Les circuits FPGAs (Field Programmable Gate Arrays) sont devenus des acteurs importants dans le domaine du traitement numérique qui a été dominé auparavant par les microprocesseurs et les circuits intégrés spéci ques. Le plus grand dé pour les FPGAs aujourd'hui est de présenter un bon compromis entre une grande souplesse et de bonnes performances (vitesse, surface et consommation). La combinaison de trois facteurs dé nit les caractéristiques d'un circuit FPGA: la qualité de l'architecture, la qualité des outils CAO de con guration et la conception électrique du FPGA. L'objet de cette thèse est l'exploration de nouvelles architectures et de structures d'interconnexion qui pourront améliorer les performances de ces circuits. En effet, les ressources d'interconnexion occupent 90% de la surface totale et occasionnent 60% de la consommation électrique. Les architectures étudiées présentent des structures matricielles et arborescentes. Les principaux résultats sont les suivants: Au départ nous explorons différentes topologies arborescentes et nous comparons leurs surfaces à celles des architectures matricielles. Pour cela, nous développons une plateforme d'outils logiciels permettant d'implanter différents circuits logiques sur l'architecture cible. En se basant sur cette étude expérimentale, nous dé nissons une nouvelle architecture arborescente. Nous montrons, en nous appuyant sur un modèle d'estimation de surface, que cette architecture permet de réduire la surface totale de 56% par rapport à une architecture matricielle. Ceci est dû essentiellement à une meilleure utilisation des ressources d'interconnexion. Nous explorons les effets des différents paramètres de l'architecture proposée: le coef cient de Rent, la taille des groupes logiques et le nombre d'entrées par bloc logique. Ceci permet de régler l'architecture pour l'adapter à des domaines d'applications qui ont des contraintes spéci ques en terme de surface, vitesse et consommation. En n, nous proposons une architecture qui rassemble les avantages des structures arborescentes et matricielles. Nous uni ons les deux structures en construisant des groupes de blocs logiques qui ont localement un réseau d'interconnexion arborescent et qui sont connectés entre eux via un réseau matriciel. Nous montrons que l'architecture obtenue présente un bon compromis entre l'évolutivité de la vue physique et la densité de la surface.

<p>This study seeks to evaluate the performance of unit trust managers based on their client affiliation classification. Worldwide, the number of investors investing in unit trusts is on the rise and increasingly they want to be able to evaluate the performance of the managers managing their funds so as to make better investment decisions. This increase in the asset size and number of unit trusts funds could be attributed but not limited to the low capital required for investment by small investors who before could not afford to invest in portfolios requiring large capital (Prather, Bertin, and Henker, 2004). In addition, the fund managers of these units are believed to have special skills such as market timing and stock selectivity which contribute to the performances they achieve. The evaluation of the performance of unit trust fund managers is a largely unexplored area in South Africa. As a result, the study focuses on South Africa fund managers and has as aim to evaluate the performance of two groups of fund managers (independent and dependent) who were classified based on their client affiliation structure. The client affiliation classification is as a result of the fund manager‟s clientele base. The dependent group are those who formed part of a group structure and offer other wealth management services for which their clients or investors in the unit trust services originate from within the group while the independent group are those whose clients are pulled together from diverse individuals or institutions and does not form part of a group or render other services other than fund management. Two fund types were selected namely / general equity funds and balanced funds. It has also examined the underlying skills the different groups of fund managers possess. The performance of unit trust has an effect on many parties who are related in one way or the other to the unit trust funds. The results of this study will inform individual investors, trustees and asset consultants in their decision making process of selecting a fund manager. The results of the study will be of value to the asset management industry in terms of assessing their structures and restructuring the investment service business to meet the expectations of their clients / the investors. It could also be used as a marketing tool. Publicly available historical data on the returns generated by fund managers for a five year period from&nbsp / 2005 to 2009 was obtained. Analyses were done using the independent sampled t-test and the Treynor Mazel model respectively for the different research questions posed. The results obtained indicated that there were no statistically significant differences between the performances of independent fund managers with those of dependent fund managers. However, dependent fund managers of equity funds performed better than their counterparts the independent fund managers. In the case of balanced funds, the independent fund managers performed better than their dependent counterparts. On average, both fund&nbsp / manager types possessed selectivity skills for equity funds and none for balanced funds. However for both fund types, the dependent fund manager demonstrated more selectivity skills than their independent counterparts. The results for market timing skills demonstrated that on average, both fund managers did not possess market timing skills for balanced funds while possessing these skills for equity funds. The dependent&nbsp / fund managers demonstrated more market timing skills for balanced funds though negative when compared to that of their counterparts. On the other hand, the equity fund independent fund&nbsp / managers demonstrated more market timing skills than the dependent fund managers.</p>

Abstract Title: Customer Perceived Value of Credit Card Rewards - A study on Canadian Consumers Level: Final assignment for Bachelor’s Degree in Business Administration Author: Lisa Smedley Supervisor: Jonas Kågström Date: 2013 - January Aim: The aim of this study is to investigate what influences Customer Perceived Value; where Canadian consumers’ preferences lie in terms of rewards in the Canadian credit card industry. Method: After researching previous studies and determining what constructs have been utilized prior on similar research topics, I implement a quantitative, and to some extend iterative, research approach. Through survey research, I investigate Canadian consumer preferences through a survey sample of 124 Canadian consumers in Calgary, Alberta, Canada. Result &amp; Conclusions: One finding in the study indicates that utilitarian benefits, which provide financial gain for the card holder, are perceived by respondents as the most valuable reward. Another finding is that inexperienced credit card holders see significantly greater value in symbolic benefits than experienced card holders do. The present study does not support the theory that customer involvement influences the customer’s perception of rewards. 2 Suggestions for future research: More extensive research is needed on the subject of whether Canadian consumers’ perceived value of rewards is influenced by their level of involvement in their credit card. Also, studies involving additional factors that could possibly determine a consumer’s perception of rewards, such as income and ethnicity should be investigated for a more well-rounded understanding of customer preferences. Contribution of the thesis: The present study contributes with new findings that can be of substantial significance for Canadian financial institutions as it provides insight into what credit card rewards Canadian consumers perceive as being valuable to them. Key words: Rewards programs, credit cards, customer loyalty, perceived customer value, timing of reward, type of reward, dimension of benefit, utilitarian, hedonic, symbolic

<p>Disclosure of HIV positive status to male partners is well established as a key element in the success of prevention of mother to child transmission of HIV programmes, as it helps improve adherence to ARVs by the women within these programme. However, partner notification rates remain low in the urban areas of Lusaka, Zambia against a high HIV prevalence of 25%. The purpose of this study was to explore the timing of disclosure as part of the process of disclosure amongst women who were part of the PMTCT services at Kaulu health centre in Lusaka. An exploratory descriptive study using qualitative research methods was conducted. 15 women, who were attending the Kaulu health centre PMTCT programme, were requested to participate in a semi-structured interview. The women, who were purposively selected with the aid of the health centre‟s PMTCT focal point nurse, had to have disclosed their HIV positive status to their partner, either before or during the course of their pregnancy or after delivery. To increase rigour,&nbsp / individual interviews were conducted with 5 health workers associated with the PMTCT programme so as to obtain their perspective and experiences on the issue of HIV disclosure amongst their PMTCT patients. Participation in the study was voluntary and all information obtained during the course of the interviews remained confidential and secure. Potential participants were each provided with an explanation of the purpose and process of the study and their informed written consent obtained before the researcher embarked on the interviews. Content analysis of the transcripts was done so as to develop coding categories and identify emerging themes. Disclosure to male partners is an important step in PMTCT and facilitates adherence to HIV care for the family and should be done as early as possible after the woman receives her HIV test result, though there exists a range of alternative times when it can be done. The relationship existing between a couple is very important in determining the timing of when a woman chooses to disclose. PMTCT services need to provide ongoing counselling for HIV positive women during pregnancy and after giving birth that supports, informs and equips them with the necessary skills to make an informed and timely decision about disclosure to a partner. In addition, the PMTCT service providers need to be encouraged to implement couple counselling as a strategy to facilitate disclosure as well as establishment of a peer support network for HIV positive pregnant women. The study findings will be used to contribute to health workers‟ capacity to support women manage the disclosure process to their male partners, thus helping to increase the disclosure rate and also contributing to improving the positive effect of the PMTCT services, in Lusaka, Zambia</p>

Abstract Title: Emerging Markets – A Case Study in Foreign Market Entry toBangladesh Keywords: emerging market, entry strategy, market entry, factors behind entry choice, entry mode, entry node, entry timing Background: Internationalism and international marketing are hot topics among the strategy discussions of the companies and as a result companies continuously look for new, unreached sales potential to their products and services as well as better use of their resources. Purpose: To find the most efficient international market entry strategy for companies moving from developed/transition economy to an emerging market. Theoretical framework: The base for the start of internationalisation process is company’s inner motives and resources. Motives and resources combined with the cultural distance, competition and general external environment of host country form potential company-specific risks for the entry to foreign market. Potential customers in combination with company resources shows how big is the match between market demand and what company can offer and therefore determines the potential reward. Risks and reward are both input to the decision making process where the potential benefits and drawbacks are analysed against each other. The output of this decision making is the entry strategy. Methodology: Internet was mainly used to collect secondary data about company resources, cultural distance and external environment. Interviews with 150 retailers inBangladesh were conducted to collect primary data about the competition and consumer behaviours in the hosiery market ofBangladesh. Then comparative analysis was made based on the model developed by the authors to reach to the decision. Conclusion: The most effective entry strategy for the entry to emerging markets is indirect exporting through an agent in case there is high location risk, moderately high competition risk, medium country risk and moderately low demand risk, the company has no surplus finances for big investments and no prior experience in doing business in an emerging market.

In this thesis, we propose an approach to estimate the Worst Case Response Time (WCRT) of each task in a preemptive multi-tasking single-processor real-time system utilizing an L1 cache. The approach combines inter-task cache eviction analysis and intra-task cache access analysis to estimate the Cache Related Preemption Delay (CRPD). CRPD caused by preempting task(s) is then incorporated into WCRT analysis. We also propose a prioritized cache to reduce CRPD by exploiting cache partitioning technique. Our WCRT analysis approach is then applied to analyze the behavior of a prioritized cache. Four sets of applications with up to six concurrent tasks running are used to test our WCRT analysis approach and the prioritized cache. The experimental results show that our WCRT analysis approach can tighten the WCRT estimate by up to 32% (1.4X) over prior state-of-the-art. By using a prioritized cache, we can reduce the WCRT estimate of tasks by up to 26%, as compared to a conventional set associative cache.

A Delay-Locked Loop (DLL) for the generation of multiple clock phases/delays is proposed. Several new techniques are used to help enhance the DLLs performance, specifically, to achieve wide lock range, short locking time, and reduced jitter. The DLL can be used for a variety of applications which require precise time intervals or phase shifts. The phase detector (PD), charge pump (CP), and voltage-controlled delay line (VCDL) are the three most important blocks in a DLL. In our research, we have proposed a novel structure which integrates the functionality of both the PD and CP. By using this structure, a fast switching speed can be achieved. Moreover, the combined PD and CP also lead to reduced chip area and better jitter performance. A novel phase detection algorithm is developed and implemented in the combined PD and CP structure. This algorithm also involves a start-control circuit to avoid locking failure or false lock to harmonics. With the help of this algorithm, the proposed DLL is able to achieve lock as long as the minimum VCDL delay is less than one reference clock cycle, which is the largest possible lock range that can be achieved by the DLL. The VCDL uses fully differential signaling to minimize jitter. The delay stage of the VCDL is built with a differential topology using symmetrical loads and replica-feedback biasing, which provides a low sensitivity to supply and substrate noise as well as a wide tuning range. In addition, a shift-averaging technique is used to improve the matching between delay stages and thus to equalize the delay of each individual stage.

The purpose of the study is to analyze the problem of offer pricing and share issuing in the issuer with a dynamic model. The result shows: First, the correlation between discount and the prosperity of the market is not a linear function. Thus, ¡¥the IPO¡¦s timing¡¦ should be the key issue of issuing firm, which can reasonably explain the fluctuations in underpricing over time. Second, we also discover that the relationship between the ownership structure and the prosperity of the market is not a liner function as what Bolton and Thadden (1998) have claimed. Third, the firm with low quality prefers to release issuing shares in IPO when the market is booming. Fourth, the less the IPO discount in traditional industry, the more the share releasing. Finally, the more the phenomena of subscription, the more the investors subscribe blindly and generate an effect of the cascade. One of the contributions in this model comprises the views between the prosperity of the market and IPO discount in the literature to integrate a profound viewpoint by dividing the IPO discount into the offer pricing error and the market pricing error. Another finding is that the offer pricing error will decrease subject to the prosperity of the market. This is consistent with the viewpoint of Chemmanur (1993) that when the market is booming, the issuers can decrease the cost of adverse choice. As for the non-linear relation between the market pricing error and the prosperity of the market is coordinate with what Ma(1999) has claimed that IPO discount is influenced by the offer pricing and the market pricing. Besides, the non-linear relationship between ownership structure and the prosperity of market can reasonably account for the real situation. Besides, the benefit of controlling authority is better than the value of the cash flow of the released shares owing to the depression of the market, even the market is turning booming. Therefore, the issuers will minimize releasing share. However, the issuers will increasing released shares up to the market is keeping booming and the value of the cash flow of the released shares is better than the benefit of controlling authority. An auction model is established to analyze the choice of firms between fixed-price offerings and auctions. Chemmanur and Liu(2003) stated that obtaining the auction information is more advanced; however, the cost is high so that it can not catch enough investors to produce information which does not favor the discovery of the secondary market price. On the contrary, we proclaim that the information transmitting effect produced by auction will be helpful to the discovery of the secondary market price. Thus, a profitable company but suffers from high levels of information asymmetry prior to the IPO might expect the secondly market price to reflect its real value by auction. Moreover, the choice of underwriting mechanisms lies in the prosperity of the market, the reputation of the company and the degree of the releasing shares. The result of the empirical study is consistant with the expectation of our model.

In this thesis, X-ray and optical observations of accretion powered pulsars are presented. By using archival RXTE observations we work on the X-ray spectral and pulse timing analysis of 4U 1538-52, 4U 1907+09, SMC X-1 to have more detailed information about their orbital and spin parameters. For 4U 1538-52 and SMC X-1, we determined new orbital epochs. By using long term pulse history of 4U 1907+09, we were able to work spin-down trend of the system and also calculate the change in the spin-down rate. Using Fermi/GBM observations we can monitor bright accreting pulsar systems. We are producing long term histories of pulse frequency and flux of 20 continuosly monitoring systems. Adding Swift/BAT observations to GBM observations, for 4U 1626-67, we did reveal the characteristics belong to spin-down trend before and spin-up behaviour after torque reversal seen in 2008 February. Two newly discovered IGRJ06074+2205 and IGRJ01583+6713 sources are identified as X-ray binary systems and we found parameters of them like distance, magnitudes, by using both optical photometric and spectroscopic observations.

It is known that concurrent secondary tasks or attentionally salient stimuli shorten reproduced temporal durations. The main aim of this thesis is to use three types of secondary tasks to see their effects on duration judgments. The Attentional Gate Model (Block &amp / Zakay, 2006) served as theoretical background for a series of 4 experiments. There were 2 baseline/control experiments for studying the effect of 2 different and novel secondary tasks which are temporal comparison and non-temporal executive tasks. Three duration lengths (short-moderate-long) were used (15, 30 and 45 sec) that subjects had to reproduce. In Exp-1 (control experiment for Exp-2) subjects had to reproduce almost empty time intervals. Exp-2, which investigated the role of a secondary temporal task, revealed significantly decreased reproduced durations as compared to Exp-1 which is in line with our hypothesis. In Exp-3 (control experiment for Exp-4) subjects carried out a non-temporal/non-executive secondary task. Exp-4, in which a Simon task was used as a non-temporal executive secondary task, resulted in significantly decreased reproduced durations as compared to Exp-3 as well. Moreover, duration length effects were found for all experiments that included an attention consuming secondary tasks (Exp-2-3-4), i.e., longer durations were more underestimated than shorter ones in the presence of attention demanding tasks. We conclude that secondary temporal tasks and even more so executive non-temporal tasks can lead to decreased temporal duration judgements, thus affecting subjects&rsquo / time perception, in line with the Attentional Gate Model.