Imagine your local creamery administers a survey asking their patrons to choose their five favorite ice cream flavors. Any data collected by this survey would be an example of partially ranked data, as the set of all possible flavors is only ranked into subsets of the chosen flavors and the non-chosen flavors. If the creamery asks you to help analyze this data, what approaches could you take? One approach is to use the natural symmetries of the underlying data space to decompose any data set into smaller parts that can be more easily understood. In this work, I describe how to use permutation representations of the symmetric group to create and study efficient algorithms that yield such decompositions.

Τα Ασύρματα Δίκτυα Αισθητήρων (ΑΔΑ) αποτελούν μία αρκετά πρόσφατη και πολλά υποσχόμενη νέα τεχνολογία που βρίσκει πληθώρα εφαρμογών. Λόγω της ευρύτατης εφαρμοσιμότητάς της και της προφανούς θέσης που βρίσκει στο σύγχρονο κατανεμημένο υπολογιστικό κόσμο, η επιστημονική τυπική θεμελίωση των νόμων που διέπουν αυτή τη νέα τεχνολογία καθίσταται απαραίτητη. Έτσι, έχουν προταθεί πολλά νέα υπολογιστικά μοντέλα για ΑΔΑ. Μία ειδική κατηγορία τέτοιων συστημάτων είναι τα Πρωτόκολλα Πληθυσμών (ΠΠ). Αυτά διέπονται από τρία ιδιαίτερα χαρακτηριστικά: Οι κόμβοι αίσθησης (πράκτορες) κινούνται παθητικά, δηλαδή δε μπορούν να ελέγξουν την κίνηση στην οποία υπόκεινται, η διαθέσιμη μνήμη κάθε κόμβου είναι πολύ περιορισμένη και οι πράκτορες αλληλεπιδρούν κατά ζεύγη. Έχει αποδειχθεί ότι ένα κατηγόρημα είναι υπολογίσιμο από το μοντέλο των ΠΠ εάν και μόνο εάν είναι ημιγραμμικό. Η κλάση των ημιγραμμικών κατηγορημάτων αποτελεί μία αρκετά μικρή κλάση. Στην παρούσα εργασία, βασικός μας στόχος είναι η επέκταση του μοντέλου των πρωτοκόλλων πληθυσμών με σκοπό το κέρδος σε υπολογιστική ισχύ. Πρώτα κάνουμε την παραδοχή ότι, πέρα των κόμβων αίσθησης, και οι ακμές του γραφήματος μπορούν να διατηρούν περιορισμένες καταστάσεις. Έτσι, σε ένα πλήρες γράφημα n κόμβων είναι σα να έχουμε προσθέσει Ο(n^2) επιπλέον θέσεις μνήμης οι οποίες διαβάζονται και γράφονται μόνο από τα άκρα της αντίστοιχης ακμής. Αποδεικνύουμε ότι το νέο μοντέλο, το οποίο καλούμε μοντέλο Πρωτοκόλλων Πληθυσμών με Διαμεσολαβητή, μπορεί να λειτουργήσει ως μία κατανεμημένη ανταιτιοκρατική μηχανή Turing (ΜΤ) που χρησιμοποιεί όλη τη διαθέσιμη μνήμη. Η μόνη διαφορά από μία συνήθη ΜΤ είναι ότι η συγκεκριμένη μηχανή υπολογίζει μόνο συμμετρικές γλώσσες. Πιο τυπικά, δείχνουμε ότι ένα κατηγόρημα είναι υπολογίσιμο από το νέο μοντέλο εάν και μόνο εάν είναι συμμετρικό και ανήκει στην NSPACE(n^2). Επιπλέον, μελετάμε και τη δυνατότητα του νέου μοντέλου να διαγιγνώσκει γλώσσες γραφημάτων (για γενικά γραφήματα). Εν συνεχεία, αγνοούμε τις καταστάσεις των ακμών και δίνουμε μία νέα βελτίωση και πάλι απευθείας απ' το μοντέλο των ΠΠ. Η υπόθεση που κάνουμε τώρα είναι ότι οι πράκτορες είναι πολυταινιακές ΜΤ με άπειρη μνήμη, που μπορούν τόσο να εκτελούν εσωτερικό υπολογισμό όσο και να αλληλεπιδρούν με άλλους πράκτορες και ορίζουμε χωρικά φραγμένους υπολογισμούς. Καλούμε το νέο αυτό μοντέλο, μοντέλο Παθητικά κινούμενων Μηχανών. Αποδεικνύουμε ότι αν χρησιμοποιείται σε κάθε πράκτορα μνήμη το πολύ f(n) για f(n)=Ω(log n) τότε ένα κατηγόρημα είναι υπολογίσιμο από το νέο μοντέλο εάν και μόνο εάν είναι συμμετρικό και ανήκει στην NSPACE(nf(n)). Δείχνουμε επίσης ότι αυτό δεν ισχύει για f(n)=o(log n). Βασιζόμενοι σε αυτά, δείχνουμε ότι για f(n)=Ω(log n) υπάρχει μία χωρική ιεραρχία ακριβώς όπως και για τις συνήθεις (συμμετρικές) ΜΤ. Δείχνουμε επίσης ότι αυτό δεν ισχύει για f(n)=o(loglog n), καθώς στην τελευταία περίπτωση η αντίστοιχη κλάση καταρρέει μέσα στην κλάση των ημιγραμμικών κατηγορημάτων, και τέλος ότι για f(n)=Ω(loglog n) η κλάση γίνεται αυστηρά μεγαλύτερη των ημιγραμμικών κατηγορημάτων. Αφήνουμε ανοικτό το πρόβλημα του τι ακριβώς συμβαίνει για χωρικά φράγματα f(n) τέτοια ώστε f(n)=Ω(loglog n) και f(n)=o(log n). / Wireless Sensor Networks (WSNs) constitute a recent and promising new technology that is widely applicable. Due to the applicability of this technology and its obvious importance for the modern distributed computational world, the formal scientific foundation of its inherent laws becomes essential. As a result, many new computational models for WSNs have been proposed. Population Protocols (PPs) are a special category of such systems. These are mainly identified by three distinctive characteristics: the sensor nodes (agents) move passively, that is, they cannot control the underlying mobility pattern, the available memory to each agent is restricted, and the agents interact in pairs. It has been proven that a predicate is computable by the PP model iff it is semilinear. The class of semilinear predicates is a fairly small class. In this work, our basic goal is to enhance the PP model in order to improve the computational power. We first make the assumption that not only the nodes but also the edges of the communication graph can store restricted states. In a complete graph of n nodes it is like having added O(n^2) additional memory cells which are only read and written by the endpoints of the corresponding edge. We prove that the new model, called Mediated Population Protocol model, can operate as a distributed nondeterministic Turing machine (TM) that uses all the available memory. The only difference from a usual TM is that this one computes only symmetric languages. More formally, we establish that a predicate is computable by the new model iff it is symmetric and belongs to NSPACE(n^2). Moreover, we study the ability of the new model to decide graph languages (for general graphs). The next step is to ignore the states of the edges and provide another enhancement straight away from the PP model. The assumption now is that the agents are multitape TMs equipped with infinite memory, that can perform internal computation and interact with other agents, and we define space-bounded computations. We call this the Passively mobile Machines model. We prove that if each agent uses at most f(n) memory for f(n)=Ω(log n) then a predicate is computable iff it is symmetric and belongs to NSPACE(nf(n)). We also show that this is not the case for f(n)=o(log n). Based on these, we show that for f(n)=Ω(log n) there exists a space hierarchy like the one for classical symmetric TMs. We also show that the latter is not the case for f(n)=o(loglog n), since here the corresponding class collapses in the class of semilinear predicates and finally that for f(n)=Ω(loglog n) the class becomes a proper superset of semilinear predicates. We leave open the problem of characterizing the classes for f(n)=Ω(loglog n) and f(n)=o(log n).

[pt] As denominadas construções tough (tough constructions - TCs), como John is easy to please, apresentam o formato superficial [SD COPULA Adj VInf], tendo como característica distintiva o fato de que o SD (sintagma determinante), sujeito do predicado, é interpretado como objeto lógico do verbo no infinitivo. Muitos estudos teóricos examinaram a sintaxe dessa estrutura e as relações semânticas tecidas a partir dela, mas até o momento nenhuma teoria explica de forma satisfatória as TCs e estruturas relacionadas, com objeto lacunado, como pretty predicates, gapped degree phrases e ready predicates. Os objetivos desta dissertação são: (a) discutir, à luz dos dados, as diferentes propostas teóricas para TCs e construções relacionadas, buscando sistematizar nosso conhecimento atual sobre a gramática dessas construções em inglês e em algumas línguas românicas, como o francês, o espanhol, e o romeno; (b) apresentar uma nova abordagem teórica, construída a partir da negação das premissas adotadas em estudos anteriores, sugerindo que, em TCs e estruturas relacionadas, a formação do predicado envolva nominalização e adjunção de projeções máximas, resultando em estruturas simétricas. Na análise proposta, o elo semântico entre o sujeito sintático do predicado e o objeto lógico do verbo no infinitivo se realiza na Forma Lógica, em função da estrutura argumental do verbo nominalizado e da categoria do adjetivo concatenado simetricamente. A modelagem teórica apresentada permite unificar TCs, pretty predicates, gapped degree phrases e ready predicates, propondo que essas construções compartilham os processos de nominalização e de adjunção de projeções máximas, diferenciando-se nas relações semânticas entre seus elementos. Permite também uma explicação mais adequada para o comportamento de for phrases e para o fato de que apenas as TCs (e algumas gapped degree phrases) aparecem em posição prenominal, com uso atributivo. / [en] Superficially, tough constructions (TCs) like John is easy to please present the format [DP COPULA A VINF], but their distinctive characteristic is that the subject DP is interpreted as the logical object of the verb in the infinitive. Many theoretical studies have examined the syntactic structure of TCs and the semantic relations supported by that structure, but to date no theory provides a good explanatory description for TCs and related object-gap constructions like pretty predicates, gapped degree phrases, and ready predicates. The objectives of this thesis are (a) to discuss, in light of the data, the different theoretical proposals on TCs and related constructions, providing a systematic overview of current knowledge on the grammar of these constructions in English and a number of Romance languages, such as French, Spanish and Romanian, and (b) to present a new theoretical approach, starting from a contrarian position on the premises adopted in earlier studies. Under the proposed analysis, formation of the predicate in TCs and related constructions involves nominalization of the verb in the infinitive and adjunction of maximal projections, resulting in a symmetrical structure. In this symmetrical structure, the semantic link between syntactic subject and logical object is established at Logical Form, in function of the nominalized verb s argument structure and the category of the concatenated adjective. The theoretical model presented in this thesis unifies TCs, pretty predicates, gapped degree phrases and ready predicates, suggesting that these constructions share the processes of nominalization and adjunction of maximal projections, and differ in the semantic relationships between their elements. It also offers a principled explanation for the behavior of for-phrases in these object-gap constructions and for the fact that only TCs (and some gapped degree phrases) appear prenominally, in attributive use.

L'accroissement d'efficacite des réseaux d'interconnexion et la vulgarisation des machines multiprocesseurs permettent la réalisation de machines parallèles a mémoire distribuée de faible coût: les grappes de multiprocesseurs. Elles nécessitent l'exploitation à la fois du parallélismeà grain fin, interne à un multiprocesseur offert par la multiprogrammation légère, et du parallélisme à gros grain entre les différents multiprocesseurs. L'exploitation simultanée de ces deux types de parallélisme exige une méthode de communication entre les processus légers qui ne partagent pas le mêmme espace d'adressage. Le travail de cette thèse porte sur le problème de l'Intégration de la multiprogrammation légère et des communications sur grappes de multiprocesseurs symétriques (SMP). II porte plus précisément sur evaluation et le reglage du noyau exécutif ATHAPASCAN-0 sur ce type d'architecture. ATHAPASCAN-0 est un noyau exécutif, portable, développé au sein du projet APACHE (CNRS-INPG-INRIA-UJF), qui combine la multiprogrammation légère et la communication par échange de messages. La portabilité est assurée par une organisation en couches basée sur les standards POSIX threads et MPI largement répandus. ATHAPASCAN-0 étend le modèle de réseau statique de processus «lourds» communicants tel que MPI, PVM, etc,à celui d'un réseau dynamique de processus légers communicants. La technique de base est la multiprogrammation lègere des communications et des calculs. La progression des communications exige la scrutation de état du reseau et l'enchainement des opérations de transferts. L'efficacité repose sur la minimisation de ces opérations. De plus, l'emploi de multiprocesseurs ajoute des problèmes spécifiques dus à l'apparition d'un parallélisme réel entre calcul et communication. Ces problèmes sont présentés et des solutions sont proposées pour l'environnement ATHAPASCAN-0. Ces solutions sont évaluées sur des grappes de multiprocesseurs. / The continuous price reduction for commodity PC multiprocessors and the availability of fast network interfaces have made cluster of multiprocessors an attractive low-price alternative to build parallel systems. Multiprocessor clusters offer two levels of parallelism: a fine grain parallelism inside a single multiprocessor and a coarse grain among them. A mechanism must be provided to exploit both levels of parallelism simultaneously. This requires to provide communications between threads belonging to different addresses spaces. This dissertation addresses the problem of integrating threads and communications on ATHAPASCAN-0 run time system. ATHAPASCAN-0 is a portable run time for cluster of multiprocessors developed as part of the APACHE project (CNRS-INPG-INRIA-UJF). Portability is achieved by a layered organization based on standards like POSIX threads and MPI. The ATHAPASCAN-0 run time system extends the heavy-weight process communication model of message passing libraries such as MPI, PVM, etc, into a lighter dynamic network of communicating threads. Multiprogramming is the key concept used. Communication progress is based on a network polling basis to handle incoming messages and to deliver outgoing communications requests. Performance is strongly dependent on the way these operations are implemented. Additionally, multiprocessors introduce some programming problems like overhead of cache coherency mechanisms, method of managing concurrent accesses and efficient mutex locking to avoid unnecessary context switching. These problems are analyzed and solutions are implemented in the ATHAPASCAN-0 run time system. An evaluation of these solutions is performed on a cluster of multiprocessors.

VASCONCELOS, Rosa Tayane de. O tensor de Ricci e campos de killing de espaços simétricos. 2017. 81 f. Dissertação (Mestrado em Matemática)- Centro de Ciências, Universidade Federal do Ceará, Fortaleza, 2017. / Submitted by Andrea Dantas (pgmat@mat.ufc.br) on 2017-09-18T13:45:50Z No. of bitstreams: 1 2017_dis_rtvasconcelos.pdf: 555452 bytes, checksum: 4ff6c8fb7950682913acabed03e9d3d7 (MD5) / Rejected by Rocilda Sales (rocilda@ufc.br), reason: Boa tarde, A Dissertação de ROSA TAYANE DE VASCONCELOS apresenta a alguns erros que devem corrigidos, os mesmos seguem listados abaixo: 1- EPÍGRAFE (coloque o nome do autor da epígrafe todo em letra maiúscula) 2- RESUMO/ ABSTRACT (retire o recuo dos parágrafos do resumo e do abstract) 3- PALAVRAS-CHAVE/ KEYWORDS (coloque a letra inicial do primeiro elemento das palavras- -chave e das Keywords em maiúscula) 4- CITAÇÕES (as citações a autores, que aparecem em todo o trabalho, não estão no padrão ABNT: se for apenas uma referência geral a uma obra, deve se colocar o último sobrenome do autor em letra maiúscula e o ano da publicação, ex.: EBERLEIN (2005). Caso seja a citação de um trecho particular da obra deve acrescentar o número da página, ex.: EBERLEIN (2005, p. 30). OBS.: as citações não devem estar entre colchetes. 5- TÍTULOS DOS CAPÍTULOS E SEÇÕES (coloque os títulos dos capítulos e seções em negrito) 6- REFERÊNCIAS (as referências bibliográficas não estão no padrão ABNT: apenas o último sobrenome do autor, que inicia a referência, deve estar em letra maiúscula, o restante do nome deve estar em letra minúscula. EX.: BROCKER, Theodor; TOM DIECK, Tammo. Representations of compact Lie groups, v. 98. Springer Science & Business Media, 2013. Atenciosamente, on 2017-09-18T15:04:06Z (GMT) / Submitted by Andrea Dantas (pgmat@mat.ufc.br) on 2017-09-19T13:33:40Z No. of bitstreams: 1 2017_dis_rtvasconcelos.pdf: 522079 bytes, checksum: ff99004fbe22e922f704a6a87365d3b6 (MD5) / Approved for entry into archive by Rocilda Sales (rocilda@ufc.br) on 2017-09-21T12:18:22Z (GMT) No. of bitstreams: 1 2017_dis_rtvasconcelos.pdf: 522079 bytes, checksum: ff99004fbe22e922f704a6a87365d3b6 (MD5) / Made available in DSpace on 2017-09-21T12:18:22Z (GMT). No. of bitstreams: 1 2017_dis_rtvasconcelos.pdf: 522079 bytes, checksum: ff99004fbe22e922f704a6a87365d3b6 (MD5) Previous issue date: 2017-09-13 / This work brings a smooth and self-contained introduction to the study of the most basic aspects of symmetric spaces, having as its nal goal the characterization of the Killing vector fields and of the Ricci tensor of such riemannian manifolds. Several of the results presented in the initial chapter are not easily found, in the Diferential Geometry literature, in a way as accessible and self-contained as here. This being said, we believe that this work embodies some didactic relevance, for it others students interested in symmetric spaces a relatively smooth first contact. We shall generally look at symmetric spaces as homogeneous manifolds G=H, where G is a Lie group and H is a closed Lie subgroup of G, such that the natural mapping : G ! G=H is a riemannian submersion. Ultimately, this map allows us to describe the relationships between the curvature, the Ricci tensor and the geodesics of G and G=H. For our purposes, the crucial remark is that, under appropriate circumstances, one guarantees the existence, in G=H, of a metric for which left translations are isometries. Hence, a one-parameter family of such isometries gives rise to a Killing vector field, which turn into a Jacobi vector eld when restricted to a geodesic. We present explicit expressions for such Jacobi vector elds, showing that they only depend on the eigenvalues of the linear operator TX : g ! g given by TX = (adX)2, for certain vector elds X 2 g. / Este trabalho traz uma introdução suave e autocontida ao estudo dos aspectos mais básicos de espaços simétricos, tendo como objetivo final a caracterização dos campos de Killing e do tensor de Ricci de tais variedades riemannianas. Vários dos resultados obtidos nos capítulos iniciais não são encontrados, na literatura de Geometria Diferencial, de maneira tão acessível e autocontida como apresentados aqui. Com isso, acreditamos que o trabalho reveste-se de alguma relevância didática, por oferecer aos alunos interessados no estudo de espaços simétricos um primeiro contato relativamente suave. Em linhas gerais, veremos espaços simétricos como variedades homogêneas G=H, onde G e um grupo de Lie e H um subgrupo de Lie fechado de G, tais que a aplicação natural: G ! G=H seja uma submersão riemanniana. Através dela, descrevemos relações entre a curvatura, o tensor de Ricci e as geodésicas de G e G=H. Para nossos propósitos, a observação crucial e que, sob certas hipóteses, garantimos a existência, em G=H, de uma métrica cujas translações a esquerda são isometrias. Portanto, uma família a um parâmetro de tais isometrias d a origem a um campo de Killing que, por sua vez, restrito a geodésicas torna-se um campo de Jacobi. Apresentamos expressões para esses campos de Jacobi, mostrando que os mesmos só dependem dos autovalores do operador linear TX : g ! g dado por TX = (adX)2, para certos campos X 2 g.

Provamos resultados sobre a geometria de hipersuperfícies em diferentes espaços ambiente. Primeiro, definimos uma aplicação de Gauss generalizada para uma hipersuperfície Mn-1 c/ Nn, onde N é um espaço simétrico de dimensão n ≥ 3. Em particular, generalizamos um resultado de Ruh-Vilms e apresentamos aplicações. Em seguida, estudamos superfícies em espaços de dimensão 3: estudamos a equação da curvatura média em um produto semidireto R2oAR e obtemos estimativas da altura e a existência de gráficos mínimos do tipo Scherk. Finalmente, no espaço ambiente de uma variedade hiperbólica de dimensão 3: nós apresentamos condições suficientes para que um mergulho completo de uma superfície ∑ de topologia finita em N com curvatura média |H∑| ≤ 1 seja próprio. / We prove results concerning the geometry of hypersurfaces on di erent ambient spaces. First, we de ne a generalized Gauss map for a hypersurface Mn-1 c/ Nn, where N is a symmetric space of dimension n ≥ 3. In particular, we generalize a result due to Ruh-Vilms and make some applications. Then, we focus on surfaces on spaces of dimension 3: we study the mean curvature equation of a semidirect product R2 oA R to obtain height estimates and the existence of a Scherk-like minimal graph. Finally, on the ambient space of a hyperbolic manifold N of dimension 3 we give su cient conditions for a complete embedding of a nite topology surface ∑ on N with mean curvature |H∑| ≤ 1 to be proper.

Chapter 1 reviews better-known factorization theorems of a square matrix. For example, a square matrix over a field can be expressed as a product of two symmetric matrices; thus square matrices over real numbers can be factorized into two diagonalizable matrices. Factorizing matrices over complex num hers into Hermitian matrices is discussed. The chapter concludes with theorems that enable one to prescribe the eigenvalues of the factors of a square matrix, with some degree of freedom. Chapter 2 proves that a square matrix over arbitrary fields (with one exception) can be expressed as a product of two diagona lizab le matrices. The next two chapters consider decomposition of singular matrices into Idempotent matrices, and of nonsingutar matrices into Involutions. Chapter 5 studies factorization of a comp 1 ex matrix into Positive-( semi )definite matrices, emphasizing the least number of such factors required / Mathematical Sciences / M.Sc. (MATHEMATICS)

Chapter 1 reviews better-known factorization theorems of a square matrix. For example, a square matrix over a field can be expressed as a product of two symmetric matrices; thus square matrices over real numbers can be factorized into two diagonalizable matrices. Factorizing matrices over complex numbers into Hermitian matrices is discussed. The chapter concludes with theorems that enable one to prescribe the eigenvalues of the factors of a square matrix, with some degree of freedom. Chapter 2 proves that a square matrix over arbitrary fields (with one exception) can be expressed as a product of two diagonalizable matrices. The next two chapters consider decomposition of singular matrices into Idempotent matrices, and of nonsingular matrices into Involutions. Chapter 5 studies factorization of a complex matrix into Positive-(semi)definite matrices, emphasizing the least number of such factors required. / Mathematical Sciences / M. Sc. (Mathematics)

Made available in DSpace on 2016-06-02T20:27:38Z (GMT). No. of bitstreams: 1 3466.pdf: 565162 bytes, checksum: 770041f07c68eda588bd0c501dabe93d (MD5) Previous issue date: 2011-02-10 / Financiadora de Estudos e Projetos / In this work we analyze the existence of radially symmetric solutions, positive solutions as well as the existence of ground state solutions for a class of Klein-Gordon-Maxwell equations when the nonlinearity exhibits critical behavior. For the positive and ground state solutions we prove existence results when a potential V is introduced. In order to obtain such results, we use variational methods / Neste trabalho analisamos a existência de soluções radialmente simétricas, soluções positivas, bem como a existência de soluções ground state para uma classe de equações do tipo Klein-Gordon-Maxwell quando a não-linearidade exibe comportamento crítico. Para as soluções positivas e do tipo ground state provamos resultados de existência quando um potencial V é introduzido. A fim de obtermos tais resultados, usamos métodos variacionais.

L'accroissement d'efficacite des réseaux d'interconnexion et la vulgarisation des machines multiprocesseurs permettent la réalisation de machines parallèles a mémoire distribuée de faible coût: les grappes de multiprocesseurs. Elles nécessitent l'exploitation à la fois du parallélismeà grain fin, interne à un multiprocesseur offert par la multiprogrammation légère, et du parallélisme à gros grain entre les différents multiprocesseurs. L'exploitation simultanée de ces deux types de parallélisme exige une méthode de communication entre les processus légers qui ne partagent pas le mêmme espace d'adressage. Le travail de cette thèse porte sur le problème de l'Intégration de la multiprogrammation légère et des communications sur grappes de multiprocesseurs symétriques (SMP). II porte plus précisément sur evaluation et le reglage du noyau exécutif ATHAPASCAN-0 sur ce type d'architecture. ATHAPASCAN-0 est un noyau exécutif, portable, développé au sein du projet APACHE (CNRS-INPG-INRIA-UJF), qui combine la multiprogrammation légère et la communication par échange de messages. La portabilité est assurée par une organisation en couches basée sur les standards POSIX threads et MPI largement répandus. ATHAPASCAN-0 étend le modèle de réseau statique de processus «lourds» communicants tel que MPI, PVM, etc,à celui d'un réseau dynamique de processus légers communicants. La technique de base est la multiprogrammation lègere des communications et des calculs. La progression des communications exige la scrutation de état du reseau et l'enchainement des opérations de transferts. L'efficacité repose sur la minimisation de ces opérations. De plus, l'emploi de multiprocesseurs ajoute des problèmes spécifiques dus à l'apparition d'un parallélisme réel entre calcul et communication. Ces problèmes sont présentés et des solutions sont proposées pour l'environnement ATHAPASCAN-0. Ces solutions sont évaluées sur des grappes de multiprocesseurs. / The continuous price reduction for commodity PC multiprocessors and the availability of fast network interfaces have made cluster of multiprocessors an attractive low-price alternative to build parallel systems. Multiprocessor clusters offer two levels of parallelism: a fine grain parallelism inside a single multiprocessor and a coarse grain among them. A mechanism must be provided to exploit both levels of parallelism simultaneously. This requires to provide communications between threads belonging to different addresses spaces. This dissertation addresses the problem of integrating threads and communications on ATHAPASCAN-0 run time system. ATHAPASCAN-0 is a portable run time for cluster of multiprocessors developed as part of the APACHE project (CNRS-INPG-INRIA-UJF). Portability is achieved by a layered organization based on standards like POSIX threads and MPI. The ATHAPASCAN-0 run time system extends the heavy-weight process communication model of message passing libraries such as MPI, PVM, etc, into a lighter dynamic network of communicating threads. Multiprogramming is the key concept used. Communication progress is based on a network polling basis to handle incoming messages and to deliver outgoing communications requests. Performance is strongly dependent on the way these operations are implemented. Additionally, multiprocessors introduce some programming problems like overhead of cache coherency mechanisms, method of managing concurrent accesses and efficient mutex locking to avoid unnecessary context switching. These problems are analyzed and solutions are implemented in the ATHAPASCAN-0 run time system. An evaluation of these solutions is performed on a cluster of multiprocessors.