909 resultados para GPU acceleration
Resumo:
As simulações que pretendam modelar fenómenos reais com grande precisão em tempo útil exigem enormes quantidades de recursos computacionais, sejam estes de processamento, de memória, ou comunicação. Se até há pouco tempo estas capacidades estavam confinadas a grandes supercomputadores, com o advento dos processadores multicore e GPUs manycore os recursos necessários para este tipo de problemas estão agora acessíveis a preços razoáveis não só a investigadores como aos utilizadores em geral. O presente trabalho está focado na otimização de uma aplicação que simula o comportamento dinâmico de materiais granulares secos, um problema do âmbito da Engenharia Civil, mais especificamente na área da Geotecnia, na qual estas simulações permitem por exemplo investigar a deslocação de grandes massas sólidas provocadas pelo colapso de taludes. Assim, tem havido interesse em abordar esta temática e produzir simulações representativas de situações reais, nomeadamente por parte do CGSE (Australian Research Council Centre of Excellence for Geotechnical Science and Engineering) da Universidade de Newcastle em colaboração com um membro da UNIC (Centro de Investigação em Estruturas de Construção da FCT/UNL) que tem vindo a desenvolver a sua própria linha de investigação, que se materializou na implementação, em CUDA, de um algoritmo para GPUs que possibilita simulações de sistemas com um elevado número de partículas. O trabalho apresentado consiste na otimização, assente na premissa da não alteração (ou alteração mínima) do código original, da supracitada implementação, de forma a obter melhorias significativas tanto no tempo global de execução da aplicação, como no aumento do número de partículas a simular. Ao mesmo tempo, valida-se a formulação proposta ao conseguir simulações que refletem, com grande precisão, os fenómenos físicos. Com as otimizações realizadas, conseguiu-se obter uma redução de cerca de 30% do tempo inicial cumprindo com os requisitos de correção e precisão necessários.
Resumo:
The Intel R Xeon PhiTM is the first processor based on Intel’s MIC (Many Integrated Cores) architecture. It is a co-processor specially tailored for data-parallel computations, whose basic architectural design is similar to the ones of GPUs (Graphics Processing Units), leveraging the use of many integrated low computational cores to perform parallel computations. The main novelty of the MIC architecture, relatively to GPUs, is its compatibility with the Intel x86 architecture. This enables the use of many of the tools commonly available for the parallel programming of x86-based architectures, which may lead to a smaller learning curve. However, programming the Xeon Phi still entails aspects intrinsic to accelerator-based computing, in general, and to the MIC architecture, in particular. In this thesis we advocate the use of algorithmic skeletons for programming the Xeon Phi. Algorithmic skeletons abstract the complexity inherent to parallel programming, hiding details such as resource management, parallel decomposition, inter-execution flow communication, thus removing these concerns from the programmer’s mind. In this context, the goal of the thesis is to lay the foundations for the development of a simple but powerful and efficient skeleton framework for the programming of the Xeon Phi processor. For this purpose we build upon Marrow, an existing framework for the orchestration of OpenCLTM computations in multi-GPU and CPU environments. We extend Marrow to execute both OpenCL and C++ parallel computations on the Xeon Phi. We evaluate the newly developed framework, several well-known benchmarks, like Saxpy and N-Body, will be used to compare, not only its performance to the existing framework when executing on the co-processor, but also to assess the performance on the Xeon Phi versus a multi-GPU environment.
Resumo:
A Digital Breast Tomosynthesis (DBT) é uma técnica que permite obter imagens mamárias 3D de alta qualidade, que só podem ser obtidas através de métodos de re-construção. Os métodos de reconstrução mais rápidos são os iterativos, sendo no en-tanto computacionalmente exigentes, necessitando de sofrer muitas optimizações. Exis-tem optimizações que usam computação paralela através da implementação em GPUs usando CUDA. Como é sabido, o desenvolvimento de programas eficientes que usam GPUs é ainda uma tarefa demorada, dado que os modelos de programação disponíveis são de baixo nível, e a portabilidade do código para outras arquitecturas não é imedia-ta. É uma mais valia poder criar programas paralelos de forma rápida, com possibili-dade de serem usados em diferentes arquitecturas, sem exigir muitos conhecimentos sobre a arquitectura subjacente e sobre os modelos de programação de baixo nível. Para resolver este problema, propomos a utilização de soluções existentes que reduzam o esforço de paralelização, permitindo a sua portabilidade, garantindo ao mesmo tempo um desempenho aceitável. Para tal, vamos utilizar um framework (FastFlow) com suporte para Algorithmic Skeletons, que tiram partido da programação paralela estruturada, capturando esquemas/padrões recorrentes que são comuns na programação paralela. O trabalho realizado centrou-se na paralelização de uma das fases de reconstru-ção da imagem 3D – geração da matriz de sistema – que é uma das mais demoradas do processo de reconstrução; esse trabalho incluiu um método de ordenação modificado em relação ao existente. Foram realizadas diferentes implementações em CPU e GPU (usando OpenMP, CUDA e FastFlow) o que permitiu comparar estes ambientes de programação em termos de facilidade de desenvolvimento e eficiência da solução. A comparação feita permite concluir que o desempenho das soluções baseadas no FastFlow não é muito diferente das tradicionais o que sugere que ferramentas deste tipo podem simplificar e agilizar a implementação de um algoritmos na área de recons-trução de imagens 3D, mantendo um bom desempenho.
Resumo:
A aceleração do ritmo de mudança verificado nas sociedades atuais, tem sido impulsionado pela globalização, fenómeno decorrente da evolução das tecnologias da informação, das telecomunicações, das comunicações e transportes e do desaparecimento de fronteiras. Viver na “aldeia global” ou à escala global como o previu Marshall McLuhan (1964) no livro Understanding Media, é hoje uma realidade inquestionável. As consequências desta transformação foram múltiplas quer do ponto de vista dos benefícios, quer do ponto de vista dos problemas gerados. No plano da segurança, face ao multiculturalismo envolvido e ao aumento crescente do crime transfronteiriço, tornou-se essencial a partilha de informação a nível internacional tendo em vista o seu combate não só olhando a situação dos cidadãos como a defesa dos princípios democráticos. Realça-se que os progressos tecnológicos e as facilidades que criam aos seus utilizadores, neste caso os criminosos, fazem com que as ações por estes praticadas sejam cada vez mais meticulosas, imprevisíveis, sofisticadas e complexas o que impõe uma resposta correspondente e adequada. Por essa razão, as políticas de segurança existentes mostraram-se insuficientes e esgotadas requerendo novas respostas capazes de produzir os efeitos desejáveis para uma efetiva prevenção da criminalidade. Pelas dimensões que tem vindo a tomar, a criminalidade tornou-se uma preocupação que ultrapassou o domínio da segurança interna de cada país para ser encarada a nível internacional ou mesmo mundial. Para o efeito urge concertar processos e procedimentos securitários agregando vontades que convirjam e defendam a unificação dos sistemas dos países a nível mundial. A verificar-se tal intento, daí resultariam significativas melhorias da segurança a todos os níveis (nacional, internacional e mundial). Além disso também resultariam ganhos em termos de tempo, redução de custos, impacto na qualidade dos serviços prestados, na gestão das pessoas e na eficiência das organizações. Realça-se que a democracia ao promover a dignidade do homem densificando os seus direitos, liberdades e garantias, criou indiretamente condições para que fosse gerada instabilidade e o desenvolvimento de comportamentos criminosos. Importa, portanto, face à situação existente e à previsível complexidade do crime no futuro, estudar profundamente a nova realidade neste domínio, para tomar as medidas preventivas tendentes a reporem a estabilidade e a promoverem a paz social. Foi neste contexto que a presente investigação, desenvolvida no âmbito académico, mas também suportado na realidade profissional, pretendeu refletir sobre o estado da segurança global e dar o seu contributo nesta matéria.
Resumo:
Dissertação de mestrado integrado em Engenharia e Gestão Industrial
Resumo:
The influence of the hip joint formulation on the kinematic response of the model of human gait is investigated throughout this work. To accomplish this goal, the fundamental issues of the modeling process of a planar hip joint under the framework of multibody systems are revisited. In particular, the formulations for the ideal, dry, and lubricated revolute joints are described and utilized for the interaction of femur head inside acetabulum or the hip bone. In this process, the main kinematic and dynamic aspects of hip joints are analyzed. In a simple manner, the forces that are generated during human gait, for both dry and lubricated hip joint models, are computed in terms of the system’s state variables and subsequently introduced into the dynamics equations of motion of the multibody system as external generalized forces. Moreover, a human multibody model is considered, which incorporates the different approaches for the hip articulation, namely ideal joint, dry, and lubricated models. Finally, several computational simulations based on different approaches are performed, and the main results presented and compared to identify differences among the methodologies and procedures adopted in this work. The input conditions to the models correspond to the experimental data capture from an adult male during normal gait. In general, the obtained results in terms of positions do not differ significantly when the different hip joint models are considered. In sharp contrast, the velocity and acceleration plotted vary significantly. The effect of the hip joint modeling approach is clearly measurable and visible in terms of peaks and oscillations of the velocities and accelerations. In general, with the dry hip model, intra-joint force peaks can be observed, which can be associated with the multiple impacts between the femur head and the cup. In turn, when the lubricant is present, the system’s response tends to be smoother due to the damping effects of the synovial fluid.
Resumo:
Tese de Doutoramento (Programa Doutoral em Engenharia Biomédica)
Resumo:
Tese de Doutoramento Plano Doutoral em Engenharia Eletrónica e de Computadores.
Resumo:
This chapter presents a general methodology for the formulation of the kinematic constraint equations at position, velocity and acceleration levels. Also a brief characterization of the different type of constraints is offered, namely the holonomic and nonholonomic constraints. The kinematic constraints described here are formulated using generalized coordinates. The chapter ends with a general approach to deal with the kinematic analysis of multibody systems.
Resumo:
"Series title: Springerbriefs in applied sciences and technology, ISSN 2191-530X"
Resumo:
Background:In chronic Chagas disease (ChD), impairment of cardiac autonomic function bears prognostic implications. Phase‑rectification of RR-interval series isolates the sympathetic, acceleration phase (AC) and parasympathetic, deceleration phase (DC) influences on cardiac autonomic modulation.Objective:This study investigated heart rate variability (HRV) as a function of RR-interval to assess autonomic function in healthy and ChD subjects.Methods:Control (n = 20) and ChD (n = 20) groups were studied. All underwent 60-min head-up tilt table test under ECG recording. Histogram of RR-interval series was calculated, with 100 ms class, ranging from 600–1100 ms. In each class, mean RR-intervals (MNN) and root-mean-squared difference (RMSNN) of consecutive normal RR-intervals that suited a particular class were calculated. Average of all RMSNN values in each class was analyzed as function of MNN, in the whole series (RMSNNT), and in AC (RMSNNAC) and DC (RMSNNDC) phases. Slopes of linear regression lines were compared between groups using Student t-test. Correlation coefficients were tested before comparisons. RMSNN was log-transformed. (α < 0.05).Results:Correlation coefficient was significant in all regressions (p < 0.05). In the control group, RMSNNT, RMSNNAC, and RMSNNDCsignificantly increased linearly with MNN (p < 0.05). In ChD, only RMSNNAC showed significant increase as a function of MNN, whereas RMSNNT and RMSNNDC did not.Conclusion:HRV increases in proportion with the RR-interval in healthy subjects. This behavior is lost in ChD, particularly in the DC phase, indicating cardiac vagal incompetence.
Resumo:
Abstract Background: Right-sided heart failure has high morbidity and mortality, and may be caused by pulmonary arterial hypertension. Fractal dimension is a differentiated and innovative method used in histological evaluations that allows the characterization of irregular and complex structures and the quantification of structural tissue changes. Objective: To assess the use of fractal dimension in cardiomyocytes of rats with monocrotaline-induced pulmonary arterial hypertension, in addition to providing histological and functional analysis. Methods: Male Wistar rats were divided into 2 groups: control (C; n = 8) and monocrotaline-induced pulmonary arterial hypertension (M; n = 8). Five weeks after pulmonary arterial hypertension induction with monocrotaline, echocardiography was performed and the animals were euthanized. The heart was dissected, the ventricles weighed to assess anatomical parameters, and histological slides were prepared and stained with hematoxylin/eosin for fractal dimension analysis, performed using box-counting method. Data normality was tested (Shapiro-Wilk test), and the groups were compared with non-paired Student t test or Mann Whitney test (p < 0.05). Results: Higher fractal dimension values were observed in group M as compared to group C (1.39 ± 0.05 vs. 1.37 ± 0.04; p < 0.05). Echocardiography showed lower pulmonary artery flow velocity, pulmonary acceleration time and ejection time values in group M, suggesting function worsening in those animals. Conclusion: The changes observed confirm pulmonary-arterial-hypertension-induced cardiac dysfunction, and point to fractal dimension as an effective method to evaluate cardiac morphological changes induced by ventricular dysfunction.
Resumo:
Mangoes, cv. Imperial, were exposed, in post harvest, to the following methods of ripening: 1) sawdust burning; 2) alcohol vaporization; 3) calcium carbide (acetylene), 4) vapour of ethylene; and, 5) immersion in ethefon. All methods resulted in acceleration of ripening, when compared to controls. Calcium carbide, ethelene and ethefon were the best, methods. Alcohol vaporization also showed good results sawdust burning method showing low efficiency.
Resumo:
Summary : The hypothalamus represents less than 1 % of the total volume of the brain tissue, yet it plays a crucial role in endocrine regulations. Puberty is defined as a process leading to physical, sexual and psychosocial maturation. The hypothalamus is central to this process, via the activation of GnRH neurons. Pulsatile GnRH secretion, minimal during childhood, increases with the onset of puberty. The primary function of GnRH is to regulate the growth, development and function of testes in boys and ovaries in girls, by stimulating the pituitary gland secretion of luteinizing hormone (LH) and follicle-stimulating hormone (FSH). Several factors contribute to the timing of puberty, including sex and ethnicity, genetics, dietary intake and energy expenditure. Kisspeptins constitute a family of small peptides arising from the proteolytic cleavage of metastin, a peptide with 54 amino acids initially purified from human placenta. These kisspeptins were the subject of much attention following their discovery because of their antimetastatic properties, but it was more recently that their determining role in the reproductive function was demonstrated. It was shown that kisspeptins are ligands of a receptor, GPR54, whose natural inactivating mutation in humans, or knockout in the mouse, lead to infertility. GnRH neurons play a pivotal role in the central regulation of fertility. Kisspeptin greatly increases GnRH release and GnRH neuron firing activity, but the neurobiological mechanisms for these actions are unknown. Gprotein-coupled receptor 54, the receptor for kisspeptin, is expressed by GnRH neurons as well as other hypothalamic neurons, suggesting that both direct and indirect effects are possible. In the first part of my thesis, we investigated a possible connection between the acceleration of sexual development induced by leptin and hypothalamic metastin neurons. However, the data generated by our preliminary experiments confirmed that the commercially available antibodies are non-specific. This finding constituted a major drawback for our studies, which relied heavily upon the neuroanatomical study of the hypothalamic metastinergic pathways to elucidate their sensitivity to exogenous leptin. Therefore, we decided to postpone any further in vivo experiment until a better antibody becomes available, and focused on in vitro studies to better understand the mechanisms of action of kisspeptins in the modulation of the activity of GnRH neurons. We used two GnRH-expressing neuronal cell lines to investigate the cellular and molecular mechanisms of action of metastin in GnRH neurons. We demonstrated that kisspeptin induces an early activation of the MAP kinase intracellular signaling pathway in both cell lines, whereas the SAP/JNK or the Akt pathways were unaffected. Moreover, we found an increase in GnRH mRNA levels after 6h of metastin stimulation. Thus, we can conclude that kisspeptin regulates GnRH neurons both at the secretion and the gene expression levels. The MAPK pathway is the major pathway activated by metastin in GnRH expressing neurons. Taken together, these data provide the first mechanism of action of kisspeptin on GnRH neurons. Résumé : L'hypothalamus est une zone située au centre du cerveau, dont il représente moins de 1 du volume total. La puberté est la période de transition entre l'enfance et l'age adulte, qui s'accompagne de transformations somatiques, psychologiques, métaboliques et hormonales conduisant à la possibilité de procréer. La fonction principale de la GnRH est la régulation de la croissance, du développement et de la fonction des testicules chez les hommes, et des ovaires chez les femmes en stimulant la sécrétion de l'hormone lutéinisante (LH) et de l'hormone folliculostimulante (FSH) par la glande hypophysaire. Plusieurs facteurs contribuent au déclanchement de la puberté, y compris le sexe et l'appartenance ethnique, la génétique, l'apport alimentaire et la dépense énergétique. Les Kisspeptines constituent une famille de peptides résultant de la dissociation proteolytique de la métastine, un peptide de 54 acides aminés initialement purifié à partir de placenta humain. Ces kisspeptines ont fait l'objet de beaucoup d'attention à la suite de leur découverte en raison de leurs propriétés anti-metastatiques, et c'est plus récemment que leur rôle déterminant dans la fonction reproductive a été démontré. Les kisspeptines sont des ligands du récepteur GPR54, dont la mutation inactivatrice chez l'homme, ou le knockout chez la souris, conduisent à l'infertilité par hypogonadisme hypogonadotrope. Les neurones à GnRH jouent un rôle central dans le règlement des fonctions reproductrices et la kisspeptine stimule l'activité des neurones à GnRH et la libération de GnRH par ces neurones. Toutefois, les mécanismes neurobiologiques de ces actions ne sont pas connus. Dans la première partie de ma thèse, nous avons étudié le lien potentiel entre l'accélération du développement sexuel induite par la leptine et les neurones hypothalamiques à metastine. Les données générées dans cette première série d'expériences ont malheureusement confirmé que les anticorps anti-metastine disponibles dans le commerce sont aspécifiques. Ceci a constitué un inconvénient majeur pour nos études, qui devaient fortement s'appuyer sur l' étude neuroanatomique des neurones hypothalamiques à metastine pour évaluer leur sensibilité à la leptine exogène. Nous avons donc décidé de focaliser nos travaux sur une étude in vitro des mécanismes d'action de la kisspeptine pour moduler l'activité des neurones à GnRH. Nous avons utilisé deux lignées de cellules neuronales exprimant la GnRH pour étudier les mécanismes d'action cellulaires et moléculaires de la metastine dans des neurones. Nous avons ainsi pu démontrer que la kisspeptine induit une activation précoce de la voie f de signalisation de la MAP kinase dans les deux lignées cellulaires, alors que nous n'avons observé aucune activation de la voie de signalisation de la P13 Kinase et de la SAP/JNK. Nous avons en outre démontré une augmentation de l'expression de la GnRH par la stimulation avec la Kisspeptine. L'ensemble de ces données contribue à élucider le mécanisme d'action avec lequel la kisspeptine agit dans les neurones à GnRH, en démontrant un effet sur l'expression génique de la GnRH. Nous pouvons également conclure que la voie de la MAPK est la voie principale activée par la metastine dans les neurones exprimant la GnRH.
Resumo:
PURPOSE: To introduce a new k-space traversal strategy for segmented three-dimensional echo planar imaging (3D EPI) that encodes two partitions per radiofrequency excitation, effectively reducing the number excitations used to acquire a 3D EPI dataset by half. METHODS: The strategy was evaluated in the context of functional MRI applications for: image quality compared with segmented 3D EPI, temporal signal-to-noise ratio (tSNR) (the ability to detect resting state networks compared with multislice two-dimensional (2D) EPI and segmented 3D EPI, and temporal resolution (the ability to separate cardiac- and respiration-related fluctuations from the desired blood oxygen level-dependent signal of interest). RESULTS: Whole brain images with a nominal voxel size of 2 mm isotropic could be acquired with a temporal resolution under half a second using traditional parallel imaging acceleration up to 4× in the partition-encode direction and using novel data acquisition speed-up of 2× with a 32-channel coil. With 8× data acquisition speed-up in the partition-encode direction, 3D reduced excitations (RE)-EPI produced acceptable image quality without introduction of noticeable additional artifacts. Due to increased tSNR and better characterization of physiological fluctuations, the new strategy allowed detection of more resting state networks compared with multislice 2D-EPI and segmented 3D EPI. CONCLUSION: 3D RE-EPI resulted in significant increases in temporal resolution for whole brain acquisitions and in improved physiological noise characterization compared with 2D-EPI and segmented 3D EPI. Magn Reson Med 72:786-792, 2014. © 2013 Wiley Periodicals, Inc.