Field-Programmable Gate Arrays (FPGAs) systems now comprise many processing elements that are processors running software and hardware engines used to accelerate specific functions. To make the programming of such a system simpler, it is easiest to think of a shared-memory environment, much like in current multi-core processor systems. This thesis introduces a novel, shared-memory, cache-coherent infrastructure for heterogeneous systems implemented on FPGAs that can then form the basis of a shared-memory programming model for heterogeneous systems. With simulation results, it is shown that the cache-coherent infrastructure outperforms the infrastructure of Woods [1] with a speedup of 1.10. The thesis explores the various configurations of the cache interconnection network and the benefit of the cache-to-cache cache line data transfer with its impact on main memory access. Finally, the thesis shows the cache-coherent infrastructure has very little overhead when using its cache coherence implementation.

Orientador: Mario Luiz Teixeira de Moraes / Resumo: O objetivo principal do trabalho foi o de quantificar a variabilidade genética em progênies de seringueira fundamentando-se na avaliação de caracteres quantitativos e na caracterização molecular do tipo microssatélites (SSR). A primeira população do estudo é originária da floresta primária de Rio Branco- Acre (população selvagem-PS), e a outra, trata-se de uma população, originada de matrizes clonais (população melhorada-PM). Encontram-se instaladas na forma de teste de progênies na Fazenda de Ensino, Pesquisa e Extensão da Faculdade de Engenharia de Ilha Solteira/UNESP (FEPE), localizada em Selvíria, MS. Para as duas populações foram avaliados os seguintes caracteres silviculturais de crescimento, altura (ALT), altura comercial (AC), diâmetro médio de copa (DMC), forma do fuste (FOR), perímetro do caule (PAP e P50) e produção de borracha seca (PBS), aos oito (PM) e 23 (PS) anos de idade. O delineamento experimental utilizado foi o de blocos casualizados, compostos por 31 famílias, quatro repetições e parcelas lineares de 10 plantas, no espaçamento de 3 x 3 m (PM). Para PS o delineamento experimental foi de blocos causalizados com 37 famílias distribuídas em três repetições, de forma desbalanceada com no máximo 10 plantas por progênies no espaçamento de 5 x 3 m (PS). As estimativas dos parâmetros genéticos foram feitas utilizando-se a metodologia de modelo linear misto univariado aditivo REML/BLUP e ganhos na seleção pelo método índice multiefeitos (IME). O DNA genômico foi e... (Resumo completo, clicar acesso eletrônico abaixo) / Abstract: The main objective of this work is to quantify the genetic variability in progenies of rubber tree, based on the quantitative characterization and the molecular characterization of the microsatellite type (SSR). The first population of the study originated in the primary forest of Rio Branco-Acre (wild population-PS), and the other population is a population originated from clonal matrices (improved population-PM). They are installed as progeny tests at the Teaching, Research and Extension Farm, That belongs to Engeneering College of the Julio Mesquita Filho State University of São Paulo, In Selvira, State of South Mato Grosso Brazil. For the two populations, the following silvicultural characteristics of growth, height (ALT), commercial height (AC), average crown diameter (DMC), stem shape (FOR), stem perimeter (PAP and P50) and dry rubber yield (PBS) at eight (PM) and 23 (PS) years old. The experimental design consisted of randomized blocks, composed of 31 families, four replications and linear plots of 10 plants, spaced 3 x 3 m (MP). For PS the experimental design was of causalized blocks with 37 families distributed in three replications, unbalanced with a maximum of 10 plants per progeny in the spacing of 5 x 3 m (PS). Estimates of the genetic parameters were made using the mixed linear univariate model (REML / BLUP) methodology and gains in selection by multi-effects index (MEI). Genomic DNA was extracted, quantified and genotyped for the two study populations. The anal... (Complete abstract click electronic access below) / Doutor

Submitted by Cecília Luzia Dourado (douradocl.cl@gmail.com) on 2017-05-23T14:32:39Z No. of bitstreams: 1 Cecília_Dourado_Tese_final.pdf: 2809054 bytes, checksum: 3b7bae6fc7d04f1ab484508fc4a7496d (MD5) / Approved for entry into archive by Luiz Galeffi (luizgaleffi@gmail.com) on 2017-05-23T18:22:13Z (GMT) No. of bitstreams: 1 dourado_cl_dr_ilha.pdf: 2809054 bytes, checksum: 3b7bae6fc7d04f1ab484508fc4a7496d (MD5) / Made available in DSpace on 2017-05-23T18:22:13Z (GMT). No. of bitstreams: 1 dourado_cl_dr_ilha.pdf: 2809054 bytes, checksum: 3b7bae6fc7d04f1ab484508fc4a7496d (MD5) Previous issue date: 2016-11-10 / Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) / O objetivo principal do trabalho foi o de quantificar a variabilidade genética em progênies de seringueira fundamentando-se na avaliação de caracteres quantitativos e na caracterização molecular do tipo microssatélites (SSR). A primeira população do estudo é originária da floresta primária de Rio Branco- Acre (população selvagem-PS), e a outra, trata-se de uma população, originada de matrizes clonais (população melhorada-PM). Encontram-se instaladas na forma de teste de progênies na Fazenda de Ensino, Pesquisa e Extensão da Faculdade de Engenharia de Ilha Solteira/UNESP (FEPE), localizada em Selvíria, MS. Para as duas populações foram avaliados os seguintes caracteres silviculturais de crescimento, altura (ALT), altura comercial (AC), diâmetro médio de copa (DMC), forma do fuste (FOR), perímetro do caule (PAP e P50) e produção de borracha seca (PBS), aos oito (PM) e 23 (PS) anos de idade. O delineamento experimental utilizado foi o de blocos casualizados, compostos por 31 famílias, quatro repetições e parcelas lineares de 10 plantas, no espaçamento de 3 x 3 m (PM). Para PS o delineamento experimental foi de blocos causalizados com 37 famílias distribuídas em três repetições, de forma desbalanceada com no máximo 10 plantas por progênies no espaçamento de 5 x 3 m (PS). As estimativas dos parâmetros genéticos foram feitas utilizando-se a metodologia de modelo linear misto univariado aditivo REML/BLUP e ganhos na seleção pelo método índice multiefeitos (IME). O DNA genômico foi extraído, quantificado e genotipado para as duas populações de estudo. As análises do sistema de reprodução e diversidade foram baseadas no modelo misto de reprodução e modelo de cruzamentos correlacionados. Foram detectadas diferenças significativas pelo teste da razão de verossimilhança na análise de deviance das procedências para os caracteres, diâmetro médio de copa (DMC), perímetro a 50 cm do solo (P50), perímetro a 1,30m do solo (PAP) e produção de borracha seca (PBS), para PM. Para população selvagem foram significativos, os caracteres ALT, DMC, P50 e PAP. Os caracteres que apresentaram maior magnitude para as herdabilidades foram a PBS, para população melhorada e PAP, PBS para a população selvagem, com herdabilidades acima de 60%. As estratégias de seleção de 50%, 40% e 22% dos indivíduos para o caractere PBS e PAP utilizando o índice multiefeitos revelaram de altos e baixos ganhos na seleção as duas populações. Para PM foi mais indicado a estratégia de seleção entre e dentro e para PS a seleção individual. Os ganhos obtidos na seleção foram de 54% para o caráter PBS na população melhorada e de 0,46% para o caráter PAP para população selvagem. A heterozigosidade observada foi de 0,839 a 0,747 para adultos e de 0,425 a 0,399 para as progênies, para as duas populações estudas. A taxa de cruzamento multiloco ( ) variou de 0,726 (PM) a 0,798 (PS), indicando que grande parte das sementes foram originadas por cruzamentos, mas houve presença de autofecundação, caracterizado por um sistema de reprodução misto. A endogamia apresentada foi gerada por cruzamentos correlacionados e não por autofecundação. Devido ao ao tamanho efetivo (Ne) menor que 4, resultou na necessidade de coletar-se sementes para fins de conservação genética, recuperação ambiental e melhoramento florestal de pelo menos 94 e 89 árvores, para PM e PS respectivamente. / The main objective of this work is to quantify the genetic variability in progenies of rubber tree, based on the quantitative characterization and the molecular characterization of the microsatellite type (SSR). The first population of the study originated in the primary forest of Rio Branco-Acre (wild population-PS), and the other population is a population originated from clonal matrices (improved population-PM). They are installed as progeny tests at the Teaching, Research and Extension Farm, That belongs to Engeneering College of the Julio Mesquita Filho State University of São Paulo, In Selvira, State of South Mato Grosso Brazil. For the two populations, the following silvicultural characteristics of growth, height (ALT), commercial height (AC), average crown diameter (DMC), stem shape (FOR), stem perimeter (PAP and P50) and dry rubber yield (PBS) at eight (PM) and 23 (PS) years old. The experimental design consisted of randomized blocks, composed of 31 families, four replications and linear plots of 10 plants, spaced 3 x 3 m (MP). For PS the experimental design was of causalized blocks with 37 families distributed in three replications, unbalanced with a maximum of 10 plants per progeny in the spacing of 5 x 3 m (PS). Estimates of the genetic parameters were made using the mixed linear univariate model (REML / BLUP) methodology and gains in selection by multi-effects index (MEI). Genomic DNA was extracted, quantified and genotyped for the two study populations. The analyzes of the breeding and diversity system were based on the mixed breeding model and correlated crosses model. Significant differences were detected by the likelihood ratio test in the deviance analysis for the characters, mean crown diameter (DMC), stem perimeter at 50 cm from soil (P50), stem perimeter at 1,30m from soil (PAP) and dry rubber yield (PBS) for PM. For wild populations were significant, the characters ALT, DMC, P50 and PAP. The characters that presented the greatest magnitude for the heritabilities were PBS, for improved population and PAP, PBS for the wild population, with heritabilities above 60%. The selection strategies of 50%, 40% and 22% of individuals for the PBS and PAP using the multi-effects index revealed high and low gains in the selection of the two populations. For PM it was more indicated the strategy of selection between and within and for PS the individual selection. The gains obtained in the selection were 54% for the PBS character in the improved population and 0,46% for the PAP character for the wild population. The observed heterozygosity was 0,839 to 0,747 for adults and 0,425 to 0,399 for the progenies, for the two populations studied. The multilocus crossing rate ( ) varied from 0,726 (MP) to 0,798 (PS), indicating that most of the seeds originated by crosses, but there was a presence of self-fertilization, characterized by a mixed mating system.The presented inbreeding was generated by correlated crosses and not by self-fertilization. Due to the effective size (Ne) of less than 4, it resulted in the need to collect seeds for purposes of genetic conservation, environmental recovery and forest improvement of at least 94 and 89 trees, for PM and PS respectively. / FAPESP: 2013/03074-5

El principal objetivo de este trabajo es evaluar las diferencias en rentabilidad obtenidas por medio de los Multifondos (Fondo Tipo 1, Fondo Tipo 2 y Fondo Tipo 3) frente a las obtenidas con el Fondo Único (Fondo Tipo 2) desde el año 2007 al 2014 con la finalidad de evaluar el impacto positivo de este nuevo esquema en el sistema pensionario y en la inversión nacional: para llevar a cabo este análisis se utiliza el índice Sharpe para cuantificar el esquema Multifondos en el SPP y la paradoja de Feldstein-Horioka para visualizar el impacto en la inversión. / The main objective of this envestigation is to evaluate the differences in profitability obtained through the Multi-Funds (Fund Type 1, Fund Type 2 and Fund type 3) compared to those obtained with the fund that exixts sice the creation of the system (Fund Type 2) from 2007 to 2014 with the purpose of evaluating the positive impact of this new scheme in the pension system and in the national investment; to carry out this analysis the Sharpe Index is used to quantify the effect of the Multi-Fund scheme in the Private Pension System and the Feldstein-Horioka Paradox to mesure the impact on the domestic investment. / Trabajo de investigación

No description available.

La prévision de l'évolution des déchets radioactifs lors d'un stockage en couche géologique argileuse profonde nécessite une bonne compréhension du transport de l'eau et des ions dans l'argile. Leur diffusion dans ce milieu poreux et chargé est décrite par des paramètres empiriques, comme le coefficient de distribution (Kd) qui rend compte des interactions avec les surfaces minérales. Notre travail a porté sur la pertinence de ce concept et sa définition à partir de processus microscopiques. <br /><br />Nous avons d'abord modélisé la contribution ionique aux propriétés diélectriques des argiles, et proposé une détermination de Kd par spectroscopie diélectrique. <br /><br />Nous avons ensuite calculé par simulations microscopiques (Monte-Carlo et dynamique moléculaire) les enthalpies libres et enthalpies d'échange ionique pour les ions alcalins, qui contrôlent Kd et ses variations avec la température T. Les résultats pour le césium sont en bon accord avec des mesures de microcalorimétrie et de Kd en fonction de T. <br /><br />Après avoir contribué au développement d'une nouvelle méthode de simulation sur réseau (Lattice Fokker-Planck), nous l'avons utilisée pour établir un lien explicite entre la dynamique microscopique des ions et le modèle de diffusion-réaction qui sous-tend la notion de Kd. <br /><br />Enfin, nous avons étudié par simulation de dynamique moléculaire la cinétique d'échange d'eau et d'ions entre les particules d'argile (porosité interfoliaire) et la porosité extra-particulaire. Les résultats confirment les hypothèses généralement admises selon lesquelles l'eau et les cations peuvent explorer toute la porosité, tandis que les anions sont exclus des espaces interfoliaires.

Moore's Law has helped Field Programmable Gate Arrays (FPGAs) scale continuously in speed, capacity and energy efficiency, allowing the integration of ever-larger systems into a single FPGA chip. This brings challenges to the productivity of developers in leveraging the sea of FPGA resources. Higher level of design abstractions and programming models are needed to improve the design productivity, which in turn require memory architectural supports on FPGAs. While previous efforts focus on computation-centric applications, we take a bandwidth-centric approach in designing memory systems. In particular, we investigate the scheduling, buffered switching and searching problems, which are common to a wide range of FPGA applications. Despite that the bandwidth problem has been extensively studied for general-purpose computing and application specific integrated circuit (ASIC) designs, the proposed techniques are often not applicable to FPGAs. In order to achieve optimized design implementations, designers need to take into consideration both the underlying FPGA physical characteristics as well as the requirements from applications. We therefore extract design requirements from four driving applications for the selected problems, and address them by exploiting the physical architectures and available resources of FPGAs. Towards solving the selected problems, we manage to advance state-of-the-art with a scheduling algorithm, a switch organization and a cache analytical model. These lead to performance improvements, resource savings and feasibilities of new approaches for well-known problems.

Moore's Law has helped Field Programmable Gate Arrays (FPGAs) scale continuously in speed, capacity and energy efficiency, allowing the integration of ever-larger systems into a single FPGA chip. This brings challenges to the productivity of developers in leveraging the sea of FPGA resources. Higher level of design abstractions and programming models are needed to improve the design productivity, which in turn require memory architectural supports on FPGAs. While previous efforts focus on computation-centric applications, we take a bandwidth-centric approach in designing memory systems. In particular, we investigate the scheduling, buffered switching and searching problems, which are common to a wide range of FPGA applications. Despite that the bandwidth problem has been extensively studied for general-purpose computing and application specific integrated circuit (ASIC) designs, the proposed techniques are often not applicable to FPGAs. In order to achieve optimized design implementations, designers need to take into consideration both the underlying FPGA physical characteristics as well as the requirements from applications. We therefore extract design requirements from four driving applications for the selected problems, and address them by exploiting the physical architectures and available resources of FPGAs. Towards solving the selected problems, we manage to advance state-of-the-art with a scheduling algorithm, a switch organization and a cache analytical model. These lead to performance improvements, resource savings and feasibilities of new approaches for well-known problems.

No description available.

No description available.