Les systèmes de calcul actuels sont généralement des grappes de machines composés de nombreux processeurs à l'architecture fortement hiérarchique. Leur exploitation constitue le défi majeur des implémentations de modèles de programmation tels MPI ou OpenMP. Une pratique courante consiste à mélanger ces deux modèles pour bénéficier des avantages de chacun. Cependant ces modèles n'ont pas été pensés pour fonctionner conjointement ce qui pose des problèmes de performances. Les travaux de cette thèse visent à assister le développeur dans la programmation d'application de type hybride. Il s'appuient sur une analyse de la hiérarchie architecturale du système de calcul pour dimensionner les ressources d'exécution (processus et threads). Plutôt qu'une approche hybride classique, créant un processus MPI multithreadé par noeud, nous évaluons de façon automatique des solutions alternatives, avec plusieurs processus multithreadés par noeud, mieux adaptées aux machines de calcul modernes.

De nos jours, des processeurs à mémoire partagée cohérente ayant jusqu’à 100 coresintégrés sur la même puce sont une réalité et des processeurs many-cores ayant plusieurs centaines, voire, un millier de cores sont à prévoir prochainement.Dans ces architectures, la question de la localité du trafic lié aux miss decaches L1 (data, instruction et TLB) est primordiale à la fois pour passer àl’échelle et pour réduire la consommation électrique (énergie consommée par bittransféré). Notre thèse est que : (i) la gestion de la localité des accès mémoiredoit être prise en compte au niveau du noyau du système d’exploitation et elle doitêtre effectuée d’une manière transparente aux applications utilisateur; et (ii) les noyaux monolithiques actuels sont incapables de renforcer la localité des accès mémoire des threads d’une même application parallèle, car la notion de threadsdans ces noyaux est intrinsèquement inadaptée pour les processeurs many-cores.Par conséquent, nous pensons que la démarche suivie jusqu’à présent pour faireévoluer les noyaux monolithiques n’est pas suffisante et qu’il est impératif demettre la question de la localité des accès mémoire au centre de cette évolution.Pour prouver notre thèse, nous avons conçu et réalisé ALMOS (Advanced Locality Management Operating System), un système d’exploitation expérimental à base de noyau monolithique distribué. ALMOS dispose d’un nouveau concept de thread, nommé Processus Hybrides. Il permet à son noyau de renforcer, d’une manière transparente, la localité des accès mémoire liés à l'exécution de chaque thread. La gestion des ressources (cores et mémoires physiques) dans le noyau d’ALMOS est distribuée renforçant la localité des accès mémoire lors de la réalisation des services systèmes. La prise de décision concernant l’allocation mémoire, le placement des tâches et l’équilibrage de charge dans le noyau d’ALMOS est décentralisée, multi-critères et sans prise de verrou. Elle repose sur une infrastructure distribuée coordonnant d’une manière scalable l’accès aux ressources.En utilisant le prototype virtuel précis au cycle et au bit près du processeur many-core TSAR, nous avons expérimentalement démontré que : (i) les performances(scalabilité et temps d’exécution) du schéma d'ordonnancement distribué du noyaud’ALMOS sur 256 cores dépassent celles des noyaux monolithiques existants; (ii) la réalisation répartie de l’appel système fork permet de passer à l’échellece service système sur 512 cores; (iii) le coût de la mise à jour de l’infrastructure distribué de prise de décisions du noyau d’ALMOS ne nécessiteque 0.05% de la puissance de calcul totale du processeur TSAR; (iv) les performances(scalabilité, temps d’exécution et trafic distant) de la stratégie d’affinitémémoire du noyau d’ALMOS, nommé Auto-Next-Touch, dépassent celles des deuxstratégies First-Touch et Interleave sur 64 cores; (v) le modèle de processushybrides d’ALMOS permet de passer à l’échelle deux applications hautementmulti-threads existantes sur 256 cores et une troisième sur 1024 cores; et enfin (vi) le couple ALMOS/TSAR (64 cores) offre systématiquement une bien meilleure scalabilité que le couple Linux/AMD (Interlagos 64 cores) pour 8 applications de classe HPC et traitement d’images. / Nowadays, single-chip cache-coherent many-core processors having up to 100 coresare a reality. Many-cores with hundreds or even a thousand of cores are planned in the near future. In theses architectures, the question of the locality of L1 cache-miss related traffic (data, instruction and TLB) is essential for both scalability and power consumption (energy by moved bit). Our thesis is that: (i) handling the locality of memory accesses should be done at kernel level of an operating system in a transparent manner to user applications; and (ii) the current monolithic kernels are not able to enforce the locality of memory accesses of multi-threaded applications, because the concept of thread in these kernels is inherently unsuitable for many-core processors. Therefore, we believe that the evolution approach of monolithic kernels undertaken until now is insufficient and it is imperative to put the question of the locality of memory accesses in the heart of this evolution.To prove our thesis, we designed and implemented ALMOS (Advanced Locality Management Operating System), an experimental operating system based on a distributed monolithic kernel. ALMOS has a new concept of thread, called Hybrid Process. It allows its kernel to enforce the locality of memory accesses of each thread. The resources (cores and physical memory) management in ALMOS's kernel is distributed enforcing the locality of memory accesses when performing system services. Decision making regarding memory allocation, tasks placement and load balancing in ALMOS's kernel is decentralized, multi-criteria and without locking. It is based on a distributed infrastructure coordinating, in a scalable manner, the accesses to resources.Using the cycle accurate and bit accurate virtual prototype of TSAR many-core processor, we experimentally demonstrated that: (i) performance (scalability and execution time) on 256 cores of the distributed scheduling scheme of ALMOS's kernel outperform those of the shared scheduling scheme found in existing monolithic kernels; (ii) distributed realization of the fork system call enables this system service to scale on 512 cores; (iii) updating the distrusted decision-making infrastructure of ALMOS's kernel costs just 0.05 % of the total computing power of TSAR processor; (iv) performance (scalability, execution time and remote traffic) of memory affinity strategy of ALMOS's kernel, called Auto-Next-Touch, outperform those of two existing strategies First-Touch and Interleave on 64 cores; (v) concept of Hybrid Process of ALMOS's kernel scales up two existing highly multi-threads applications on 256 cores and a third one on 1024 cores; and finally (vi) the couple ALMOS/TSAR (64 cores) gives systematically much better scalability than the couple Linux/AMD (Interlagos 64 cores) for 8 multi-threads applications belonging to HPC and image processing domains.

SkePU 3 is a framework for parallel program execution that uses higher order functions called skeletons, which provide a layer of abstraction between user code and the parallel implementation it provides through its backends. The backend that enables SkePU to run on an HPC cluster has a slowdown of a factor two. This reduces the viability of SkePU as an alternative for HPC, and as such, warrants an investigation. Programs written in SkePU are sequential-looking, single-source C++ programs where skeleton calls can transparently execute on multiple different types of processing units, such as CPU cores, GPUs and clusters, using different backends. In this thesis, a strategy for improving the performance of SkePU on clusters is presented, and with it, the design and implementation of a new cluster backend that is simpler and more closely integrated with the non-cluster SkePU code base. Runtime measurements are made, which show that the new cluster backend sees a relative speedup of about a factor of two, which effectively eliminates the slowdown.

Cette thèse s'intéresse à la performance des serveurs de données sur les architectures multi-cœurs. Nous avons choisi d'étudier ce problème sous deux aspects différents. Premièrement, nous étudions un support d'exécution événementiel. Nous montrons notamment que le mécanisme de vol de tâches, utilisé pour équilibrer la charge entre les cœurs, peut pénaliser la performance d'un serveur Web. Nous proposons donc diverses optimisations pour améliorer les performances de ce mécanisme sur les processeurs multi-cœurs. Deuxièmement, nous étudions la performance du serveur Web Apache, exploitant à la fois un ensemble de threads et de processus, sur une architecture multi-cœurs NUMA. Nous montrons notamment que, sous une charge réaliste, ce serveur Web ne passe pas idéalement à l'échelle. Grâce à une analyse détaillée des coûts, nous déterminons les raisons de ce manque de passage à l'échelle et présentons un ensemble de propositions visant à améliorer la performance de ce serveur sur une architecture NUMA.

Made available in DSpace on 2016-04-28T19:34:04Z (GMT). No. of bitstreams: 1 LETICIA MORALES WANDERLEY MARINHO.pdf: 529924 bytes, checksum: 824c8d76f71c943a8e72de48186b4d38 (MD5) Previous issue date: 2006-04-26 / This work has the intention of identifying the strategies used by the characters of a constructed dialog. The selected corpus contains dialogs between characters from the play of playwright Plínio Marcos, Dois perdidos numa noite suja (Both lost in a dirty night). Here we pointed in the dialogs the conversational schemas that the caracthers use to interact with their in the most diversified conversational situations. Dois perdidos numa noite suja (Both lost in a dirty night) is a play fiction, elaborated with devices from spoken language. The text show two characters life that have fiwest economic resorces, have suffers with the social differences and live a life in a hostile way. To highlight the feelings like rebilion, non-conformism and hate, that the characthers fiils the author employs devices that produce effects bringing the fictional text closer to reality. By the selects dialogs, we can to confirm how the literary dialogs can offer expressive examples of interaction, like it happens in a natural conversation / Este trabalho tem por objetivo identificar as estratégias conversacionais utilizadas pelas personagens do diálogo de ficção. O corpus selecionado contém diálogos retirados da peça de Plínio Marcos, Dois perdidos numa noite suja, onde destacamos os esquemas conversacionais que as personagens utilizam para interagirem nas mais variadas situações comunicativas. Dois perdidos numa noite suja é uma obra ficcional em que os diálogos foram construídos com recursos da língua oral. O texto retrata a vida de duas personagens com poucos recursos financeiros, que sofrem com as diferenças sociais e vivem em um ambiente hostil. Para realçar o sentimento de revolta, inconformismo e o ódio que as personagens sentem, o autor utiliza recursos lingüísticos que aproxima o texto ficcional da realidade, dando maior ênfase as emoções. Assim, por meio dos diálogos selecionados, podemos afirmar que os diálogos de ficção podem fornecer-nos exemplos expressivos de interação, aproximando-se de uma conversação natural

Submitted by Angela Maria de Oliveira (amolivei@uepg.br) on 2017-12-01T13:39:21Z No. of bitstreams: 2 license_rdf: 811 bytes, checksum: e39d27027a6cc9cb039ad269a5db8e34 (MD5) Jose Wander de Paula.pdf: 846750 bytes, checksum: 4b196f858fbc02cabdfbc08b6cee6553 (MD5) / Made available in DSpace on 2017-12-01T13:39:21Z (GMT). No. of bitstreams: 2 license_rdf: 811 bytes, checksum: e39d27027a6cc9cb039ad269a5db8e34 (MD5) Jose Wander de Paula.pdf: 846750 bytes, checksum: 4b196f858fbc02cabdfbc08b6cee6553 (MD5) Previous issue date: 2017-09-25 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / Esta dissertação trata de uma leitura do livro Se um viajante numa noite de inverno de Italo Calvino (1979) e o objetivo é analisar três categorias trazidas pelo livro como personagens: o autor, o editor e o leitor e a “voz narrativa” que fala no romance. Para o escritor “a relação de quem narra com a matéria narrada e com o leitor” é a “problemática mais árdua da narrativa contemporânea” (CALVINO, 2009, p. 199). Como fundamentação teórica para esta temática utilizamos alguns textos fomentadores destas discussões: Roger Chartier (1998) e seu estudo sobre autor e editor, Compagnon sobre estilo, autor e leitor (2001), Michel Foucault (2009) sobre o que é um autor, a “morte do autor” prenunciada por Roland Barthes (2004), Umberto Eco (1994) e a questão do leitor, entre outros. A respeito da categoria “autor”, entendemos ser este uma figura que Umberto Eco renomeou em “autor-modelo”, “autor-empírico”, entendemos que a “voz narrativa” criada por Italo Calvino no romance e as estratégias narrativas utilizadas pelo escritor ecoam este viés. Destas duas categorias, “autor” e “voz narrativa”, relacionamos à teoria que fala a respeito de “estilo”, reconhecendo as estratégias de Italo Calvino como um “tom” que dificilmente poderíamos desvincular da figura histórica do autor ao compararmos com seu experimentalismo em outros livros como Cidades Invisíveis (CALVINO, 2003). Este estudo nos possibilitou compreender, também, que o “Editor” faz parte da autoria no sentido de que a Casa Editorial é o local em que são aceitos para publicação textos que passam pelo crivo de estéticas estabelecidas por um Editor que circunscreve, delimita e corrige o que será publicado. Por último, estudar a morte do autor anuncia que o Leitor deve ser enfatizado em sua atividade de leitura. A conclusão a que chegamos é que em Se um viajante numa noite de inverno Italo Calvino criou um romance armadilha, tanto do leitor-médio, como dos demais leitores. / This paper is about our reading of If on a traveler on a winter’s night, book was wrote by Italo Calvino (1979), and the aim is to analyze three characters from the book: the editor, the author and the reader, as well as we study the book “narrative voice”. To the writer “a relação de quem narra com a matéria narrada e com o leitor” is the “problemática mais árdua da narrativa contemporânea” (CALVINO, 2009, p. 199). This discussions in this text is based on Roger Chartier about author and editor (1998), Antoine Compagnon (2001) about author, reader and style, the “author’s dead” by Roland Barthes (2004), Michel Foucault and what is an author? (2009), Umberto Eco (1994) and the reader, among others. We understand the author is a figure divided by Umberto Eco in autor-modelo and autor-empírico and the narrative strategies used by the writer talks with these categories – and these categories are very hard to separate on historical author figure. The Editor is part of the authorship in the sense that the Editorial House is the place where texts that pass through the aesthetic list established by an Editor are accepted for publication. Studying the death of the author announces that the reader should be emphasized in his reading activity. The conclusion we reach is that in If a traveler on a winter's night Italo Calvino created a trappings romance, be it the average reader or the other readers.

Tese de Doutoramento. Líderes para Indústrias Tecnológicas. Faculdade de Engenharia. Universidade do Porto. 2012

Les architectures des machines de calcul sont de plus en plus complexes et hiérarchiques, avec des processeurs multicœurs, des bancs mémoire distribués, et de multiples bus d'entrées-sorties. Dans le cadre du calcul haute performance, l'efficacité de l'exécution des applications parallèles dépend du coût de communication entre les tâches participantes qui est impacté par l'organisation des ressources, en particulier par les effets NUMA ou de cache. Les travaux de cette thèse visent à l'étude et à l'optimisation des communications haute performance sur les architectures hiérarchiques modernes. Ils consistent tout d'abord en l'évaluation de l'impact de la topologie matérielle sur les performances des mouvements de données, internes aux calculateurs ou au travers de réseaux rapides, et pour différentes stratégies de transfert, types de matériel et plateformes. Dans une optique d'amélioration et de portabilité des performances, nous proposons ensuite de prendre en compte les affinités entre les communications et le matériel au sein des bibliothèques de communication. Ces recherches s'articulent autour de l'adaptation du placement des tâches en fonction des schémas de transfert et de la topologie des calculateurs, ou au contraire autour de l'adaptation des stratégies de mouvement de données à une répartition définie des tâches. Ce travail, intégré aux principales bibliothèques MPI, permet de réduire de façon significative le coût des communications et d'améliorer ainsi les performances applicatives. Les résultats obtenus témoignent de la nécessité de prendre en compte les caractéristiques matérielles des machines modernes pour en exploiter la quintessence.

Submitted by Automa??o e Estat?stica (sst@bczm.ufrn.br) on 2017-11-22T20:56:04Z No. of bitstreams: 1 MariaDoPerpetuoSocorroRochaSousaSeverino_TESE.pdf: 5405679 bytes, checksum: 33fdaee514b9e930abbff55ced8d0616 (MD5) / Approved for entry into archive by Arlan Eloi Leite Silva (eloihistoriador@yahoo.com.br) on 2017-11-24T19:10:38Z (GMT) No. of bitstreams: 1 MariaDoPerpetuoSocorroRochaSousaSeverino_TESE.pdf: 5405679 bytes, checksum: 33fdaee514b9e930abbff55ced8d0616 (MD5) / Made available in DSpace on 2017-11-24T19:10:38Z (GMT). No. of bitstreams: 1 MariaDoPerpetuoSocorroRochaSousaSeverino_TESE.pdf: 5405679 bytes, checksum: 33fdaee514b9e930abbff55ced8d0616 (MD5) Previous issue date: 2017-05-22 / Esta Tese inscreve-se na tem?tica da avalia??o de pol?ticas p?blicas. Demarcou como objeto de estudo o Programa Incluir: acessibilidade na educa??o superior, orientando-se pela seguinte pergunta de pesquisa: Como o Programa Incluir est? sendo implementado na Universidade Federal do Rio Grande do Norte (UFRN)? Estabeleceu como objetivo (geral) avaliar o processo de implementa??o do Programa Incluir na UFRN, e como objetivos espec?ficos: identificar os fatores que facilitam e os fatores que obstacularizam a implementa??o do Programa Incluir e que condicionam o alcance dos objetivos do mesmo na UFRN; conhecer os servi?os, recursos e apoios t?cnicos existentes e disponibilizados pela Programa Incluir na perspectiva de viabilizar a inclus?o educacional dos acad?micos com defici?ncia nessa IFES; detectar as barreiras (arquitet?nica, pedag?gica, de comunica??o e informa??o e atitudinal) que dificultam o processo de implementa??o do Programa Incluir na UFRN; identificar as estrat?gias de implementa??o do Programa Incluir nessa IFES. A hip?tese enunciada para nortear o estudo afirmou que O Programa Incluir implementa recursos, servi?os, apoios t?cnicos que minimizam as barreiras arquitet?nicas, pedag?gicas e de comunica??o e informa??o, viabilizando parcialmente a inclus?o educacional de acad?micos com defici?ncia na UFRN. Os procedimentos metodol?gicos utilizados para o alcance do objetivo proposto consistiram em revis?o da literatura, pesquisa documental, observa??o sistem?tica e pesquisa de campo. A pesquisa de natureza qualitativa utilizou para a coleta de dados entrevista semiestruturada aplicada aos sujeitos constituintes do quadro amostral, quais sejam: discentes com defici?ncia f?sica, auditiva, visual, intelectual e m?ltipla, professores, presidente e t?cnicos da Comiss?o Permanente de Apoio a Estudantes com Necessidades Especiais (CAENE), bolsistas e monitor. A tese conclui que o processo de implementa??o do Programa Incluir na UFRN efetiva apoios t?cnicos, servi?o de int?rprete de LIBRAS, recursos pedag?gicos em formato acess?vel, recursos financeiros sistem?ticos, os quais, associados a um conjunto de estrat?gias institucionais, como cria??o de normatiza??es relativas ? inclus?o educacional, instala??o de laborat?rios de acessibilidade, disponibiliza??o de tecnologias assistivas, materializa??o de parcerias intrainstitucionais, apoio da Reitoria, constituem fatores facilitadores do processo de implementa??o do programa avaliado. Por outro lado, constatamos desproporcionalidade na produ??o, adapta??o e disponibilidade de recursos did?ticos pedag?gicos em formato acess?vel em rela??o aos tipos de defici?ncia, certo desconhecimento/invisibilidade do Programa Incluir, multiplicidade de objetivos, p?blico-alvo distinto do preconizado no documento oficial, subestima??o da participa??o dos professores na condi??o de implementadores (in)diretos do Programa, o que, associado ao formalismo do parecer t?cnico emitido pela CAENE, sem a devida media??o entre t?cnicos da CAENE, professores e coordenadores de cursos, evidencia o que cunhamos como barreira burocr?tica. Essa coexiste com as remanescentes barreiras atitudinal, de infraestrutura, pedag?gica e de comunica??o e informa??o. Esses fatores inscrevem-se como obst?culos ? implementa??o do Programa Incluir na UFRN. A despeito dos mesmos, podemos inferir que o Programa Incluir contribui com o percurso de forma??o profissional e a participa??o parcial dos acad?micos com defici?ncia na UFRN. / This thesis is a part of the public policies evaluation. It is demarcated as study object the Include Program: accesibility in higher education, being guided by the following research question: How the Include Program is being implemented at the Federal University of Rio Grande do Norte (UFRN)? It established as (general) objective to evaluate the process of implementation of the Include Program at UFRN, and as specific objectives: to identify the factors that facilitate and the factors that obstruct the implementation of the Include Program, and that condition the achievement of its objectives at UFRN; to know the services, resources and technical support existing and made available by the Include Program in the perspective of enabling the educational inclusion of disabled students in this IFES; to detect barriers (architectural, pedagogical, communication and information and attitudinal) that hinder the implementation process of the Include Program at UFRN; to identify the implementation strategies of the Include Program in this IFES. The hypothesis enunciated to guide the study stated that the Include Program implements resources, services, technical support that minimize architectural, pedagogical and communication and information barriers, partially enabling the educational inclusion of disabled students in UFRN. The methodological procedures used to reach the proposed objective consisted of literature review, documentary research, systematic observation and field research. The qualitative research used for the data collection semistructured interview applied to the constituent subjects of the sample frame, namely: students with physical, auditory, visual, intellectual and multiple disabilities; teachers; president and technicians of the Permanent Commission to Support Students with Special Needs (CAENE); and scholarship holders and monitor. The thesis concludes that the implementation process of the Include Program at UFRN provides technical support, a LIBRAS interpreter service, pedagogical resources in an accessible format, systematic financial resources, which, together with a set of institutional strategies, such as creation of norms related to educational inclusion, Installation of accessibility laboratories, provision of assistive technologies, materialization of intra-institutional partnerships, support of the Rectory, are factors that facilitate the implementation process of the Include Program. Besides, we find disproportionality in the production, adaptation and availability of pedagogical didactic resources in relation to the types of disability, a lack of awareness/invisibility in the Include Program, multiplicity of objectives, different target audience from the official document, underestimation of the participation of teachers in the condition of (in)direct implementers of the Program, witch, together with the technical opinion formalism issued by CAENE, without proper mediation between CAENE technicians, teachers and course coordinators, shows what we have classified as a bureaucratic barrier. This coexists with the remaining attitudinal, infrastructure, pedagogical and communication and information barriers. These factors are inscribed as obstacles to the implementation of the Include Program at UFRN. Notwithstanding to this, we can infer that the Include Program contributes to the vocational training course and the partial participation of disabled academics in UFRN.

Submitted by ANA KARLA PEREIRA RODRIGUES (anakarla_@hotmail.com) on 2017-08-23T13:08:07Z No. of bitstreams: 1 arquivototal.pdf: 1209071 bytes, checksum: 1eddcf2f494891c2466f5052f15d1ced (MD5) / Made available in DSpace on 2017-08-23T13:08:07Z (GMT). No. of bitstreams: 1 arquivototal.pdf: 1209071 bytes, checksum: 1eddcf2f494891c2466f5052f15d1ced (MD5) Previous issue date: 2016-09-14 / Conselho Nacional de Pesquisa e Desenvolvimento Científico e Tecnológico - CNPq / In 1943 Beniamino Segrebelievedtohaveshownthatthemaximumnumberof lines containedinasmoothquarticsurfacein P3 is 64, ([16]).Butrecently,therewasa majoroverturnonthatthemewhenthemathematiciansRamsandSchuttfoundthat Segre hadmadeamistakeinhisworktoforgetthequartic'sfamily Z , ([14]),which essentiallycorrespondstothosequarticscontainingalinesthatcanbeincidenttomore than 18 lines containedinthesurface.Inthiswork,basedon([14]),weshowthatevery smoothquarticsurface,whichdoesnotbelongtofamily Z containsamaximumof 64 lines. Oneofthemostimportanttoolstoshowthisresult,isthestudyof_brations _l induced byaline l containedonthesurface,andtherelationshipbetweentheEuler characteristicofthebase(P1 in ourcase),the_bersandthesurfaceconcerned. / Em 1943,BeniaminoSegreacreditouterdemonstradoqueonúmeromáximo de retascontidasnumasuperfíciequárticanãosingularem P3 é 64; ([16]). Mas recentemente,houveumareviravoltanessetema,quandoosmatemáticosSªawomir Rams eMatthiasSchüttconstataramqueSegretinhacometidoumerroemseutrabalho ao esquecerasquárticasdafamília Z; ([14]), quecorrespondemessencialmenteas quárticas quepossuemretasquepodemserincidentesamaisde 18 retas contidas na superfície.Nestetrabalho,tendocomobase[14],mostramosquetodaquártica não singular,quenãopertenceafamília Z; contémnomáximo 64 retas. Umadas ferramentasmaisimportantes,paramostraresseresultado,éoestudodas_brações _l induzida porumareta l contidanasuperfície,earelaçãoqueexisteentrea característica deEulerdabase(emnossocaso P1), das_brassingulareseadasuperfície em questão.