The macro-dataflow model of execution has been used in scheduling heuristics for directed acyclic graphs. Since this model was developed for the scheduling of parallel applications on distributed computing systems, it is inadequate when applied to the multicore shared-memory computers prevalent in the market today. The pulled-macro-dataflow model is put forth as an alternative to the macro-dataflow model, having been designed specifically to accurately describe the memory bandwidth limitations and request-driven nature of communications characteristic of today's machines. The performance of the common scheduling heuristics DSC and CASS-II are evaluated under the pulled-macro-dataflow model and it is shown that their poor performance motivates the development of a new scheduling heuristic. The Concurrent Tournament Reducer (ConTouR) is developed as a scheduling heuristic which operates well with the pulled-macro-dataflow model. ConTouR is compared to the existing heuristics Load Balancing and Communication Minimization in scheduling two programs. For both programs, the other reducers are shown to outperform ConTouR.

Thesis (PhD)--Stellenbosch University, 2015. / ENGLISH ABSTRACT: Macro synthetic fibres are known to significantly improve the toughness and energy absorption capacity of conventional concrete in the short term. However, since macro synthetic fibre are flexible and have relatively low modulus of elastic compared to steel fibres, it is uncertain if the improved toughness and energy absorption could be sustained over a long time, particularly under sustained tensile loadings. The main goal of this study is to investigate the time-dependent crack mouth opening response of macro synthetic fibre reinforced concrete (FRC) under sustained uniaxial tensile loadings, and to simulate the flexural creep behaviour. For the purpose of simulating the in-service time-dependent condition, all specimens were pre-cracked. Experimental investigations were carried out at three levels (macro, single fibre and structural) to investigate the time-dependent behaviour and the mechanisms causing it. At the macro level, compressive strength, uniaxial tensile strength and uniaxial tensile creep test at 30 % to 70 % stress levels of the average residual tensile strength were performed. To understand the mechanism causing the time-dependent response, fibre tensile test, single fibre pullout rate test, time-dependent fibre pullout test and fibre creep test were done. Flexural test and flexural creep test were done to simulate the structural level performance. The results of this investigation have shown significant drop in stress and increase in crack width of uniaxial tensile specimens after the first crack. The post cracking response has shown significant toughness and energy absorption capacity. Under sustained load at different stress levels, significant crack opening has been recorded for a period of 8 month even at a low stress level of 30 %. Creep fracture of specimens occurred at 60 % and 70 % indicating that these stress levels are not sustainable for cracked macro synthetic FRC. The single fibre level investigations have revealed two mechanisms responsible for the time-dependent crack widening of cracked macro synthetic FRC under sustained loading: time-dependent fibre pullout and fibre creep. In all cases of investigation, fibre failure was by complete pullout without rupture. Flexural creep results have shown that the crack opening increases over time. After 8 months of investigation, the total crack opening was 0.2 mm and 0.5 mm at 30 % and 50 % stress levels respectively. Since the crack opening of tensile creep and flexural creep specimens cannot be compared due to differences in geometry, specimen size, load transfer mechanisms and stress distribution in the cracked plane, a finite element analysis (FEA) was conducted. Material model parameters obtained from the uniaxial tensile test and viscoelastic parameters from curve fitting to experimental uniaxial creep results have been implemented to successfully predict the time-dependent crack opening of specimens subjected to sustained flexural loading. Analyses results correspond well with experimental result at both 30 % and 50 % stress levels. / AFRIKAANSE OPSOMMING: Makro sintetiese vesels is bekend daarvoor dat dit die taaiheid en energie absorpsie van konvensionele beton beduidend verbeter in die kort termyn. Aangesien makro sintetiese vesels buigsaam is met 'n relatiewe lae styfheidsmodulus in vergeleke met staalvesels, is dit onseker of die verhoogde kapasiteit vir energie absorpsie en taaiheid volgehou kan word oor die langer termyn, veral in gevalle waar dit aan volgehoue trekkragte blootgestel is. Die hoofdoel van die studie is om die tydafhanklike-kraakvergrotingsgedrag van makro sintetiese veselversterkte beton (VVB) wat blootgestel is aan volgehoue trekkragte te ondersoek asook die simulasie van die kruipgedrag in buig. Ten einde die werklike toetstande te simuleer is al die proefstukke doelbewus gekraak in 'n beheerde manier voor die aanvang van die toetse. Die eksperimentele ondersoek is uitgevoer op drie vlakke (makro, enkelvesel en strukturele) om die tydafhanklike gedrag sowel as die meganismes verantwoordelik vir hierdie gedrag te ondersoek. Op die makro-vlak is druktoetse gedoen saam met eenassige trek- en eenassige kruiptoetse met belastings tussen 30 % en 70 % van die gemiddelde residuele treksterkte. Om die meganisme wat die tydafhanklike gedrag veroorsaak te verstaan is veseltoetse, enkel vesel uittrektoetse, enkel vesel uittrek kruiptoetse asook kruiptoetse op vesels gedoen. Buigtoetse en buig kruiptoetse is ook gedoen om die gedrag op die strukturele vlak te ondersoek. Die resultate van hierdie ondersoek wys dat daar 'n beduidende val in spanning is en dat daar gepaardgaande kraak opening in die eenassige trek proefstukke plaasgevind het na die vorming van 'n kraak. Die na-kraak gedrag wys beduidende taaiheid en energie absorpsie kapasiteit. Gedurende die volgehoue trekbelasting by verskillende spanningsvlakke is beduidende kraakvergroting opgemerk, selfs by 30 % belasting na 8 maande. Kruipfaling het plaasgevind by proefstukke met belastings van 60 % en 70 % wat daarop wys dat hierdie spanningsvlakke nie geskik is vir gekraakte makro sintetiese VVB nie. Op die enkel veselvlak is twee meganismes geïdentifiseer wat verantwoordelik is vir die kraakvergroting oor tyd vir gekraakte makro sintetiese VVB met volgehoue trekbelasting: tydafhanklike vesel uittrek en vesel-kruip. In alle gevalle in hierdie ondersoek was die falingsmeganisme vesels wat uittrek. Buig kruiptoets resultate wys dat die krake vergroot oor tyd. Na 8 maande van ondersoek was die kraakwydtes 0.2 mm en 0.5 mm by 30 % en 50 % spanningsvlakke onderskeidelik. Aangesien die kraak opening van eenassige trek kruiptoetse en die buig kruiptoetse nie direk met mekaar vergelyk kan word nie weens die verskille in geometrie, proefstuk grootte en spanningsverdeling in die kraakvlak, is 'n eindige element analises (EEA) gedoen. Materiaal eienskappe is bepaal deur gebruik te maak van die eenassige kruip trektoets se resultate en viskoelastiese parameters is bepaal deur middel van kurwepassing van die resultate. Dit was gebruik om suksesvol die buig kruip kraak opening gedrag te simuleer. Die analises se resultate vergelyk goed met die eksperimentele data by beide 30 % en 50 % spanningsvlakke.

This thesis discusses the relationship between the two forces—the Internet and the conventional media, and the rules and etiquette that have governed these media from the mid-1990s to present day. I divide the history of interaction between them into three phases, corresponding to different power balances and stances in the field of cultural production. By analyzing three Internet-themed TV talk show episodes, I probe how the Internet has gradually evolved into an active competitor, and how the conventional media have correspondingly changed their stance in response to their thriving counterpart. / text

With the prevalence of legacy C/C++ code, issues of readability and maintainability have become increasingly important. When we consider the problem of refactoring or migrating C/C++ code, we see the significant role that preprocessor directives play. It is partially because of these preprocessor directives that code maintenance has become extremely difficult. This thesis describes a method of fact extraction and code manipulation to create a set of transformations which will remove preprocessor directives from the original source, converting them into regular C/C++ code with as few changes as possible, while maintaining readability in the code. In addition, some of the subtle issues that may arise when migrating preprocessor directives are explored. After discussing the general architecture of the test implementation, an examination of some metrics gathered by running it on two software systems is given.

This thesis contributes to the empirical literature about how uncertainty affects firm-level investment behavior and capital accumulation using a structural econometric approach. Chapter 2 surveys the literature and highlights that there are two key channels through which uncertainty may affect investment decisions. One reflects the non-linearity of operating profits in stochastic demand or productivity parameters, summarized as the Hartman-Abel-Caballero (HAC) effect. Another reflects frictions in capital adjustment, summarized by different forms of capital adjustment costs: partial irreversibility, a fixed cost of undertaking any investment and quadratic adjustment costs. Chapter 3 presents simulation evidence about the effects of uncertainty on investment dynamics and capital accumulation through different forms of adjustment costs. Using the Method of Simulated Moments, Chapters 4 and 5 estimate fully parametric structural investment models, for panels of Brazilian and UK manufacturing firms, respectively. Chapter 4 investigates the effects of reducing capital adjustment costs. Counterfactual simulations indicate that investment would be much more responsive to new information about profitability if firms in Brazil faced a lower level of adjustment costs. A lower level of adjustment costs would also induce firms to operate with substantially higher capital stocks. Both these effects are mainly due to the importance of the estimated quadratic adjustment costs. Chapter 5 then investigates the effects of changing the level of uncertainty. The estimated investment models predict a small effect of uncertainty on investment dynamics in the short-run, and a negative and potentially large effect of uncertainty on capital accumulation in the long-run. The long-run effect of uncertainty operates through the negative effect of quadratic adjustment costs in the baseline model, or through a richer combination of effects in an extended model that allows discount rates to vary with the level of uncertainty.

Typescript, etc. / Digitized by Kansas Correctional Industries

O termo macro-rugosidade é empregado para caracterizar os elementos contínuos ou esparsos de formas diversas, que são aplicados artificialmente no fundo e paredes dos condutos hidráulicos, em oposição à rugosidade de grão, relacionada às asperezas das paredes destes condutos. Neste trabalho apresenta-se o estudo teórico-experimental do comportamento do escoamento livre com a presença de macro-rugosidade constituída de elementos prismáticos de seção transversal retangular, denominados travessões, que encontram aplicações práticas na estabilização de leitos e regularização de canais para navegação. O emprego da macro-rugosidade resulta na ocorrência de maiores perdas de energia quando as profundidades são pequenas e na redução progressiva deste efeito à medida que a lâmina d\'água se eleva. O estudo teórico compreende o tratamento dado pela literatura ao problema da resistência ao escoamento em canais, desenvolvido a partir da integração da tradicional lei logarítmica para o para o perfil bidimensional de velocidades. Também foi abordado o modelo de distribuição probabilística de velocidades de CHIU, baseado na maximização da entropia. Estudos específicos sobre macro-rugosidade de fundo, desenvolvidos a partir da década de 1960, foram pesquisados para orientação dos estudos experimentais desenvolvidos, notadamente os trabalhos de MORRIS & WIGGERT, SAYRE & ALBERTSON e KNIGHT & MACDONALD. Ensaios experimentais foram desenvolvidos em canal de laboratório para acaracterização das perdas de carga dos travessões de fundo, em diferentes condições de profundidade e espaçamento longitudinal. Nestes ensaios desenvolveu-se uma técnica para medição das perdas de energia e cálculo das tensões de atrito em escoamentos permanentes e gradualmente variados no espaço, situação esta usual nos canais empregados em engenharia. ) Foi também considerado o efeito de parede, resultante do emprego de um canal de laboratório de pequenas dimensões, de modo a poder-se generalizar os resultados para o caso bidimensional. Os dados experimentais permitiram o teste dos modelos para cálculo de perdas de carga de Manning e do Perfil Logarítmico, estabelecendo-se correlações entre as perdas e as características geométricas da macro-rugosidade. Como contribuição ao tema foi proposta uma adaptação do modelo logarítmico considerando a hipótese de MORRIS & WIGGERT da separação do escoamento em uma zona de interferência, denominada de zona de hiper-turbulência e outra, de turbulência normal. Os resultados obtidos indicaram a validade do modelo proposto. Ainda como contribuição ao tema, a lei de distribuição probabilística das velocidades foi aplicada, obtendo-se uma estimativa dos parâmetros para as diferentes configurações geométricas. / The designation large-scale roughness is generally employed to describe spot or continuous elements artificially disposed in channel bottom, different from wall roughness or grain roughness, associated to the wall material. This work presents a theoretical and experimental study of the free flow in channels with prismatic rectangular elements regularly disposed on the bottom, named traverse blocks, which have large application in river stabilization and navigation works. The use of the large-scale roughness results in a great head loss for small depths and a progressive reduction of these losses as the depth increases. The theoretical analysis initially focused the traditional approach considering the logarithmic velocity distribution law. It was also analyzed the probabilistic velocity distribution law by CHIU, based on the maximum entropy principle. Specific studies about large-scale roughness, developed since 1960, where considered to guide the model testes, especially those by MORRIS & WIGGERT, SAYRE & ALBERTSON and KNIGHT & MACDONALD. Experimental data where collected from a laboratory channel and a technique for computing bottom shear stress were developed allowing to consider permanent gradually varied flows, a regular situation in engineering applications. The wall interference in total channel shear stress where considered in order to make data comparable with those from real two-dimensional flow. Practical correlations using geometriccharacteristics of the traverse block roughness and flow parameters where established considering Manning\'s Formula and the Logarithmic Law model. In order to achieve a better modeling of the phenomenon, a complementary term to the Logarithmic law was proposed, considering the hyper-turbulent layer over the roughness, and fit to the data. Finally, parameter estimation for the use of the CHIU\'s law with large-scale roughness where developed as another contribution to the matter.

Cette thèse propose de reconstituer l'histoire environnementale holocène du plateau du Béage en Ardèche (massif duMézenc, Massif Central, France). Ce travail repose sur l'étude des sédiments issus des tourbières de La Narce du Béage et de Pialeloup, mais aussi du marais de Pré-du-Bois et des séquences de la moyenne vallée du Gage. Notre approche vise (1) à caractériser le contenu paléoécologique des sédiments (macro-restes et palynologie), (2) à appréhender les évolutions du bassin versant via une analyse multi-proxies (susceptibilité magnétique, granulométrie,taux de matière organique, géochimie) et (3) à confronter ces éléments aux données locales et régionales disponibles.Durant !'Holocène ancien et moyen, des fluctuations locales d'humidité paraissent concomitantes de changements climatiques majeurs dans l'hémisphère nord (épisodes humides: 10400-10000, 9600-9100, 7900-7400, 6400-6000,5400-5000 cal. BP; secs: 10900-10500, 9900-9600 cal. BP; contrasté: 8500-8150 cal. BP). Cette analyse montre des impacts agro-pastoraux entre 6900 et 6450 cal. BP (Néolithique ancien), entre 5700 et 5200cal. BP (Néolithique moyen 2/récent) puis vers 4800 - 4100 cal. BP (Néolithique final). La pression anthropique s'intensifie entre 2400 et 1250 cal. BP (IVème s. av. J.-C.-Vlème s. apr. J.-C.). Le développement agricole médiéval s'opère à partir de 1050 cal. BP (Xème s.) : culture du seigle, cheptel ovin et bovin, productions de noix et de châtaignes à partir des XIIIème et XIVème s. L'érosion des versants semble aussi stimulée par la péjoration du Petit Âge Glaciaire (Xlllème-XVIIIème s.).Aux XIXème et XXème siècles, les activités d'élevage bovin perdurent mais la céréaliculture finit par disparaître. / This PhD thesis has resulted in the reconstruction of the holocene environmental history of the Béage plateau in Ardèche (massif du Mézenc, Massif Central, France). It focussed on the study of organic sediments from the La Narce du Béage and Pialeloup peatlands. Two complementary sites were also considered: the Pré-du-Bois swamp and the alluvial stratigraphies from the middle Gage valley. Our approach aimed (1) to perform a palaeoecological study of these sites (macrofossils and palynology); (2) to better understand watershed evolution thanks to multi-proxies analyses (magnetic susceptibility, grain size, organic matter content, geochemistry); (3) to compare these results with other local to regional data. During the Barly and the Middle Holocene, local humidity variations occuned concurrently with rapid climate changes(RCCs) recorded in the Northern Hemisphere (dry episodes: 10,900-10,500; 9,900-9,600; moist: 10,400-10,000;9,600-9,100; 7,900-7,400; 6,400-6,000 and 5,400-5,000 cal. BP; contrasted: 8,500-8,150 cal. BP). This study also allowed to identify agro-pastoral impacts around 6,900-6,450 cal. BP (Barly Neolithic), around 5, 700-5,000 cal. BP (Middle to Recent Neolithic) and close to 4,800-4,100 cal. BP (Final Neolithic). Anthropogenic impact got stronger between 2,400 and 1,250 cal. BP (IV'h century BC-Vlth century AD). A new stage of landscape management was recorded after 1,050 cal. BP (Xth century AD), involving rye cultivation, extensive livestock grazing(sheep and cattle) and fruit tree farming (walnut and chestnut: XIIIth-XIV'h centuries). Severe soil erosion occurred, alsodriven by the Little Ice Age deterioration (from the XIIlth to the XVth century AD). Finally, during the X!Xth and theXXth centuries cattle grazing remained strong, but cultivation ended due to demographic exodus during the xxth century.

The objective of this dissertation is to investigate the impact of important market participants such as Mutual Funds, Hedge Funds and the Federal Reserve Bank on the equilibrium equity premium, risk free rate and asset volatility and to analyze the effect of these institutions on risk shifting, portfolio allocation and financial stability. Specific features of institutional investors and central banks as well as their role in financial markets are reviewed and analyzed in Chapter 1. In Chapter 2, it is shown that the competitive pressure to beat a benchmark may induce institutional trading behavior that exposes retail investors to tail risk. In our model, institutional investors are different from a retail investor because they derive higher utility when they outperform the benchmark. This forces institutions to take on leverage to over-invest in the benchmark. Institutions execute fire sales when the benchmark asset experiences negative shocks. This behavior increases market volatility, raising the tail risk exposure of the retail investor. Nevertheless, ex-post, tail risk is only short lived, all investors survive in the long run under standard conditions, and the most patient investor dominates in the sense that she has the highest consumption wealth ratio. Ex-ante, however, benchmarking is welfare reducing for the retail investor, and beneficial only to the impatient institutional investor. Chapter 3 presents an analysis on how monetary authorities seeking to stabilize inflation, output and smooth interest rates distort the term structure of interest rates and prices of risk relative to an economy where central authorities adjust the money supply without taking into consideration the slope of the yield curve. Closed-form expressions for all equilibrium quantities are presented and the impact of quantitative easing on prices, risk premium and volatility of financial markets instruments, such as stocks and bonds, are evaluated. The changes in macroeconomic variables such as consumption, money demand and investment policies are also investigated. Under the adopted parametrization, quantitative easing is welfare improving. In addition, quantitative easing increases nominal bond and equity volatility, while reducing both real and nominal bond yields for all maturities.

Thesis (M. Sc. (Zoology)) -- University of the North, 1994 / Refer to the document