As we approach exascale systems, power is turning from an optimization goal to a critical operating constraint. With power bounds imposed by both stakeholders and the limitations of existing infrastructure, achieving practical exascale computing will therefore rely on optimizing performance subject to a power constraint. However, this requirement should not add to the burden of application developers; optimizing the runtime environment given restricted power will primarily be the job of high-performance system software. In this dissertation, we explore this area and develop new techniques that extract maximum performance subject to a particular power constraint. These techniques include a method to find theoretical optimal performance, a runtime system that shifts power in real time to improve performance, and a node-level prediction model for selecting power-efficient operating points. We use a linear programming (LP) formulation to optimize application schedules under various power constraints, where a schedule consists of a DVFS state and number of OpenMP threads for each section of computation between consecutive message passing events. We also provide a more flexible mixed integer-linear (ILP) formulation and show that the resulting schedules closely match schedules from the LP formulation. Across four applications, we use our LP-derived upper bounds to show that current approaches trail optimal, power-constrained performance by up to 41%. This demonstrates limitations of current systems, and our LP formulation provides future optimization approaches with a quantitative optimization target. We also introduce Conductor, a run-time system that intelligently distributes available power to nodes and cores to improve performance. The key techniques used are configuration space exploration and adaptive power balancing. Configuration exploration dynamically selects the optimal thread concurrency level and DVFS state subject to a hardware-enforced power bound. Adaptive power balancing efficiently predicts where critical paths are likely to occur and distributes power to those paths. Greater power, in turn, allows increased thread concurrency levels, CPU frequency/voltage, or both. We describe these techniques in detail and show that, compared to the state-of-the-art technique of using statically predetermined, per-node power caps, Conductor leads to a best-case performance improvement of up to 30%, and an average improvement of 19.1%. At the node level, an accurate power/performance model will aid in selecting the right configuration from a large set of available configurations. We present a novel approach to generate such a model offline using kernel clustering and multivariate linear regression. Our model requires only two iterations to select a configuration, which provides a significant advantage over exhaustive search-based strategies. We apply our model to predict power and performance for different applications using arbitrary configurations, and show that our model, when used with hardware frequency-limiting in a runtime system, selects configurations with significantly higher performance at a given power limit than those chosen by frequency-limiting alone. When applied to a set of 36 computational kernels from a range of applications, our model accurately predicts power and performance; our runtime system based on the model maintains 91% of optimal performance while meeting power constraints 88% of the time. When the runtime system violates a power constraint, it exceeds the constraint by only 6% in the average case, while simultaneously achieving 54% more performance than an oracle. Through the combination of the above contributions, we hope to provide guidance and inspiration to research practitioners working on runtime systems for power-constrained environments. We also hope this dissertation will draw attention to the need for software and runtime-controlled power management under power constraints at various levels, from the processor level to the cluster level.

Large-scale wireless sensor networks consist of thousands of tiny and low cost nodes with very limited energy, computing power and communication capabilities. They have a myriad of possible applications. They can be used in hazardous and hostile environments to sense for deadly gases and high temperatures, in personal area networks to monitor vital signs, in military and civilian environments for intrusion detection and tracking, emergency operations, etc. In large scale wireless sensor networks the protocols need to be scalable and energy-efficient. Further, new strategies are needed to address the well-known energy depletion problem that nodes close to the sink node face. In this thesis the Scalable Energy-efficient Location-Aided Routing (SELAR) protocol for wireless sensor networks is proposed to solve the above mentioned problems. In SELAR, nodes use location and energy information of the neighboring nodes to perform the routing function. Further, the sink node is moved during the network operation to increase the network lifetime. By means of simulations, the SELAR protocol is evaluated and compared with two very well-known protocols - LEACH (Low-Energy Adaptive-Clustering Hierarchy) and MTE (Minimum Transmission Energy). The results indicate that in realistic senarios,SELAR delivers up to 12 times more and up to 1.4 times more data packets to the base station than LEACH and MTE respectively. It was also seen from the results that for realistic scenarios, SELAR with moving base station has up to 5 times and up to 27 times more lifetime duration compared to MTE and LEACH respectively.

In this thesis, we investigate two different system infrastructures in underlay cognitive radio network, in which two popular techniques, cross-layer design and cooperative communication, are considered, respectively. In particular, we introduce the Aggressive Adaptive Modulation and Coding (A-AMC) into the cross-layer design and achieve the optimal boundary points in closed form to choose the AMC and A-AMC transmission modes by taking into account the Channel State Information (CSI) from the secondary transmitter to both the primary receiver and the secondary receiver. What’s more, for the cooperative communication design, we consider three different relay selection schemes: Partial Relay Selection, Opportunistic Relay Selection and Threshold Relay Selection. The Probability Density Functions (PDFs) of the Signal-to- Noise Ratio (SNR) in each hop for different selection schemes are provided, and then the exact closed-form expressions for the end-to-end packet loss rate in the secondary link considering the cooperation of the Decode-and-Forward (DF) relay for different relay selection schemes are derived.

L’étude des réseaux de microcapteurs sans fil met clairement en évidence la contrainte principale de l’autonomie en énergie. En effet, ces microcomposants autonomes et communicants appelés aussi nœuds du réseau sont dispersés dans des lieux parfois peu ou pas accessibles. L’objectif de notre travail est de proposer une méthode de conception d’un émetteur-récepteur adapté à ce type de réseaux. Partant d’une modélisation au niveau système mettant en relief la part prépondérante du module radiofréquence sur la consommation moyenne d’un nœud, nous avons développé trois nouveaux outils de conception correspondant à différents niveaux de modélisation de la chaine de communication. Leur utilisation conjointe et les résultats de simulations obtenus nous offrent la possibilité de mettre en relation les spécifications et les performances d’un module radiofréquence avec la consommation. L’association de ces outils dans une méthode de conception itérative nous a permis de dimensionner une chaine de communication en fonction d’une contrainte de consommation. Finalement, nous avons conçu, fabriqué et testé, un amplificateur faible bruit (LNA ou Low Noise Amplifier) à 868 MHz qui présente des caractéristiques très intéressantes en termes de consommation. / Survey on wireless microsensor networks highlights the main constraint of energy autonomy. In fact, these autonomous and communicating microcomponents named network nodes are scattered into few or not open environment. The goal of our work is to propose a transceiver design method adapted to microsensor networks. After a demonstration of predominant part of RF into the mean power consumption of a microsensor node, we developed three new simulation tools which correspond to different level of transceiver modelling. Their use and obtained simulation results demonstrate the relation between transceiver specifications and performances with power consumption. The association of these tools was used to propose a new design method under power consumption constraint. Finally, we designed, produced and tested a 868 MHz Low Noise Amplifier which presents interesting power consumption characteristics.

In recent years, global mobile data traffic has been increasing exponentially as mobile devices pervade our daily lives. To cope with the ever growing demands for higher data rates and seamless connectivity, one solution is to drastically increase the number of macro base stations in the conventional cellular architecture. However, this results in high deployment costs. Deploying low-power nodes such as relays that do not require a wired backhaul connection within a macrocell is one of cost-effective ways to extend high data rate coverage range. Relays are typically deployed to increase signal strength in poor coverage areas or to eliminate dead spots. But more importantly, relays provide a natural diversity, called cooperative diversity. In addition to a direct signal from a base station, extra copies of the same signal are forwarded from relays. Utilizing this diversity at the destination can yield significant performance enhancements. Thus, cooperative relay strategies need to be considered to enable high data rate coverage in a cost-effective manner. In this dissertation, we consider a simple single-relay network and present low-complexity and power-efficient cooperative relay designs that can achieve low error rate. We first study decode-and-forward (DF) relay networks with a single antenna at each node, where the relay decodes the received signal and forwards the re-encoded information to the destination. In DF relay scheme, decoding at the relay is not perfect and the error-propagation phenomenon is a detrimental problem, preventing the destination from collecting the cooperative diversity. To enable cooperative diversity in DF relay networks, we adopt link-adaptive power-scaling relay strategies where the relay scales the transmission power of the re-encoded signal based on the reliability of the source-relay link. We generalize power-profile designs and analyze the diversity order enabled by the general power-profile designs. We provide necessary and sufficient conditions for the designs to enable full cooperative diversity at the destination. In the second part of this dissertation, we extend the power-scaling relay strategy to DF multi-input multi-output (MIMO) relay networks, where multiple antennas are adopted at each node, and show that full cooperative diversity can also be achieved here. To collect spatial diversity provided by multiple antennas without using maximum-likelihood equalizers (MLEs) or near-ML detectors which exhibit high complexity, channel-controlled automatic repeat request (CC-ARQ) scheme is developed for DF MIMO relay networks to enable spatial diversity with linear equalizers (LEs) maintaining low-complexity. We also show that joint cooperative and spatial diversity can be achieved at the destination when the power-scaling strategy and the CC-ARQ with LEs are combined. Finally, amplify-and-forward (AF) MIMO relay designs, where the relay simply amplifies the received signal and forwards it to the destination, are studied with consideration of peak-power constraints at the relay. One practical concern for AF relaying is that the output signal at the relay may suffer from large peak-to-average power ratio (PAR), which may cause nonlinear distortion and/or saturation in the transmitted signal due to the limited linear range of power amplifiers. Thus, we first investigate peak-power constrained power-scaling strategies and find a sufficient condition to enable joint cooperative and spatial diversity at the destination. Based on this study, we propose simple and practical AF MIMO relay designs with peak-power constraint at the relay. CC-ARQ is also applied to AF MIMO relay networks to reduce the decoding complexity.

Τα τελευταία χρόνια και κυρίως μετά το 1996 έχει παρουσιαστεί ένα έντονο ενδιαφέρον γύρω από τα ασύρματα συστήματα MIMO (Συστήματα Πολλών Εισόδων και Πολλών Εξόδων). Η βασική αιτία που δημιούργησε αυτό το ενδιαφέρον ήταν η θεωρητική ανάλυση της χωρητικότητας που επιτυγχάνεται με τη χρήση πολλαπλών κεραιών τόσο στο δέκτη όσο και στον πομπό. Ωστόσο πέρα από την αύξηση της χωρητικότητας τα συστήματα MIMΟ έχουν μια σειρά από επιπλέον πλεονεκτήματα, επιτυγχάνουν ανεκτικότητα στην εξασθένιση, αυξάνουν την φασματική αποδοτικότητα, μειώνουν την κατανάλωση ενέργειας καθώς και το κόστος χρήσης και κατασκευής ασύρματων δικτύων. Έχοντας λάβει υπ’ όψιν τα παραπάνω και έχοντας κάνει ήδη μια αρχική έρευνα στο τι υπάρχει στην επιστημονική βιβλιογραφία γύρω από τα συστήματα MIMO, επιλέξαμε να ασχοληθούμε με την συγκεκριμένη τεχνολογία και πιο συγκεκριμένα με την προ-κωδικοποίηση πομπού, όταν τα κανάλια που διαθέτουμε είναι συχνοτικά επιλεκτικά (frequency selective). Το πρόβλημα που μας απασχόλησε είναι η διερεύνηση τεχνικών κωδικοποίησης των δεδομένων του πομπού, πριν αυτά μεταδοθούν, ούτως ώστε να μπορέσει να απλοποιηθεί η σχεδίαση του δέκτη. Στην περίπτωση των συχνοτικά επιλεκτικών καναλιών πέρα από τον πανταχού παρόντα Γκαουσιανό θόρυβο έχουμε να αντιμετωπίσουμε και τη διασυμβολική παρεμβολή (Intersymbol Interference - ISI) . Στόχος μας είναι να μπορέσουμε να μετατοπίσουμε τη διαδικασία της ισοστάθμισης από το δέκτη στον πομπό, όπου, στην περίπτωση που ο πομπός είναι ένας σταθμός βάσης, η δυνατότητα υλοποίησης πολύπλοκων διεργασιών είναι μεγαλύτερη, τόσο μεγαλύτερης ανοχής στην κατανάλωση ενέργειας όσο κυρίως, και στη δυνατότητα χρήσης πολύπλοκου υλικού. Το πρόβλημα που τίθεται στην περίπτωση που η ισοστάθμιση γίνει στην μεριά του πομπού είναι η ποιότητα και η ποσότητα της γνώσης του πραγματικού καναλιού. Είναι σχεδόν αδύνατο ο πομπός να έχει πλήρη γνώση του πραγματικού καναλιού, με αποτέλεσμα να πρέπει να αρκεστεί σε γνώση, που στην καλύτερη περίπτωση προσεγγίζει αυτή του πραγματικού καναλιού. Επίσης, πολλές φορές, η ποσότητα της γνώσης που μπορεί να αποσταλεί από το δέκτη – όταν δεν ισχύει η αρχή της αμοιβαιότητας (reciprocity) – είναι περιορισμένη. Από τα παραπάνω καταλαβαίνουμε ότι θα είχε ενδιαφέρον η μελέτη τεχνικών που παρουσιάζουν ανοχή στην ποιότητα της γνώσης του καναλιού καθώς και η συμπεριφορά τους όταν αυτές διαθέτουν μερική ή και περιορισμένη γνώση του πραγματικού καναλιού. Ένας από τους τρόπου αντιμετώπισης όταν συναντούνται συχνοτικά επιλεκτικά κανάλια είναι η χρήση του OFDM με όσα αρνητικά αυτό συνεπάγεται (π.χ. η δυσκολία συγχρονισμού πομπού και δέκτη). Ένας άλλος τρόπος αντιμετώπισης, στην περίπτωση μονής φέρουσας (single carrier), είναι η χρήση προκωδικοποιητων Bezout. Επιλέξαμε να ασχοληθούμε με αυτόν τον τύπου προκωδικοποίητων. Πιο συγκεκριμένα, όταν ο αριθμός των κεραιών του πομπού είναι μεγαλύτερος από τον αριθμό των κεραιών του δέκτη, είναι δυνατόν να εφαρμόσουμε την ταυτότητα Bezout (Bezout Identity) στην μεριά του δέκτη. Με απλά λόγια θα εφαρμόσουμε ένα προ-ισοσταθμιστή επιβολής μηδενικών (zero forcing). Όπως είναι αναμενόμενο, μιας και ο προκωδικοποιητής εφαρμόζεται στην μεριά του πομπού, πριν εμφανιστεί ο Γκαουσιανός θόρυβος, δεν έχουμε ενίσχυση αυτού του θορύβου, ωστόσο είναι δυνατόν να έχουμε σημαντική αύξηση της μεταδιδόμενης ισχύος. Για να λυθεί αυτό το πρόβλημα έχει προταθεί στη βιβλιογραφία η σχεδίαση προκωδικοποιητών Bezout με την χρήση περιορισμών ισχύος. Αυτό που εμείς μελετήσαμε στην περίπτωση των προκωδικοποιητών Bezout (ή FIR προκωδικοποιητών), είναι η συμπεριφορά τους στην περίπτωση που ο πομπός διαθέτει πλήρη ή μερική γνώση του καναλιού. Είδαμε την περίπτωση του σχεδιασμού ενός FIR προκωδικοποιητή κάτω από τον περιορισμό ισχύος μετάδοσης μέσω της μεθόδου Tikhonov Regularization, ενός σημαντικού εργαλείου επίλυσης του προβλήματος των ελαχίστων τετραγώνων κάτω από τη δι-κριτήριο διατύπωση. Τέλος, προτείναμε δύο FIR σχεδιασμούς προκωδικοποιητών για την περίπτωση που ο πομπός διαθέτει παραμετρική (στατιστική) περιγραφή του καναλιού. / The last years, but mainly after 1996 there is an intensive interest in MIMO systems. The reason that created this interest was the capacity that can be achieved by the use of multiple antennas to the transmitter and to receiver. Furthermore, the use of multiple antennas has a number of advantages except from the capacity increase; MIMO can achieve tolerance to fading, spectral efficiency, increased coverage etc. By having considered the above information we decided to work with MIMO systems and especially with precoding techniques for frequency selective channels. In the case of frequency selective channels except from the Gaussian noise we have to treat and with intersymbol interference (ISI) which is an important degradation factor. A very serious reason to handle intersymbol interference at the transmitter side - especially when transmitter is a base station - is to simplify receiver's design. When transmitter is a base station there is a tolerance in power consuming and in the implementation of more complicated hardware. After an extensive bibliographical we concluded to use the theory of Bezout Identity. In the case where the number of antennas at the transmitter is greater than the number of antennas at the receiver it is formed a left-coprime FIR channel. By using the Bezout identity matrix, it is possible to be designed a FIR MIMO precoder that reduce intersymbol interference. Bezout precoder (pre-equalizer) is a zero-forcing (ZF) equalizer. As it is known from bibliography ZF equalizer has the disadvantage that amplifies noise power in deep fades. For the case of Bezout precoder it is not true since there is no noise - noise will appear at the receiver. But there is another disadvantage, Bezout precoder may increase transmit power significantly to overcome deep fades in the singular values of the channel matrix. The solution to the previous phenomenon can be taken by designing ZF precoders under power constraints. The main problem in the case of transmit precoding is the channel knowledge - when reciprocity principle is not valid. In most cases it is not possible to have perfect channel knowledge or even the knowledge that we have may be of bad quality. Another drawback that appears in the case of frequency selective channels is the amounts of information (number of bits) that can be send back to the transmitter - limited feedback. Many works assumed a quasi-static channel - channel doesn't changes for a number of symbols - and perfect channel knowledge. In real world this may not be true. In this master thesis we investigated three things. Firstly, the behavior of Bezout precoders under channel mismatch at the side of transmitter. Secondly, ways that will allows us to decrease channel feedback. And thirdly, which was the most interesting, we proposed two FIR designs that uses statistical channel knowledge.

No description available.

The importance of cooperative relaying communication in substituting for, or complementing, multiantenna systems is described, and a brief literature review is presented. Amplify-and-forward (AF) and decode-and-forward (DF) relaying are investigated and compared for a dual-hop relay channel. The optimal strategy, source and relay optimal power allocation, and maximum cooperative gain are determined for the relay channel. It is shown that while DF relaying is preferable to AF relaying for strong source-relay links, AF relaying leads to more gain for strong source-destination or relay-destination links. Superimposed and selection AF relaying are investigated for multirelay, dual-hop relaying. Selection AF relaying is shown to be globally strictly outage suboptimal. A necessary condition for the selection AF outage optimality, and an upper bound on the probability of this optimality are obtained. A near-optimal power allocation scheme is derived for superimposed AF relaying. The maximum instantaneous rates, outage probabilities, and average capacities of multirelay, dual-hop relaying schemes are obtained for superimposed, selection, and orthogonal DF relaying, each with parallel channel cooperation (PCC) or repetition-based cooperation (RC). It is observed that the PCC over RC gain can be as much as 4 dB for the outage probabilities and 8.5 dB for the average capacities. Increasing the number of relays deteriorates the capacity performance of orthogonal relaying, but improves the performances of the other schemes. The application of rateless codes to DF relaying networks is studied by investigating three single-relay protocols, one of which is new, and three novel, low complexity multirelay protocols for dual-hop networks. The maximum rate and minimum energy per bit and per symbol are derived for the single-relay protocols under a peak power and an average power constraint. The long-term average rate and energy per bit, and relay-to-source usage ratio (RSUR), a new performance measure, are evaluated for the single-relay and multirelay protocols. The new single-relay protocol is the most energy efficient single-relay scheme in most cases. All the multirelay protocols exhibit near-optimal rate performances, but are vastly different in the RSUR. Several future research directions for fixed-rate and rateless coded cooperative systems, and frameworks for comparing these systems, are suggested. / Communications

No description available.