Nous avons étudié l'influence de la dissolution anodique sur le comportement en fluage à température ambiante du cuivre mono et polycristallin en solution tampon acide acétique - acétate de sodium. Les résultats montrent que l'accélération du fluage dans le cas du cuivre monocristallin orienté pour glissement multiple et dans le cas du cuivre polycristallin est essentiellement dûe à la perte de section causée par la dissolution anodique selon la loi de Faraday (effet mécanique). En outre, les résultats sur des monocristaux orientés pour glissement simple montrent que cet effet mécanique justifie l'effet d'accélération à hauteur de 50%. Les résultats d'une approche microscopique - microscopie à force atomique et microscopie électronique en transmission - mettent en lumière un mécanisme de multiplication des dislocations en surface.

Cette thèse s'intéresse aux méthodologies d'optimisation d'algorithmes de traitement de signal sur les architectures parallèles de processeurs embarqués. L'état de l'art des différentes architectures destinées au milieu embarqué permet de mettre en évidence les différents outils d'optimisation mis à disposition par les concepteurs de processeurs. L'accent est particulièrement mis sur les solutions bénéfiques aux calculs flottants intensifs, tout en notifiant les points communs et les divergences entre les différents processeurs. Le choix de l'algorithme de transformée de Fourier, comme algorithme représentatif des applications de traitement de signal, permet de détailler étape par étape les différents choix d'optimisation dans le cas d'une implémentation sur un PowerPC 970FX. Nous montrons comment à partir d'un algorithme radix-2, il est possible de réduire au plus prés du minimum la complexité de calcul grâce à l'usage de l'instruction de multiplication addition fusionnée. Nous proposons enfin une méthodologie de programmation multi-architectures utilisant le retour d'expérience précédent afin d'optimiser l'ordonnancement des instructions constituant l'algorithme.

Cette thèse présente des méthodes et des algorithmes pour l'évaluation de performance de systèmes à grand espace d'états décrits par des formalismes de haut niveau. Parmi les différents formalismes de haut niveau normalement utilisés, nous nous sommes intéressés au formalisme des Réseaux d'Automates Stochastiques (SAN). Le formalisme SAN se caractérise par la modélisation de systèmes complexes, où un système est représenté par la composition de sous-systèmes (automates) qui interagissent entre eux. Cette interaction se réalise par l'occurrence des événements synchronisants ou des taux fonctionnels. Lorsqu'on calcule l'espace d'états atteignables de systèmes complexes, le principal problème qui surgit est l'explosion combinatoire de l'espace d'états du modèle. Dans la première partie de cette thèse, nous proposons des méthodes pour la génération de l'espace d'états atteignables de modèles compositionnels qui utilisent des taux fonctionnels. Nous utilisons les Diagrammes de Décision Multi-valués (MDD) pour représenter et manipuler les espaces d'états et le formalisme SAN pour la modélisation de systèmes. Un MDD est une structure de donnée arborescente qui permet de représenter et de manipuler de façon performante un très grand espace d'états. L'avancée par rapport à l'état de l'art a été de proposer de méthodes qui prennent en compte ces fonctions qui expriment des relations entre les composants des modèles. Des études d'exemples sont présentées afin d'illustrer les apports de ces méthodes. Dans la deuxième partie de cette thèse, nous nous sommes intéressés à la résolution d'un modèle SAN à temps discret dont la matrice de transition est représentée par une formule tensorielle (appelée descripteur discret). A cet effet, nous présentons l'Algèbre Tensorielle compleXe (ATX) adaptée à la composition parallèle des SAN à temps discret pour la représentation du descripteur et nous démontrons des propriétés qui servent de base aux méthodes itératives pour la résolution de la chaîne de Markov associée au modèle SAN. Un des avantages de représenter un modèle SAN par un descripteur est la façon compacte par laquelle on peut représenter les transitions du modèle: on remplace une description dans un espace produit par un unique produit tensoriel portant sur des facteurs qui décrivent ce qui se passe sur une seule dimension (une composante du modèle SAN). Afin de profiter de cette représentation, nous présentons une méthode de multiplication d'un vecteur de probabilité par un descripteur discret adaptée à cette algèbre. Cette méthode vise à exploiter des propriétés du produit tensoriel complexe de façon à ce que la multiplication par un opérateur sur l'espace produit soit remplacée par une suite d'opérations qui manipulent des données de la taille d'une composante (et pour toutes les composantes).

Nous nous plaçons dans le cadre de l'anneau des polynômes sur un corps fini. Si P est un polynôme de degré au moins 1, tout polynôme Q se décompose de manière unique sous la forme d'une combinaison linéaire de puissances de P, dont les coefficients sont des polynômes dont le degré est strictement inférieur à celui de P. À une telle décomposition, nous associons un mot que nous appelons la P-représentation du polynôme Q. Un ensemble de polynômes est alors qualifié de P-reconnaissable si il existe un automate fini déterministe qui accepte l'ensemble des P-représentations de ses éléments.<BR><BR> Dans cette thèse, nous montrons que les ensembles P-reconnaissables sont exactement ceux qui sont définissables par une formule du premier ordre dans une certaine structure S(P) basée sur un prédicat dépendant du polynôme P. Nous donnons aussi une caractérisation des ensembles P-reconnaissables en terme de suites P-automatiques. Nous apportons également une réponse partielle à la question de savoir quels sont les ensembles reconnaissables simultanément dans toutes les bases de degré au moins 1. Finalement, nous montrons que si P et Q sont deux polynômes de degré au moins 1 et multiplicativement indépendants, alors la multiplication est définissable dans la réunion des structures S(P) et S(Q).

Finite impulse response (FIR) digital filters are frequently used in many digital signal processing and communication applications, such as IS-95 CDMA, Digital Mobile Phone Systems (D-AMPS), etc. FIR filter achieves the frequency response of system requirement using a series of multiplications and additions. Previous papers on FIR hardware implementations usually focus on reducing area and delay of the multiple constant multiplications (MCM) through common sub-expression elimination (CSE) in the transpose FIR filter structure. In this thesis, we first perform optimization for the quantization of FIR filter coefficients that satisfy the target frequency response. Then suitable encoding methods are adopted to reduce the height of the partial products of the MCM in the direct FIR filter structure. Finally, by jointly considering the errors in the truncated multiplications and additions, we can design the hardware-efficient FIR filter that meets the bit accuracy requirement. Experimental results show that although CSE in the transpose FIR structure can reduce more area in MCM, the direct form takes smaller area in registers. Compared with previous approaches, the proposed FIR implementations with direct form has the minimum area cost.

As demand for higher bandwidth I/O grows, the front end design of serial link becomes significant to overcome stringent timing requirements on noisy and bandwidthlimited channels. As a clock reconstructing module in a receiver, the recovered clock quality of Clock and Data Recovery is the main issue of the receiver performance. However, from unknown incoming jitter, it is difficult to optimize loop dynamics to minimize steady-state and dynamic jitter. In this thesis a 10 Gb/s adaptive loop bandwidth clock and data recovery circuit with on-chip loop filter is presented. The proposed system optimizes the loop bandwidth adaptively to minimize jitter so that it leads to an improved jitter tolerance performance. This architecture tunes the loop bandwidth by a factor of eight based on the phase information of incoming data. The resulting architecture performs as good as a maximum fixed loop bandwidth CDR while tracking high speed input jitter and as good as a minimum fixed bandwidth CDR while suppressing wide bandwidth steady-state jitter. By employing a mixed mode predictor, high updating rate loop bandwidth adaptation is achieved with low power consumption. Another relevant feature is that it integrates a typically large off-chip filter using a capacitance multiplication technique that employs dual charge pumps. The functionality of the proposed architecture has been verified through schematic and behavioral model simulations. In the simulation, the performance of jitter tolerance is confirmed that the proposed solution provides improved results and robustness to the variation of jitter profile. Its applicability to industrial standards is also verified by the jitter tolerance passing SONET OC-192 successfully.

The purpose of this study was to examine pre-service mathematics teachers&amp / #8217 / subject matter knowledge, pedagogical content knowledge, and the relationships between the two on multiplication and division of fractions. For this purpose, pre-service teachers&amp / #8217 / understanding of key facts, concepts, principles, and proofs, their knowledge on common (mis)conceptions and difficulties held by the elementary students, their strategies of teaching these concepts, and representations they use to reason their understanding on multiplication and division of fractions were examined. Qualitative case study was performed where / data was collected from the pre-service teachers at the end of the spring semester of 2004-2005. Pre-service teachers were senior students enrolled in a teacher education program at a public university. Results revealed that pre-service teachers could easily symbolize and solve the basic questions on multiplication and division of fractions. However, in terms of pre-service teachers&amp / #8217 / interpretation and reasoning of key facts and principles on multiplication and division of fractions, their subject matter knowledge could not be regarded as conceptually deep. Furthermore, although pre-service teachers have strong belief that they should teach multiplication and division of fractions conceptually, where the logical background of the operations is explained, they do not have sufficient knowledge to represent and explain these topics and relationships conceptually.

TO achieve high level of integration in order to reduce cost, heterodyne architecture has made way for low-IF and zero-IF (direct conversion) receiver architectures. However, a very serious issue in implementing both zero and low-IF receiver is of local oscillator (LO) pulling. Another challenge is on-chip generation of high-precision quadrature LO signals for image-rejection. We have addressed both these issues in this thesis. Regarding the first problem, we have developed a lowpower frequency multiplication technique which uses a low frequency ring oscillator and multiplies its frequency in power e cient way to generate the desired frequency. We then use this differential LO signal to generate high-precision quadrature phases by using polyphase filter and an injection-locked quadrature oscillator. Design examples are presented for 2.4 GHz band of IEEE 802.15.4 standard which is a low-data rate WPAN standard. The standard o ers relaxed performance specifications in order to help achieve low power of operation. Contributions in the thesis • The problem of local oscillator (LO) pulling can be addressed by running LO at a much reduced frequency and use a frequency multiplier (FM) to generate the desired frequency. Also, use of low-frequency LO saves power in VCO and helps eliminate first few dividers leading to significant power savings. In addition, the entire frequency synthesizer can be run at a lower supply voltage saving additional power. The frequency multiplier involves combining edges from the lower frequency ring oscillator. It improves upon the prior work by proposing a new lower-power edge-combiner. The overall power is reduced by exploiting the relaxed phase noise specification of IEEE 802.15.4 standard. Simulations using SpectreRF show that the circuit consumes only 550 ｵW of power in 0.13 ｵm RF-CMOS technology with 1.2 V supply voltage, and provides 950 VP-P sinusoidal output with phase noise of -85.5 dBc/Hz at 1 MHz offset. • An injection-locking based quadrature desensitization circuit is designed for precision quadrature generation. The differential (two phase) output of the frequency multiplier is fed to a polyphase filter to generate nearly quadrature signals. Output of polyphase filter is in turn fed to the desensitizer circuit to obtain high-precision quadrature signals. Designed for 2.4 GHz band in 0.13 µm RF-CMOS technology, it achieves a phase error of 0.5 for 1% mismatch in LC tanks. It achieves a phase noise of -84.3 dBc/Hz at 1 MHz o set and provides quadrature sinusoids of 475 mV amplitude while consuming 1.56 mW of power. • We have analyzed the popular cross-coupled LC-VCOs to generate quadrature sinusoids. In practical LC-oscillators built using low/moderate quality factor on-chip inductors, the actual frequency of oscillation is a little less than 1/2pvLC . This is known as Groszkowski effect. On the other hand, in quadrature oscillator topologies, consisting of two, cross-coupled, negative resistance LC-VCOs using parallel coupling transistors, an upward shift in frequency of oscillation from the free-running frequency of each LC-VCO is observed. This is because in order to satisfy the Barkhausen’s criteria, the LC-tanks have to operate at a frequency away from the frequency of resonance. This e ect called as quadrature detuning effect results in higher phase noise and reduced amplitude. We have shown that the old treatment given in literature is quite inaccurate for practical LC oscillators that are built using low/mo derate Q on-chip inductors. Also the prior work ignores Groszkowski effect which could be significant for low Q LC tanks. We have provided simple, accurate and closed-form expressions of associated frequency-shifts and amplitude of oscillation including both the effects. Our results show excellent match with results obtained from SpectreRF and Matlab simulations.

Rinderpest virus (RPV) belongs to the order Mononegavirale which comprises non segmented negative sense RNA viruses including human pathogens such as Measles, Ebola and Marburg virus. RPV is the causative agent of Rinderpest disease in large ruminants, both domesticated and wild. The viral genome contains a non segmented negative sense RNA encapsidated by nucleocapsid protein (N-RNA). Viral transcription/replication is carried out by the virus encoded RNA dependent RNA polymerase represented by the large protein L and phosphoprotein P as (L-P) complex. Viral transcription begins at the 3’ end of the genome 3’le-N-P-M-F-H-N-tr-5’ with the synthesis of 55nt leader RNA followed by the synthesis of other viral mRNAs. A remarkable feature common to all members of Paramyxoviridae family is the gradient of transcription from 3’ end to the 5’ end of the genome due to attenuation of polymerase transcription at each gene junction. The present study aims at functional characterization of Rinderpest virus transcription and the associated activities required for viral mRNA capping. In addition, an attempt has been made to understand the novel role of a host factor, Ebp1, playing a key role in virus multiplication in infected cells. The specific aims of the study are presented in detail below. 1. Development of in vitro transcription system for RPV mRNA synthesis and role of phosphorylation of P protein in transcription. The transition of viral polymerase from transcription to replication in infected cells has been a long standing puzzle in all paramyxoviruses. Earlier work carried out using RPV minigenome with a CAT reporter gene and studies with phosphorylation null mutant P, has revealed the importance of P phosphorylation for viral transcription in vivo. However, the contribution of other cellular factors in the viral transcription/replication switch could not be ruled out in these assays. In order to understand the specific role of P protein in transcription/replication, it was necessary to develop a cell free transcription system for viral mRNA synthesis. Hence, viral genomic RNA (N-RNA) was purified from RPV infected cells using CsCl density gradient centrifugation. The viral RNA polymerase consisting of L-P complex was separately expressed in insect cells and partially purified by glycerol gradient centrifugation. Glycerol gradient fraction containing the L-P complex was found to be active in viral transcription. Notably, the gradient of transcription of viral mRNA was observed in vitro with the partially purified recombinant L-P complex similar to in vivo. However, the recombinant polymerase complex failed to synthesis the 55nt leader RNA, in agreement with the recent finding in VSV that the transcriptase complex was unable to synthesize leader RNA and viral transcription is initiated at the N gene start site unlike the conventional 3’ entry mode. The newly developed in vitro reconstituted transcription system was used to analyze the effect of P phosphorylation on viral transcription. The results presented in chapter 2, indicate that phosphorylated P supports transcription whereas unphosphorylated P transdominantly inhibits the transcription in vitro suggesting the possible role of the status of P protein phosphorylation in determining transcription/replication switch. 2. Enzymatic activities associated with RPV L protein- role in viral mRNA capping. Post transcriptional modification of mRNA such as capping and methylation determines the translatability of viral mRNA by cellular ribosome. In negative sense RNA viruses, synthesis of viral mRNA is carried out by the viral encoded RNA polymerase in the host cell cytoplasm. Since the host capping and methylation machinery is localized to the nucleus, viruses should either encode their own mRNA modification enzymes or adopt alternative methods as has been reported for orthomyxoviruses (cap snatching) and picornaviruses (presence of IRES element). In order to test, if RPV RNA polymerase possesses any of the capping and methylation activities, both virus as well as the RNP complex containing the viral N-RNA and RNA polymerase (L-P) were purified from infected cells. Using the purified virus and RNP complex, the first two activities required for mRNA capping vis-à-vis, RNA triphosphatase and guanylyltransferase were tested and the results are described in chapter 3 and 4. Purified virus as well as the RNP complex showed both RNA triphosphatase (RTPase) and Nucleotide triphosphatase activities. Neither purified N-RNA or recombinant P proteins show these activities suggesting that it is indeed mediated by viral L protein. By the metal dependency of the reaction and by the motif conservation with other reported RTPases, RPV L protein was assigned to the metal dependent RTPase tunnel family. Capping activity was also seen with the L protein present in RNP complex by its ability to form a covalent complex with GMP moiety of GTP. The specificity of the reaction with GTP, inhibition of Enzyme-GMP complex formation by the inorganic pyrophosphate and the susceptibility of Enzyme-GMP complex under acidic conditions clearly indicated that RPV L represents the viral guanylyl transferase. Further confirmation was obtained by the indirect capping assay in which Enzyme-GMP complex was formed when recombinant L protein was incubated with the cap labeled RNA due to the reversible nature of capping reaction. Owing to the large size of L protein (240 KDa), it is conceivable that the L protein functions in a modular fashion for different activities pertaining to RNA synthesis and modification. Sequence comparison of L proteins from different morbilliviruses revealed the presence of three conserved domains namely domain I (aa 1-606), domain II (aa 650-1694) and domain III (aa 1717-2183). Since domain II has already been assigned as the viral RNA dependent RNA polymerase, domain I and domain III were chosen for further characterization. Both domains were cloned, expressed and purified to homogeneity using recombinant baculovirus expression system. However, the recombinant domain III alone showed the NTPase activity where as neither domain I or III showed RTPase activity. This is expected since a part of the conserved RTPase motif was located in domain II in the multiple sequence alignment with other viral and yeast RTPases. In addition, the recombinant domain III also showed the characteristic enzyme-GMP complex formation but failed to be active in the indirect capping assay. Therefore, both domain II and domain III are likely to be involved in the co-transcriptional capping of viral mRNA. In support of this view, recent report in VSV suggests the presence of additional motif in domain II which is essential for viral mRNA capping. Preliminary evidence has been presented in the appendix section for the presence of N7 guanine methyl transferase activity with L protein although further experiments are needed to confirm this activity. 3. Role of host factor Ebp1 in negative sense RNA virus replication - a possible antagonist In recent years, many cellular factors such as actin, tubulin and profilin have been shown to be involved in viral transcription. Ebp1-ErbB3 binding protein was initially isolated as a cellular protein which binds to Influenza viral polymerase subunit PB1. Ebp1 selectively inhibits the influenza virus transcription in vitro whereas the cap binding and endonuclease activity of PB1 subunit of viral polymerase is unaffected. Till now there are no reports of the role of Ebp1 in non segmented negative sense RNA virus infection. The fifth chapter describes the role of Ebp1 in RPV infection and vice versa. RPV infection leads to down regulation of Ebp1 mRNA levels which in turn leads to decreased protein synthesis. Subsequently, it was found that Ebp1 interacts presumably with viral N protein, being a part of the viral RNP complex in both infected cells as well as in purified virion. Further, over expression of Ebp1 inhibits viral transcription and as a consequence the virus multiplication in vivo suggesting a mutual antagonism between virus and the host cell through Ebp1 protein.

This work is concerned with the applicability of several approximation methods (finite section method, Galerkin and collocation methods with maximum defect splines for uniform and non uniform meshes) to operators belonging to the closed subalgebra generated by operators of multiplication bz piecewise continuous functions and convolution operators also with piecewise continuous generating function.