While exact methods, such as DDA or T-matrix, can be applied to particles withsizes comparable to the wavelength, computational demands mean that they are size limited. For particles much larger than the wavelength, the Geometric Optics approximation can be employed, but in doing so wave effects, such as interference and diffraction, are ignored. In between these two size extremes there exists a need for computational techniques which are capable of handling the wide array of ice crystal shapes and sizes that are observed in cirrus clouds. The Beam Tracing model developed within this project meets these criteria. It combines aspects of geometric optics and physical optics. Beam propagation is handled by Snell's law and the law of reflection. A beam is divided into reflected and transmitted components each time a crystal facet is illuminated. If the incident beam illuminates multiple facets it is split, with a new beam being formed for each illuminated facet. The phase-dependent electric field amplitude of the beams is known from their ampli- tude (Jones) matrices. These are modified by transmission and reflection matrices, whose elements are Fresnel amplitude coefficients, each time a beam intersects a crystal facet. Phase tracing is carried out for each beam by considering the path that its 'centre ray' would have taken. The local near-field is then mapped, via a surface integral formulation of a vector Kirchhoff diffraction approximation, to the far-field. Once in the far-field the four elements of the amplitude matrix are trans- formed into the sixteen elements of the scattering matrix via known relations. The model is discussed in depth, with details given on its implementation. The physical basis of the model is given through a discussion of Ray Tracing and how this leads to the notion of Beam Tracing. The beam splitting algorithm is described for convex particles followed by the necessary adaptations for concave and/or ab- sorbing particles. Once geometric aspects have been established details are given as to how physical properties of beams are traced including: amplitude, phase and power. How diffraction is implemented in the model is given along with a review of existing diffraction implementations. Comparisons are given, first against a modified Ray Tracing code to validate the geometric optics aspects of the model. Then, specific examples are given for the cases of transparent, pristine, smooth hexagonal columns of four different sizes and orientations; a highly absorbing, pristine, smooth hexagonal column and a highly absorbing, indented, smooth hexagonal column. Analysis of two-dimensional and one-dimensional intensity distributions and degree of linear polarisation results are given for each case and compared with results acquired through use of the Amster- dam Discrete-Dipole Approximation (ADDA) code; with good agreement observed. To the author's best knowledge, the Beam Tracer developed here is unique in its ability to handle concave particles; particles with complex structures and the man- ner in which beams are divided into sub-beams of quasi-constant intensity when propagating in an absorbing medium. One of the model's potential applications is to create a database of known particle scattering patterns, for use in aiding particle classification from images taken by the Small Ice Detector (SID) in-situ probe. An example of creating such a database for hexagonal columns is given.

Le concept de processeur multicœurs constitue le facteur dominant pour offrir des hautes performances aux applications parallèles. Afin de développer des applications parallèles capable de tirer profit de ces plate-formes, les développeurs doivent prendre en compte plusieurs aspects, allant de l'architecture aux caractéristiques propres à l'application. Dans ce contexte, la Mémoire Transactionnelle (Transactional Memory – TM) apparaît comme une alternative intéressante à la synchronisation basée sur les verrous pour ces plates-formes. Elle permet aux programmeurs d'écrire du code parallèle encapsulé dans des transactions, offrant des garanties comme l'atomicité et l'isolement. Lors de l'exécution, les opérations sont exécutées spéculativement et les conflits sont résolus par ré-exécution des transactions en conflit. Bien que le modèle de TM ait pour but de simplifier la programmation concurrente, les meilleures performances ne pourront être obtenues que si l'exécutif est capable de s'adapter aux caractéristiques des applications et de la plate-forme. Les contributions de cette thèse concernent l'analyse et l'amélioration des performances des applications basées sur la Mémoire Transactionnelle Logicielle (Software Transactional Memory – STM) pour des plates-formes multicœurs. Dans un premier temps, nous montrons que le modèle de TM et ses performances sont difficiles à analyser. Pour s'attaquer à ce problème, nous proposons un mécanisme de traçage générique et portable qui permet de récupérer des événements spécifiques à la TM afin de mieux analyser les performances des applications. Par exemple, les données tracées peuvent être utilisées pour détecter si l'application présente des points de contention ou si cette contention est répartie sur toute l'exécution. Notre approche peut être utilisée sur différentes applications et systèmes STM sans modifier leurs codes sources. Ensuite, nous abordons l'amélioration des performances des applications sur des plate-formes multicœurs. Nous soulignons que le placement des threads (thread mapping) est très important et peut améliorer considérablement les performances globales obtenues. Pour faire face à la grande diversité des applications, des systèmes STM et des plates-formes, nous proposons une approche basée sur l'Apprentissage Automatique (Machine Learning) pour prédire automatiquement les stratégies de placement de threads appropriées pour les applications de TM. Au cours d'une phase d'apprentissage préliminaire, nous construisons les profiles des applications s'exécutant sur différents systèmes STM pour obtenir un prédicteur. Nous utilisons ensuite ce prédicteur pour placer les threads de façon statique ou dynamique dans un système STM récent. Finalement, nous effectuons une évaluation expérimentale et nous montrons que l'approche statique est suffisamment précise et améliore les performances d'un ensemble d'applications d'un maximum de 18%. En ce qui concerne l'approche dynamique, nous montrons que l'on peut détecter des changements de phase d'exécution des applications composées des diverses charges de travail, en prévoyant une stratégie de placement appropriée pour chaque phase. Sur ces applications, nous avons obtenu des améliorations de performances d'un maximum de 31% par rapport à la meilleure stratégie statique. / Multicore processors are now a mainstream approach to deliver higher performance to parallel applications. In order to develop efficient parallel applications for those platforms, developers must take care of several aspects, ranging from the architectural to the application level. In this context, Transactional Memory (TM) appears as a programmer friendly alternative to traditional lock-based concurrency for those platforms. It allows programmers to write parallel code as transactions, which are guaranteed to execute atomically and in isolation regardless of eventual data races. At runtime, transactions are executed speculatively and conflicts are solved by re-executing conflicting transactions. Although TM intends to simplify concurrent programming, the best performance can only be obtained if the underlying runtime system matches the application and platform characteristics. The contributions of this thesis concern the analysis and improvement of the performance of TM applications based on Software Transactional Memory (STM) on multicore platforms. Firstly, we show that the TM model makes the performance analysis of TM applications a daunting task. To tackle this problem, we propose a generic and portable tracing mechanism that gathers specific TM events, allowing us to better understand the performances obtained. The traced data can be used, for instance, to discover if the TM application presents points of contention or if the contention is spread out over the whole execution. Our tracing mechanism can be used with different TM applications and STM systems without any changes in their original source codes. Secondly, we address the performance improvement of TM applications on multicores. We point out that thread mapping is very important for TM applications and it can considerably improve the global performances achieved. To deal with the large diversity of TM applications, STM systems and multicore platforms, we propose an approach based on Machine Learning to automatically predict suitable thread mapping strategies for TM applications. During a prior learning phase, we profile several TM applications running on different STM systems to construct a predictor. We then use the predictor to perform static or dynamic thread mapping in a state-of-the-art STM system, making it transparent to the users. Finally, we perform an experimental evaluation and we show that the static approach is fairly accurate and can improve the performance of a set of TM applications by up to 18%. Concerning the dynamic approach, we show that it can detect different phase changes during the execution of TM applications composed of diverse workloads, predicting thread mappings adapted for each phase. On those applications, we achieve performance improvements of up to 31% in comparison to the best static strategy.

Realistic Image synthesis, usually, requires long computations and the simulation of the light interacting with a virtual scene. One of the most computationally intensive simulation in this area is the visualization of solid participating media. This media can describe many different types of object with the same physical parameters (e.g. marble, air, fire, skin, wax ...). Simulating the light interacting with it requires the computation of many independent photons interactions inside the medium. However, those interactions can be computed in parallel, using the power of modern Graphic Processor Unit, or GPU, computing. This work present an overview over different methodologies, that can affect the performance of this type of simulations on the GPU. Different existing ideas are analyzed, compared and modified with the scope of speeding up the computation respect to the classic CPU implementation. 1

Made available in DSpace on 2014-06-12T16:01:28Z (GMT). No. of bitstreams: 2 arquivo9410_1.pdf: 4384561 bytes, checksum: 4ebaaa7cbd8455ac2eed9a38c2530cf4 (MD5) license.txt: 1748 bytes, checksum: 8a4605be74aa9ea9d79846c1fba20a33 (MD5) Previous issue date: 2012 / A Realidade Aumentada é um campo de pesquisa que trata do estudo de técnicas para integrar informações virtuais com o mundo real. Algumas aplicações de Realidade Aumentada requerem fotorrealismo, onde os elementos virtuais são tão coerentemente inseridos na cena real que o usuário não consegue distinguir o virtual do real. Para a síntese de cenas 3D existem diversas técnicas, entre elas o ray tracing. Ele é um algoritmo baseado em conceitos básicos da Física Ótica, cuja principal característica é a alta qualidade visual a um custo computacional elevado, o que condicionava a sua utilização a aplicações offline. Contudo, com o avanço do poder computacional das GPUs este algoritmo passou a ser viável para ser utilizado em aplicações de tempo real, devido principalmente ao fato de ser um algoritmo com a característica de poder ser massivamente paralelizado. Levando isto em consideração, esta dissertação propõe um pipeline para renderização fotorrealística em tempo real utilizando a técnica ray tracing em aplicações de Realidade Aumentada. O ray tracer utilizado foi o Real Time Ray Tracer, ou RT2, de Santos et al., que serviu de base para a construção de um pipeline com suporte a sombreamento, síntese de diversos tipos de materiais, oclusão, reflexão, refração e alguns efeitos de câmera. Para que fosse possível obter um sistema que funciona a taxas interativas, todo o pipeline de renderização foi implementado em GPU, utilizando a linguagem CUDA, da NVIDIA. Outra contribuição importante deste trabalho é a integração deste pipeline com o dispositivo Kinect, da Microsoft, possibilitando a obtenção de informações reais da cena, em tempo real, eliminando assim a necessidade de se conhecer previamente os objetos pertencentes à cena real

O hipotálamo modula comportamentos relacionados à motivação, recompensa e punição através de projeções para a área tegmental ventral (VTA), o núcleo dorsal da rafe (DR) e o núcleo tegmental rostromedial (RMTg). Nesse estudo, investigamos através de métodos de rastreamento retrógrado as entradas hipotalâmicas da VTA, do DR e do RMTg e, se neurônios hipotalâmicos individuais inervam mais do que uma dessas regiões. Também determinamos uma possível assinatura GABAérgica ou glutamatérgica das aferências hipotalâmicas, através de rastreamento retrógrado combinado com métodos de hibridação in situ. Observamos que VTA, DR e RMTg recebem um padrão bastante semelhante de entradas hipotalâmicas originando de neurônios de projeção glutamatérgicas e GABAérgicas, a maioria deles (> 90%) inervando somente um desses três alvos. Nossos achados indicam que entradas hipotalâmicas são importantes fontes de sinais homeostáticos para a VTA, o DR e o RMTg. Eles exibem um alto grau de heterogeneidade que permite de excitar ou inibir as três estruturas de forma independente ou em conjunto. / The hypothalamus modulates behaviors related to motivation, reward and punishment via projections to the ventral tegmental area (VTA), dorsal raphe nucleus (DR), and rostromedial tegmental nucleus (RMTg). In this study we investigated by retrograde tracing methods hypothalamic inputs to the VTA, DR, and RMTg, and whether individual hypothalamic neurons project to more than one of these structures. We also determined a possible GABAergic or glutamatergic phenotype of hypothalamic afferents, by combining retrograde tracing with in situ hybridization methods. We found that VTA, DR, and RMTg receive a very similar set of hypothalamic afferents originating from glutamatergic and GABAergic hypothalamic projection neurons, the majority of them (> 90%) only innervating one of these structures. Our findings indicate that hypothalamic inputs are important sources of homeostatic signals for the VTA, DR, and RMTg. They exhibit a high degree of heterogeneity which permits to activate or inhibit the three structures either independently or jointly.

Submitted by Renata Lopes (renatasil82@gmail.com) on 2017-03-03T11:28:48Z No. of bitstreams: 1 gildodealmeidaleonel.pdf: 24098922 bytes, checksum: 16845fd58b93cf751e3ef6da19f65159 (MD5) / Approved for entry into archive by Adriana Oliveira (adriana.oliveira@ufjf.edu.br) on 2017-03-06T20:02:53Z (GMT) No. of bitstreams: 1 gildodealmeidaleonel.pdf: 24098922 bytes, checksum: 16845fd58b93cf751e3ef6da19f65159 (MD5) / Made available in DSpace on 2017-03-06T20:02:53Z (GMT). No. of bitstreams: 1 gildodealmeidaleonel.pdf: 24098922 bytes, checksum: 16845fd58b93cf751e3ef6da19f65159 (MD5) Previous issue date: 2011-01-17 / Campos tensoriais arbitrários são úteis em diversas áreas do conhecimento como a física, engenharias e áreas da saúde. Um dos principais interesses de profissionais destas áreas é a investigação de objetos colineares e coplanares representados pelos tensores. Esses objetos são formados por subconjuntos estruturados de tensores presentes no campo e que capturam alguma continuidade geométrica. Pela sua natureza multivariada, a visualização de elementos organizados é uma tarefa desafiadora. Geralmente, utilizam-se métodos de detecção direta destas estruturas para que o observador possa analisá-las. A proposta desta dissertação é explorar o fato de que o movimento estimula percepções complexas de forma inata no sistema visual humano. A abordagem desenvolvida utiliza um sistema de rastreamento de partículas e é parametrizado por campos tensoriais de forma que o comportamento das partículas represente as características do campo e tenha um aprimoramento que possibilite o melhor entendimento e a interpretação da informação proveniente dos tensores. / Arbitrary tensor fields are useful in several areas as physics, engineering and medicine. The investigation of collinear and coplanar objects represented by tensors is the main focus of research in these areas. These objects are formed by structured tensorial fields which captures some geometric continuity. The visualization of strutured elements is a challenging task because of their multivariate nature. To be analysed by the user, direct methods are usually used for detecting these structures. The proposal of this dissertation is to explore the fact that movement increases the perception of complex shapes, that are observed in a innate form by the human visual system. The approach developed uses a particle tracing system and is parameterized by tensor fields, so the particles flow represents the characteristics of the field and make an improvement that enables better understanding and interpretation of information derived from tensors.

Submitted by Pedro Henrique Rodrigues (pedro.henriquer@ufpe.br) on 2015-03-04T18:12:29Z No. of bitstreams: 2 Dissertacao_completa_Diego_Lemos.pdf: 4382725 bytes, checksum: 304625beefcdb33f03bb97376f48c770 (MD5) license_rdf: 1232 bytes, checksum: 66e71c371cc565284e70f40736c94386 (MD5) / Made available in DSpace on 2015-03-04T18:12:29Z (GMT). No. of bitstreams: 2 Dissertacao_completa_Diego_Lemos.pdf: 4382725 bytes, checksum: 304625beefcdb33f03bb97376f48c770 (MD5) license_rdf: 1232 bytes, checksum: 66e71c371cc565284e70f40736c94386 (MD5) Previous issue date: 2012-03-09 / A Realidade Aumentada é um campo de pesquisa que trata do estudo de técnicas para integrar informações virtuais com o mundo real. Algumas aplicações de Realidade Aumentada requerem fotorrealismo, onde os elementos virtuais são tão coerentemente inseridos na cena real que o usuário não consegue distinguir o virtual do real. Para a síntese de cenas 3D existem diversas técnicas, entre elas o ray tracing. Ele é um algoritmo baseado em conceitos básicos da Física Ótica, cuja principal característica é a alta qualidade visual a um custo computacional elevado, o que condicionava a sua utilização a aplicações offline. Contudo, com o avanço do poder computacional das GPUs este algoritmo passou a ser viável para ser utilizado em aplicações de tempo real, devido principalmente ao fato de ser um algoritmo com a característica de poder ser massivamente paralelizado. Levando isto em consideração, esta dissertação propõe um pipeline para renderização fotorrealística em tempo real utilizando a técnica ray tracing em aplicações de Realidade Aumentada. O ray tracer utilizado foi o Real Time Ray Tracer, ou RT2, de Santos et al., que serviu de base para a construção de um pipeline com suporte a sombreamento, síntese de diversos tipos de materiais, oclusão, reflexão, refração e alguns efeitos de câmera. Para que fosse possível obter um sistema que funciona a taxas interativas, todo o pipeline de renderização foi implementado em GPU, utilizando a linguagem CUDA, da NVIDIA. Outra contribuição importante deste trabalho é a integração deste pipeline com o dispositivo Kinect, da Microsoft, possibilitando a obtenção de informações reais da cena, em tempo real, eliminando assim a necessidade de se conhecer previamente os objetos pertencentes à cena real.

Context. Ray tracing has always been a simple but effective way to create a photorealistic scene but at a greater cost when expanding the scene. Recent improvements in GPU and CPU hardware have made ray tracing faster, making more complex scenes possible with the same amount of time needed to process the scene. Despite the improvements in hardware ray tracing is still rarely run at a interactive speed. Objectives. The aim of this experiment was to implement a new kdtree generation algorithm using DirectX 11 compute shaders. Methods. The implementation created during the experiment was tested using two platforms and five scenarios where the generation time for the kd-tree was measured in milliseconds. The results where compared to a sequential implementation running on the CPU. Results. In the end the kd-tree generation algorithm implemented did not run within our definition of real-time. Comparing the generation times from the implementations shows that there is a speedup for the GPU implementation compared to our CPU implementation, it also shows linear scaling for the generation time as the number of triangles in the scene increase. Conclusions. Noticeable limitations encountered during the experiment was that the handling of dynamic structures and sorting of arrays are limited which forced us to use less memory efficient solutions.

Free-form surfaces and implicit surfaces must be tessellated before being rendered with rasterization techniques. However ray tracing provides the means to directly render such objects without the need to first convert into polygonal meshes. Since ray tracing can handle triangle meshes as well, the question of which method is most suitable in terms of performance, quality and memory usage is addressed in this thesis. Bézier surfaces and NURBS surfaces along with basic algebraic implicit surfaces are implemented in order to test the performance relative to polygonal meshes approximating the same objects. The parametric surfaces are implemented using an iterative Newtonian method that converges on the point of intersection using a bounding volume hierarchy that stores the initial guesses. Research into intersecting rays with parametric surfaces is surveyed in order to find additional methods that speed up the computation. The implicit surfaces are implemented using common direct algebraic methods. All of the intersection tests are implemented using the Embree ray tracing API as well as a SIMD library in order to achieve interactive framerates on a CPU. The results show that both Bézier surfaces and NURBS surfaces can achieve interactive framerates on a CPU using SIMD computation, with Bézier surfaces coming close to the performance of polygonal counterparts. The implicit surfaces implemented outperform even the simplest polygonal approximations.

Ray tracing is a popular algorithm used to simulate the behavior of light and is commonly used to render images with high levels of visual realism. Modern multicore CPUs and many-core GPUs can take advantage of the parallel nature of ray tracing to accelerate the rendering process and produce new images in real-time. For non-specialized hardware however, such implementations are often limited to low screen resolutions, simple scene geometry and basic graphical effects. In this work, a C++ framework was created to investigate how the ray tracing algorithm can be implemented and accelerated on the CPU and GPU, respectively. The framework is capable of utilizing two third-party ray tracing libraries, Intel’s Embree and NVIDIA’s OptiX, to ray trace various 3D scenes. The framework also supports several effects for added realism, a user controlled camera and triangle meshes with different materials and textures. In addition, a hybrid ray tracing solution is explored, running both libraries simultaneously to render subsections of the screen. Benchmarks performed on a high-end CPU and GPU are finally presented for various scenes and effects. Throughout these results, OptiX on a Titan X performed better by a factor of 2-4 compared to Embree running on an 8-core hyperthreaded CPU within the same price range. Due to this imbalance of the CPU and GPU along with possible interferences between the libraries, the hybrid solution did not give a significant speedup, but created possibilities for future research.