tabaré Vázquez se constituíram em parte do mecanismo causal do racha do 26M; e (iii), a diminuição das oportunidades do 26M de incidir na toma de decisões do FA, combinada com sua perda de zonas de incerteza dentro do partido, tornou- se parte do mecanismo causal do racha desta fração. As evidências apresentadas para cada hipótese suportaram o Hoop test, o que deu um alto poder confirmatório das hipóteses, porém, não permitiu descartar hipóteses alternativas. Com base nos resultados do caso Uruguaio, se construiu uma hipótese de mecanismo causal dos processos de esquerdismo nos governos de esquerda renovadora da América Latina que contém 8 partes, as entidades (atores) e ações envolvidas em cada uma delas. Por fim, utilizando theory-testing process-tracing, se fizeram apontamentos sobre o funcionamento da hipótese no caso do processo das frações que racharam com o PT e formaram o Partido Socialismo e Liberdade (PSOL). As evidências apresentadas neste caso não foram submetidas a testes empíricos, porém, de maneira geral, não diminuíram nossa confiança na hipótese de mecanismo causal. Inclusive, se produzindo uma trajetória divergente entre as frações que derivou na estruturação em dois momentos do PSOL, não foi necessário adicionar de maneira ad-hoc partes para o mecanismo, pois as foças causais foram transmitidas pelas partes nele estabelecidas, encurtando-o no caso das frações que racharam no primeiro momento, e percorrendo-o parte por parte, no caso da fração que rachou no segundo momento. / The aim of this Master\'s dissertation is to construct a causal mechanism hypothesis of leftism in the Latin American renovation of left-wing governments. This will be reviewed in two parts: starting with the case study of the split in the left of the Frente Amplio (FA) party of Uruguay, and furthermore to test the hypothesis with the case study of the split in the left of the Partido dos Trabalhadores (PT) of Brazil. Using theory- building process-tracing from literature about political parties, three hypotheses of parts of causal mechanism were elaborated and tested in the Uruguayan case, which established that: (i) the movement towards the center made by the FA to achieve electoral objectives was part of the causal mechanism of the split of the fraction 26M; (ii) crisis situations with the government of President Tabaré Vázquez were part of the causal mechanism of the split of 26M; and (iii) the reduction of 26M\'s opportunities to influence FA decision-making process, combined with the loss of zones of uncertainty within the party, became part of the causal mechanism of fracture of this fraction. The evidences presented for each hypothesis withstood the Hoop test, which gave a high confirmatory power to the hypotheses, however, they did not allow for the disregard of alternative hypotheses. Based on the results in the case of Uruguay, a hypothesis of causal mechanism of the leftist processes in the Latin American renovation of left-wing governments was constructed, which contains 8 parts, the entities (actors) and actions involved in each one of them. Finally, using theory-testing process-tracing, notes were made on the functionality of the hypothesis in the process of fractions that broke with the PT and formed the Partido Socialismo e Liberdade (PSOL). The evidence presented in this case was not subjected to empirical tests, however, in general, it did not diminish our confidence in the causal mechanism hypothesis. Even if there was a divergent trajectory between the fractions that resulted in the structuring in two moments of PSOL, it was not necessary to add parts to the mechanism ad-hoc, because the causal forces were transmitted to the parties in it, shortening it in the case of the fractions that broke initially, and going part by part, in the case of the fraction that broke after.

This thesis is pointed on simulation of electron microscope. It uses knowledge  from ray-tracing methods and physically based rendering with physics of electron microscopy. The main object of this thesis is to create realistic simulator for electron microscope which can generate satisfying realistic images. This output can be used for another research in range of machine learning or microscope application development. This method should replace long and more expensive way of getting test images from real system for those research activities.

This thesis is dedicated to ray tracing based rendering methods, primarily the original ray tracing. Besides introducing a brief historical overview of algorithms from the family, it presents all the essential tools, techniques and physics needed for designing a rendering application in detail. A significant part of the document consists of an implementation of a photorealistic rendering application for interactive graphics 3D virtual scenes. The focus is on rendering without using any additional model information. The thesis includes descriptions and explanations of specific problems and their solutions.