Call number: LD2668 .T4 1956 K31 / Master of Science

No description available.

Multi-core processors are now ubiquitous and are widely seen as the most viable means of delivering performance with increasing transistor densities. However, this potential can only be realised if the application programs are suitably parallel. Applications can either be written in parallel from scratch or converted from existing sequential programs. Regardless of how applications are parallelised, the code must be efficiently mapped onto the underlying platform to fully exploit the hardware’s potential. This thesis addresses the problem of finding the best mappings of data and streaming parallelism—two types of parallelism that exist in broad and important domains such as scientific, signal processing and media applications. Despite significant progress having been made over the past few decades, state-of-the-art mapping approaches still largely rely upon hand-crafted, architecture-specific heuristics. Developing a heuristic by hand, however, often requiresmonths of development time. Asmulticore designs become increasingly diverse and complex, manually tuning a heuristic for a wide range of architectures is no longer feasible. What are needed are innovative techniques that can automatically scale with advances in multi-core technologies. In this thesis two distinct areas of computer science, namely parallel compiler design and machine learning, are brought together to develop new compiler-based mapping techniques. Using machine learning, it is possible to automatically build highquality mapping schemes, which adapt to evolving architectures, with little human involvement. First, two techniques are proposed to find the best mapping of data parallelism. The first technique predicts whether parallel execution of a data parallel candidate is profitable on the underlying architecture. On a typical multi-core platform, it achieves almost the same (and sometimes a better) level of performance when compared to the manually parallelised code developed by independent experts. For a profitable candidate, the second technique predicts how many threads should be used to execute the candidate across different program inputs. The second technique achieves, on average, over 96% of the maximum available performance on two different multi-core platforms. Next, a new approach is developed for partitioning stream applications. This approach predicts the ideal partitioning structure for a given stream application. Based on the prediction, a compiler can rapidly search the program space (without executing any code) to generate a good partition. It achieves, on average, a 1.90x speedup over the already tuned partitioning scheme of a state-of-the-art streaming compiler.

O tema da presente dissertação é a noção de Equivalência (Äquivalenz) de Werner Koller. Frequentemente atrelado às abordagens linguísticas à tradução, que, na segunda metade do século XX, buscaram conferir à pesquisa em tradução o estatuto de ciência, o conceito de equivalência foi alçado à condição de princípio teórico que garantiria a objetividade exigida das disciplinas que aspirassem à cientificidade. Para isso, porém, adotouse uma concepção abstrata e artificial de equivalência, que relegava o tradutor ao lugar de mero reprodutor de conteúdos pré-fixados pelo autor do original. Com o surgimento de novas vertentes de pesquisa no âmbito dos estudos de tradução, tais pressupostos foram postos em xeque, o que levou ao descarte e à estigmatização dessa noção. Se tomarmos por base a oposição elucidada por Kopetzki (1996), tal virada nos estudos de tradução pode ser interpretada como um movimento de concepções universalistas para concepções relativistas de língua e tradução, que rejeitam a ideia de tradução como transporte de significados imobilizados a priori. Diante disso, o objetivo desta pesquisa é demonstrar que a Equivalência de Koller não está calcada no universalismo que serve de base a outros modelos de equivalência tradutória, e que a noção defendida pelo autor tem validade prática, na medida em que funciona como um conceito operacional para tradutores profissionais e estudiosos da tradução. O presente trabalho se divide em duas grandes partes: a primeira é de caráter teórico e se destina à apresentação e à análise de trechos selecionados da obra de Koller, ainda inédita em português, a fim de evidenciar o caráter antiuniversalista de seus pressupostos; na segunda parte, mais voltada à prática, proponho uma maneira de aplicar seu conceito de Equivalência a um par de textos empíricos. O corpus que serve de base ao exemplo de aplicação é composto das ocorrências de cores presentes no romance Buddenbrooks: Verfall einer Familie (1909), de Thomas Mann, e na sua tradução para o português, de autoria de Herbert Caro e publicada no Brasil sob o título Os Buddenbrook: decadência duma família (1942). / The topic of this dissertation is Werner Kollers notion of equivalence (Äquivalenz). Often associated with the linguistic approaches to translation, which, in the second half of the 20th century, sought to make a scientific discipline out of the subject, equivalence turned into the one theoretical principle that would presumably guarantee the objectivity expected of any aspiring science. The consequence, however, was an abstract and artificial concept of equivalence, which portrayed translators as mere reproducers of contents that had been predetermined by the original author. With the advent of new research branches in the field of Translation Studies, such assumptions were called into question, which led to the discard and stigmatization of the concept. Bearing in mind the opposition brought to light by Kopetzki (1996), such a turn in Translation Studies may be interpreted as a move away from universalistic conceptions and a step towards relativistic conceptions of language and translation, which reject the idea of translation as the transport of pre-existing meanings. In view of that, the aim of this research is to demonstrate that Kollers Äquivalenz is not based upon the universalism that characterizes other theoretical frameworks to translation equivalence and can therefore be considered of practical relevance, inasmuch as it functions as an operational concept to professional translators and translation scholars. The present dissertation is divided in two main parts: the first part is of theoretical nature and is dedicated to the presentation and analysis of selected passages from Kollers work, still unedited in Portuguese, in order to demonstrate the antiuniversalistic character of his premises; in the second, more practice-oriented part, I propose a way of applying his concept of Equivalence to a pair of empirical texts. The corpus upon which this application example is based consists of instances of colour names found in the novel Buddenbrooks: Verfall einer Familie (1909), by Thomas Mann, and its translation to Portuguese, written by Herbert Caro and published in Brazil under the title Os Buddenbrook: decadência duma família (1942).

Title also in Chinese. / Thesis (M.Phil.)--Chinese University of Hong Kong. / Bibliography: leaves 159-161.

O tema da presente dissertação é a noção de Equivalência (Äquivalenz) de Werner Koller. Frequentemente atrelado às abordagens linguísticas à tradução, que, na segunda metade do século XX, buscaram conferir à pesquisa em tradução o estatuto de ciência, o conceito de equivalência foi alçado à condição de princípio teórico que garantiria a objetividade exigida das disciplinas que aspirassem à cientificidade. Para isso, porém, adotouse uma concepção abstrata e artificial de equivalência, que relegava o tradutor ao lugar de mero reprodutor de conteúdos pré-fixados pelo autor do original. Com o surgimento de novas vertentes de pesquisa no âmbito dos estudos de tradução, tais pressupostos foram postos em xeque, o que levou ao descarte e à estigmatização dessa noção. Se tomarmos por base a oposição elucidada por Kopetzki (1996), tal virada nos estudos de tradução pode ser interpretada como um movimento de concepções universalistas para concepções relativistas de língua e tradução, que rejeitam a ideia de tradução como transporte de significados imobilizados a priori. Diante disso, o objetivo desta pesquisa é demonstrar que a Equivalência de Koller não está calcada no universalismo que serve de base a outros modelos de equivalência tradutória, e que a noção defendida pelo autor tem validade prática, na medida em que funciona como um conceito operacional para tradutores profissionais e estudiosos da tradução. O presente trabalho se divide em duas grandes partes: a primeira é de caráter teórico e se destina à apresentação e à análise de trechos selecionados da obra de Koller, ainda inédita em português, a fim de evidenciar o caráter antiuniversalista de seus pressupostos; na segunda parte, mais voltada à prática, proponho uma maneira de aplicar seu conceito de Equivalência a um par de textos empíricos. O corpus que serve de base ao exemplo de aplicação é composto das ocorrências de cores presentes no romance Buddenbrooks: Verfall einer Familie (1909), de Thomas Mann, e na sua tradução para o português, de autoria de Herbert Caro e publicada no Brasil sob o título Os Buddenbrook: decadência duma família (1942). / The topic of this dissertation is Werner Kollers notion of equivalence (Äquivalenz). Often associated with the linguistic approaches to translation, which, in the second half of the 20th century, sought to make a scientific discipline out of the subject, equivalence turned into the one theoretical principle that would presumably guarantee the objectivity expected of any aspiring science. The consequence, however, was an abstract and artificial concept of equivalence, which portrayed translators as mere reproducers of contents that had been predetermined by the original author. With the advent of new research branches in the field of Translation Studies, such assumptions were called into question, which led to the discard and stigmatization of the concept. Bearing in mind the opposition brought to light by Kopetzki (1996), such a turn in Translation Studies may be interpreted as a move away from universalistic conceptions and a step towards relativistic conceptions of language and translation, which reject the idea of translation as the transport of pre-existing meanings. In view of that, the aim of this research is to demonstrate that Kollers Äquivalenz is not based upon the universalism that characterizes other theoretical frameworks to translation equivalence and can therefore be considered of practical relevance, inasmuch as it functions as an operational concept to professional translators and translation scholars. The present dissertation is divided in two main parts: the first part is of theoretical nature and is dedicated to the presentation and analysis of selected passages from Kollers work, still unedited in Portuguese, in order to demonstrate the antiuniversalistic character of his premises; in the second, more practice-oriented part, I propose a way of applying his concept of Equivalence to a pair of empirical texts. The corpus upon which this application example is based consists of instances of colour names found in the novel Buddenbrooks: Verfall einer Familie (1909), by Thomas Mann, and its translation to Portuguese, written by Herbert Caro and published in Brazil under the title Os Buddenbrook: decadência duma família (1942).

Les étoiles sont des éléments fondamentaux de l'Univers. Elles émettent de l'énergie en forme de lumières et rendent les matériaux dans le ciel visible. Les étoiles se regroupent pour former les galaxies, en déterminant l'évolution et la dynamique de ce dernier. En même temps, l'étoile est le centre d'un système planétaire. Le disque de débris autour d'une jeune étoile se refroidi et forme un système de planète. Les caractéristiques de ce système, notamment la masse de l'étoile centrale, jouent un rôle important en ce qui concerne l'apparition de la vie. Cette thèse a pour objectif de comprendre comment la massed'une étoile est assemblée et déterminée, donnant une distribution de masse apparemmentuniverselle quel que soit l'environnement de leur formation..La thèse est constituée de deux chapitres introductifs sur la physique de formation stellaire et sur les méthodes numériques. Les trois chapitres suivants sont constitués des projets menés durant la thèse: la formation des proto-amas, l'effet de condition initiale dans le nuage moléculaire, et la formation des coeurs préstellaires par la fragmentation des filaments, suivis par les articles publiés dans les journaux scientifiques. Le dernier chapitre conclu la thèse et donne les perspectifs pour la future recherche / Stars are building blocks of the Universe. They emit energy in form of light and make the material in the night sky visible. They are the elementary constituents of galaxies, determining their evolution and dynamics. On the other hand, stars are the hosts o planetary systems. The debris disc around a new-born star eventually cools down and form planets. The characteristic of the planetary system, essentially the mass of the central star, plays a major role in the formation of living being on planets. The formation of stars often occur in a clusters manner, and one of the important issues constantly under debate is the distribution of the mass of newly-born stars. This thesis is aimed to understand the Initial Mass Function which seems to be universal among different environments.This manuscripts comprises two introductory chapters on the physics of star formation and the numerical methods, respectively. Three following chapters present the projets carried out during the thesis: formation of proto-clusters, effects of initial condition in the molecular cloud, and the formation of prestellar cores from filament fragmentation, all followed by published journal articles. The last chapter concludes the manuscript and discuss the perspectives

Mineral dust both influences and is influenced by climate on many timescales, from seconds to epochs. Its complex interactions with the climate system are still being unraveled. For example, dust fluxes change in tandem with other records of past changes in climate, and dust source is often presumed to change as well, in response to shifts in climate conditions in source regions; changes in wind regimes; or changes in atmospheric transport pathways. In this work, I investigate dust records from the Southern Hemisphere from ice core and marine sediment core climate archives, looking at both flux and provenance in order elucidate the conditions that allowed for those particles to travel from source to sink. Using multiple radiogenic isotope systems as tracers (87Sr/86Sr, εNd(0), 206Pb/207Pb, and 208Pb/207Pb), I geochemically “fingerprint” of dust particles from Southern Hemisphere climate archives over the last glacial cycle. I compare the dust fingerprints to potential source areas (PSA’s) from across the Southern Hemisphere in order to identify the sources of dust found in the WAIS Divide and Taylor Glacier ice cores from West Antarctica, as well as from marine sediment core ELT39.75 in the Tasman Sea. I use endmember mixing theory to determine the relative contribution of different sources to the climate archives over time. In West Antarctica, I geochemically identify specific local volcanoes from Marie Byrd Land as significant particle contributors to the WAIS Divide ice core during the previous glacial period. In the Tasman Sea, I identify a specific region of southeastern Australia as primary the dust source over the past glacial cycle, with the source remaining constant across glacial-interglacial climate transitions. This clarifies that the “fingerprint” of Australian dust is relatively invariant over time and allows a single Australian signature to be used as an endmember for identifying dust provenance in climate archives downwind. I also identify the dust sources in the WAIS Divide during the Last Glacial Maximum and through the early deglacial, identifying southern South America as the predominant source during cold stages. WAIS Divide and Taylor Glacier dust records do not record dust source changes across millennial-scale climate events, suggesting that a) the source regions did not change, b) the transport pathways remained pinned, or c) the proxy is not sensitive to changes in these variables. Contributions from local volcanoes are also inferred from the WAIS dust record using mixing theory. In summary, I find that the radiogenic isotope fingerprint of dust samples from the archives analyzed show subtle or no changes in source over climate transitions, and therefore the strategy of dust particles as a tracer of past atmospheric circulation pathways should be approached cautiously.

This work is an investigation into the curvature related potential for flux to flow in the radial direction in the back-iron of laminated axial flux machine cores. Analytical and numerical models are presented. Analysis based on these models has shown that, in practical axial flux machines, the radial component of the flux density can be neglected with respect to the flux density distribution in the core back-iron. It has also been found that if the core permeability, core conductivity and number of poles are sufficiently high then power loss due to curvature related cross-lamination flux is negligible compared to normal eddy current losses. A closed form expression to predict losses due to curvature related radial flux is also presented. This expression allows axial flux machine designers to make quick assessment of the need to consider these losses when designing axial flux machines.

<p>This report documents the work done by the author to design and implement a 3D graphics system on an FPGA (Field Programmable Gate Array). After a preamble with a background presentation to the project, a very brief introduction in computer graphics techniques and computer graphics theory is given. Then, the hardware available to the project, along with an analysis of general requirements is examined. The following chapter contains the proposed graphics system design for FPGA implementation. A broad approach to separate the design and the eventual implementation was used. Two 3D pipelines are suggested - one fully capable high-end version and one which use minimal resources. The documentation of the effort to implement the minimal graphics system previously discussed then follows. The documentation outlines the work done without going too deep into detail, and is followed by the largest of the tests conducted. Finally, chapter seven concludes the project with the most important project conclusions and some suggestions for future work.</p>