Cette recherche s'inscrit dans le domaine de la didactique des langues étrangères et porte sur l'apprentissage du français oral soutenu par la technologie mobile. En nous appuyant sur la recherche-action, nous avons créé un dispositif de formation destiné à des apprenants coréens de FLE leur permettant de s'entraîner à prendre la parole librement. La culture coréenne étant influencée par la philosophie confucéenne, l'enseignement-apprentissage des langues est caractérisé par une relation hiérarchique entre le « maître savant » et « son disciple » et souvent centré sur l'écrit. Notre scénario didactique prévoyait un travail de la langue en dehors d'un cours formel et soutenu par un tuteur à distance. Deux expériences ont été menées en 2008 et 2010 avec deux groupes d'apprenants coréens de FLE résidants en Corée auxquels nous avons demandé de réaliser des tâches langagières par enregistrement vidéo. Les apprenants ont été amenés à réaliser leurs productions orales en réponse à une vidéo enregistrée par un locuteur natif vivant en France. Grâce au recours à la téléphonie mobile intégrant l'accès à Internet et l'utilisation d'un réseau social en ligne coréen (Cyworld), les apprenants pouvaient échanger avec l'interlocuteur en temps différé leur permettant de s'enregistrer à plusieurs reprises et ainsi s'entraîner à la prise de parole tout en étant accompagnés par un tuteur. L'analyse des discours et des stratégies de réalisation des vidéos ainsi que les résultats des entretiens semi-directifs menés en amont et en aval avec les apprenants-participants, nous ont permis de répondre à nos questions de recherche, à savoir de quelle manière le tutorat en ligne, les tâches communicatives réalisées avec un outil mobile et la communication asynchrone sur un réseau social en ligne influent sur la production orale et l'autonomisation des apprenants coréens de FLE. / This research is in the field of foreign language didactics. It is particularly about learning spoken French via mobile technology. We carried out a research action while we elaborated a training device which would allow Korean learners of French to practice oral speech. As Korean culture is influenced by Confucian philosophy, teaching-learning foreign languages is characterized by a hierarchical relationship between the "master" and his "disciple" and is often focused on writing. Our didactic scenario included language learning tasks outside of formal learning and supported by distance tutoring. A couple of experiments were conducted in 2008 and 2010 with two Korean sets of French language learners living in Korea who were asked to perform language tasks using video recording. The learners were asked to perform oral productions in response to videos recorded by a native French speaker living in France. Through a mobile phone connected to the Internet and an Korean online social network (Cyworld), learners could exchange with the French native asynchronically. This allowed them to record themselves several times and practice speaking skills by the way, benefiting from advice given by a tutor. We analyzed discourse, video-implemented strategies and semi-directed interviews conducted upstream and downstream with the learners. This brought us to find out to which extent online tutoring, communicative tasks performed with a mobile device, and asynchronous communication via an online social network may influence oral production and autonomy of Korean learners of French.

Submitted by Alison Vanceto (alison-vanceto@hotmail.com) on 2017-06-05T11:55:43Z No. of bitstreams: 1 TeseDEE.pdf: 16153724 bytes, checksum: 96c77e1d454c6b99b64dc1effdee0484 (MD5) / Approved for entry into archive by Ronildo Prado (ronisp@ufscar.br) on 2017-06-05T17:14:25Z (GMT) No. of bitstreams: 1 TeseDEE.pdf: 16153724 bytes, checksum: 96c77e1d454c6b99b64dc1effdee0484 (MD5) / Approved for entry into archive by Ronildo Prado (ronisp@ufscar.br) on 2017-06-05T17:14:32Z (GMT) No. of bitstreams: 1 TeseDEE.pdf: 16153724 bytes, checksum: 96c77e1d454c6b99b64dc1effdee0484 (MD5) / Made available in DSpace on 2017-06-05T17:19:25Z (GMT). No. of bitstreams: 1 TeseDEE.pdf: 16153724 bytes, checksum: 96c77e1d454c6b99b64dc1effdee0484 (MD5) Previous issue date: 2017-02-15 / Não recebi financiamento / The aim of this study is to propose an online English course using a thematic task-based syllabus, on the Moodle plataform. It intends to identify the characteristics of the tasks elaborated and analyze the students’ performances. From the methodological point of view, this is a qualitative-interpretative research, with a virtual ethnographic basis and it was carried out on the Moodle platform, which is offered by a Federal Technology Institution of the interior of São Paulo. The subjects were students of technical and undergraduate courses. The data were collected through semi-structured interview, questionnaire and record of conversations on the platform, Facebook and Skype. Besides these, the tasks elaborated for this study and those that were done by the students were also used. The 40-hour-online English course had a macro-theme called Inventors and Inventions and consisted of three thematic units Top 10 inventors of all time, Nikola Tesla and Leonardo da Vinci. One of the reasons for using task-based thematic syllabus comes from results of researches (BARBIRATO, 1999, 2005; XAVIER, 1999; CASSOLI, 2011; FRANSCESCHINI, 2014) that show the use of tasks can give great opportunities for learners to use the target language. When elaborating the tasks on the Moodle platform, the teacher came across with two main problems: her approach, still with structural influence, and the limitation of some platform tools. The students also encountered some technical and linguistic difficulties in carrying out the tasks. Despite these problems, the empirical results of this research may propose alternatives to foreign language professionals interested in using thematic task-based syllabus in a virtual learning environment. / Esta pesquisa tem o intuito de propor um curso de inglês a distância por meio de um planejamento temático baseado em tarefas (PTBT), na plataforma Moodle. Pretende-se também identificar as características das tarefas elaboradas pela professora e analisar a realização delas pelos alunos. Do ponto de vista metodológico, trata-se de uma pesquisa qualitativa-interpretativista, de base etnográfica virtual e foi realizada na plataforma Moodle, disponibilizada por uma Instituição Federal de Tecnologia do interior de São Paulo. Os participantes eram alunos de cursos técnicos e superiores da área de exatas. Como instrumentos de coleta de dados foram utilizados entrevista semiestruturada, questionário e registro das conversas na plataforma, no Facebook e no Skype. Além desses, utilizou-se, para a análise, as tarefas elaboradas pela professora e as que foram realizadas pelos alunos. O curso de inglês foi oferecido a distância e tinha como tema macro Inventores e Invenções e era constituído de três unidades temáticas Os 10 melhores inventores de todos os tempos, Nikola Tesla e Leonardo da Vinci, distribuídas em 40 horas. Uma das justificativas para o uso do PTBT provém do fato de que os resultados de algumas pesquisas (BARBIRATO, 1999, 2005; XAVIER, 1999; CASSOLI, 2011; FRANSCESCHINI, 2014) apontam que o uso de tarefas proporciona maiores oportunidades de uso da língua-alvo pelos aprendizes. Ao elaborar as tarefas na plataforma Moodle, a professora deparou-se com duas problemáticas principais: a sua abordagem, ainda com resquícios da estrutural, e a limitação de algumas ferramentas da plataforma. Os alunos também se depararam com dificuldades técnicas, cognitivas e linguísticas, ao realizarem as tarefas. Apesar dessas dificuldades, os resultados empíricos desta pesquisa podem propor encaminhamentos para os profissionais de línguas estrangeiras, interessados em utilizar o PTBT em curso a distância.

Submitted by Aline Batista (alinehb.ufpel@gmail.com) on 2015-04-02T19:23:46Z No. of bitstreams: 2 Aprendizagem da língua inglesa como segunda língua baseada em tarefas.pdf: 10361270 bytes, checksum: 50839ecc3d133d5fe19ddf26c6cd568f (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) / Approved for entry into archive by Aline Batista (alinehb.ufpel@gmail.com) on 2015-04-03T01:21:36Z (GMT) No. of bitstreams: 2 Aprendizagem da língua inglesa como segunda língua baseada em tarefas.pdf: 10361270 bytes, checksum: 50839ecc3d133d5fe19ddf26c6cd568f (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) / Made available in DSpace on 2015-04-03T01:23:20Z (GMT). No. of bitstreams: 2 Aprendizagem da língua inglesa como segunda língua baseada em tarefas.pdf: 10361270 bytes, checksum: 50839ecc3d133d5fe19ddf26c6cd568f (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) Previous issue date: 2014-01-31 / Sem bolsa / Em 1997, a autora Diane Larsen-Freeman publicou um inquietante artigo chamado “Ciência do Caos / Complexidade e Aquisição de Segunda Língua” (LARSENFREEMAN, 1997). Nesse trabalho, a linguista argumenta que há muitas similaridades entre a nova ciência do caos/complexidade e a aquisição de segunda língua. Diante das reflexões que tais ideias suscitam, pretende-se analisar a estrutura e a teoria da aprendizagem baseada em tarefas (Task Based Learning), sob a ótica da teoria do caos/complexidade. De acordo com Willis, J. (1996), essa abordagem comunicativa apresenta seis etapas, a saber: pré-tarefa, tarefa, planejamento, relatório, análise e prática. Como recorte em um universo tão amplo de pesquisa, tem-se a relação da teoria do caos/complexidade com o ciclo da tarefa – tarefa, planejamento e relatório (task-planning-report), enunciado por Willis, J. (1996).De acordo com Larsen-Freeman e Cameron (2008), uma abordagem complexa não se traduz em um método de ensino de L2 complexo. Este trabalho visa aproximar o trabalho seminal da linguista Diane Larsen-Freeman sobre a Teoria do Caos/Complexidade e Aquisição de Segunda Língua (1997) com a Aprendizagem Baseada em Tarefas propostanove anos antes por Willis, J. e Willis, D. (1988). Esta dissertação de mestrado concluiu que o ciclo da tarefa pode ser caracterizado como um sistema adaptativo complexo, uma vez que satisfaz suas regras de baixo nível, já que se trata de um sistema que é dinâmico, complexo, não linear, caótico, imprevisível, sensível às condições iniciais, aberto, auto-organizável, sensível ao feedback e adaptável. / In 1997, the linguist Diane Larsen-Freeman published a disturbing article called “Chaos/Complexity Science and Second Language Acquisition”(LARSENFREEMAN, 1997). In this article, the author states that there are many similarities between the new science of chaos/complexity and second language acquisition. Faced with the reflections, which give rise to such ideas, one intends to analyse the structure and theory of Task Based Learning under the light of the Chaos/Complexity Theory. According to Willis, J. (1996), this communicative approach presents itself with six stages, namely: pre-task, task, planning, report, analyses and practice. As part of a so vast research area, one intends to relate the Chaos/Complexity Theory with the task cycle – task, planning and report – formulated by Willis, J. (1996). According to Larsen-Freeman and Cameron (2008), a complexity approach does not translate into a complexity method for teaching language. This work aims to approximate the seminal article by Diane Larsen-Freeman about the Chaos/Complexity Theory and Second Language Acquisition (1997) with Task Based Learning proposed nine years before by Willis, J. and Willis, D. (1988).To sum up, this dissertation aims to show that the task cycle may be characterized as a complex adaptive system, once it complies with the rules of such systems, that is, dynamic, complex, nonlinear, chaotic, unpredictable, sensitive to the initial conditions, open, self-organizing, feedback sensitive and adaptive.

Curriculum development involves considerations about curriculum relevance. Thus it is necessary that the curriculum should identify with the needs of the graduate and with professional practice. This requirement shoNAt-hsa t curriculum relevance is not fixed, a view that is consistent with a dynamic, situational approach to curriculum development. The basic categories which define librarianship curriculum development are library activities, theory, innovation, teaching and learning, employers, students. lecturers. These are influential factors in curriculum relevance. Variables in curriculum content such as theory and practice affect curriculum relevance. Librarianship curriculum development should aim at relevance by integrating academic study and practice. Hence, the goal of study towards librarianship education must be focused on the activities performed in library and information services. The need for innovation in library and information services means that novel viewpoints and solutions must be practical. For example, this requirement indicates that curriculum development must take into consideration leamt attributes which are aerieral and transferable in a changing world. This is in view of the employers' requirement that graduates should have critical intellectual ability and the capability to learn rather than their just immediate attributes, skills and knowledge. With teaching and leaming there is abundant rationale for the development of more effective delivery systems than traditional lecturing. If outcome-based learning is valued, individualised, self-directed learning is a prerequisite. The practices of the task-based curriculum, with its focus on student learning and on the development of transferable skills more closely approximate the ideal approaches to librarianship education. The teaching of transferable skills is more likely to define the conditions under which critical reasoning can develop. It has an advantage over the students' abilities to learn to function in the profession outside the university and for continuous development. In this respect task-based education has much wider implications than that of simply providing students with skills. Professional practice does not always fit with the curriculum that is developed by the experts. The expert-developed curriculum also poses a problem for those who interpret it, learn it and receive the products. Thus, a strong joint partnership in which the library and the library school are both recognised in curriculum development is essential if the profession is to fulfil effectively its unique role in society.

La résolution de grands systèmes linéaires creux est un élément essentiel des simulations numériques.Ces résolutions peuvent représenter jusqu’à 80% du temps de calcul des simulations.Une parallélisation efficace des noyaux d’algèbre linéaire creuse conduira donc à obtenir de meilleures performances. En mémoire distribuée, la parallélisation de ces noyaux se fait le plus souvent en modifiant leschéma numérique. Par contre, en mémoire partagée, un parallélisme plus efficace peut être utilisé. Il est doncimportant d’utiliser deux niveaux de parallélisme, un premier niveau entre les noeuds d’une grappe de serveuret un deuxième niveau à l’intérieur du noeud. Lors de l’utilisation de méthodes itératives en mémoire partagée,les graphes de tâches permettent de décrire naturellement le parallélisme en prenant comme granularité letravail sur une ligne de la matrice. Malheureusement, cette granularité est trop fine et ne permet pas d’obtenirde bonnes performances à cause du surcoût de l’ordonnanceur de tâches.Dans cette thèse, nous étudions le problème de la granularité pour la parallélisation par graphe detâches. Nous proposons d’augmenter la granularité des tâches de calcul en créant des agrégats de tâchesqui deviendront eux-mêmes des tâches. L’ensemble de ces agrégats et des nouvelles dépendances entre lesagrégats forme un graphe de granularité plus grossière. Ce graphe est ensuite utilisé par un ordonnanceur detâches pour obtenir de meilleurs résultats. Nous utilisons comme exemple la factorisation LU incomplète d’unematrice creuse et nous montrons les améliorations apportées par cette méthode. Puis, dans un second temps,nous nous concentrons sur les machines à architecture NUMA. Dans le cas de l’utilisation d’algorithmeslimités par la bande passante mémoire, il est intéressant de réduire les effets NUMA liés à cette architectureen plaçant soi-même les données. Nous montrons comment prendre en compte ces effets dans un intergiciel àbase de tâches pour ainsi améliorer les performances d’un programme parallèle. / Solving large sparse linear system is an essential part of numerical simulations. These resolve can takeup to 80% of the total of the simulation time.An efficient parallelization of sparse linear kernels leads to better performances. In distributed memory,parallelization of these kernels is often done by changing the numerical scheme. Contrariwise, in sharedmemory, a more efficient parallelism can be used. It’s necessary to use two levels of parallelism, a first onebetween nodes of a cluster and a second inside a node.When using iterative methods in shared memory, task-based programming enables the possibility tonaturally describe the parallelism by using as granularity one line of the matrix for one task. Unfortunately,this granularity is too fine and doesn’t allow to obtain good performance.In this thesis, we study the granularity problem of the task-based parallelization. We offer to increasegrain size of computational tasks by creating aggregates of tasks which will become tasks themself. Thenew coarser task graph is composed by the set of these aggregates and the new dependencies betweenaggregates. Then a task scheduler schedules this new graph to obtain better performance. We use as examplethe Incomplete LU factorization of a sparse matrix and we show some improvements made by this method.Then, we focus on NUMA architecture computer. When we use a memory bandwidth limited algorithm onthis architecture, it is interesting to reduce NUMA effects. We show how to take into account these effects ina task-based runtime in order to improve performance of a parallel program.

La complexification des architectures matérielles pousse vers l’utilisation de paradigmes de programmation de haut niveau pour concevoir des applications scientifiques efficaces, portables et qui passent à l’échelle. Parmi ces paradigmes, la programmation par tâches permet d’abstraire la complexité des machines en représentant les applications comme des graphes de tâches orientés acycliques (DAG). En particulier, le modèle de programmation par tâches soumises séquentiellement (STF) permet de découpler la phase de soumission des tâches, séquentielle, de la phase d’exécution parallèle des tâches. Même si ce modèle permet des optimisations supplémentaires sur le graphe de tâches au moment de la soumission, il y a une préoccupation majeure sur la limite que la soumission séquentielle des tâches peut imposer aux performances de l’application lors du passage à l’échelle. Cette thèse se concentre sur l’étude du passage à l’échelle du support d’exécution StarPU (développé à Inria Bordeaux dans l’équipe STORM), qui implémente le modèle STF, dans le but d’optimiser les performances d’un solveur d’algèbre linéaire dense utilisé par le CEA pour faire de grandes simulations 3D. Nous avons collaboré avec l’équipe HiePACS d’Inria Bordeaux sur le logiciel Chameleon, qui est une collection de solveurs d’algèbre linéaire portés sur supports d’exécution à base de tâches, afin de produire un solveur d’algèbre linéaire dense sur StarPU efficace et qui passe à l’échelle jusqu’à 3 000 coeurs de calcul et 288 accélérateurs de type GPU du supercalculateur TERA-100 du CEA-DAM. / The ever-increasing supercomputer architectural complexity emphasizes the need for high-level parallel programming paradigms to design efficient, scalable and portable scientific applications. Among such paradigms, the task-based programming model abstracts away much of the architecture complexity by representing an application as a Directed Acyclic Graph (DAG) of tasks. Among them, the Sequential-Task-Flow (STF) model decouples the task submission step, sequential, from the parallel task execution step. While this model allows for further optimizations on the DAG of tasks at submission time, there is a key concern about the performance hindrance of sequential task submission when scaling. This thesis’ work focuses on studying the scalability of the STF-based StarPU runtime system (developed at Inria Bordeaux in the STORM team) for large scale 3D simulations of the CEA which uses dense linear algebra solvers. To that end, we collaborated with the HiePACS team of Inria Bordeaux on the Chameleon software, which is a collection of linear algebra solvers on top of task-based runtime systems, to produce an efficient and scalable dense linear algebra solver on top of StarPU up to 3,000 cores and 288 GPUs of CEA-DAM’s TERA-100 cluster.

Cette thèse se propose de démontrer combien un apprentissage hybride en «Ticéitude» (notion que nous avons inventée) peut être source de progrès, de motivation et d’autonomie et aussi de développer l’autotélisme des professeurs et des apprenants, c’est-à-dire une notion de bien être quand on apprend ou quand on enseigne.L’éducation est en train de prendre un tournent de son histoire grâce au numérique. On voit apparaître de nouvelles approches innovantes un peu partout dans le monde. Le défi est de pouvoir améliorer la qualité des apprentissages. Dans l’enseignement des langues et du FLE en particulier, le défi est de taille car il s’agit à la fois de suivre les curriculum en place et d’inventer de nouvelles approches sans choquer les uns ou effrayer les autres. Le modèle que nous avons mis en place et que nous appelons «Ticéitude» s’inscrit dans le nouveau paradigme de l’apprenance et d’un enseignement des langues humaniste. La «Ticéitude» permet de tisser des liens dans le cadre d’un environnement pédagogique riche et diversifié. Il s’agit d’un modèle socio-constructiviste humaniste qui favorise l’apprentissage collectif ou collaboratif; ainsi renforçant la motivation et l’autonomie des apprenants dans un contexte australien, suivant ainsi les différents curriculum de ce pays. L’approche hybride s’est vite imposée à nous comme un modèle à la fois contemporain mais aussi futuriste car elle permet un apprentissage renforcé, voire accéléré. Nous visons dans une société cognitive du réseautage où on peut apprendre ce qu’on veut, quand on le veut individuellement ou avec les autres. La notion de professeur est en train de changer car nous devenons de plus en plus des «coach» ou des «facilitateurs pédagogiques». Le dispositif que nous avons mis en place permet l’apprentissage quand on a envie de le faire, mais avec le guidage du professeur. Le professeur devient alors un allié! Il guide, il propose, mais aussi il participe avec les apprenants sur le réseau social mis en place au sein du dispositif que nous avons mis en place. Finalement, la «Ticéitude» implique l’utilisation des TICE mais de manière humaniste afin que les outils numériques puissent prendre toute leur valeur dans un environnement durable propice à l’apprentissage du FLE en Australie mais aussi dans le monde entier. / This thesis aims to demonstrate the effective implementation of a hybrid approach for learning French as a second language with the application of “Ticéitude”, a practical new concept for language learning in the 21st Century. This concept is a process of self-progression that targets the development of autonomy in the learner and for self-motivation, while also developing an autotelic personality, which means a certain “well-being” in the process of learning and teaching. Education is evolving due to today’s fast progress in digital technologies and is changing the course of history. Also because of this fast paced progress, we are currently witnessing new innovative approaches almost everywhere around the world. These challenges imply the importance of improving the quality of learning materials that is made available to the general populace. Our model, which is called Ticéitude, is aimed towards a humanistic approach of learning in the paradigm of “apprenance”. The notion of Ticéitude is to create links in a rich and diversified network. It is based on a socio-constructivist model, which encourages collaborative learning; thus reinforcing the motivation and autonomy of learners. In this instance, the model is based on the Australian context following the different curriculum in the various states and territory of this country. A hybrid approach has been put in place to create accelerated learning, based on the process of self-progression that targets the development of autonomy in the learner and for self-motivation, while also encouraging the development of an autotelic personality. In today’s world, we live in a cognitive society of networking when one can learn whenever or wherever he/she wants. The notion of “teacher” is changing as we focus more about coaching than teaching. The platform, which we created, allows individual learning with the guidance of the coach (teacher); thus teachers becoming more like peers. Teachers provide guidance to the students, suggest new approaches and at the same time participate to the network of a “group focused social learning”. Ticéitude implies the utilisation of new technologies but with a humanistic approach; so those tools are used effectively in a sustainable environment aimed to develop French learning not only in Australia but also around the world.

Du fait des énormes capacités de calculs des accélérateurs tels que les GPUs et les Xeon Phi, l’utilisation de machines multicoques pourvues d’accélérateurs est devenue commune dans le domaine du calcul haute performance (HPC). La complexité induite par ces accélérateurs a suscité le développement de systèmes d’exécution à base de tâches, dans lesquels les dépendances entre les applications sont exprimées sous la forme de graphe de tâches et où les tâches sont ordonnancées dynamiquement sur les ressources de calcul. La difficulté est alors de concevoir des stratégies d’ordonnancement qui font une utilisation efficace des ressources de calculs et le développement de telles stratégies, même pour un unique noeud hybride, est un enjeu essentiel de la performance des systèmes HPC. Nous considérons dans cette thèse l’ordonnancement de noyaux d’algèbre linéaire dense sur des noeuds complètement hétérogènes et constitués de CPUs et de GPUs. Les performances relatives des accélérateurs par rapport aux coeurs classique dépend très fortement du noyau considéré. Par exemple, les accélérateurs sont beaucoup plus efficaces pour les produits de matrices, par exemple, que pour les factorisations. Dans cette thèse, nous analysons les performances de stratégies statiques et dynamiques d’ordonnancement et nous proposons un ensemble de stratégies intermédiaires, en ajoutant des composantes statiques (respectivement dynamiques) à des stratégies d’ordonnancements dynamique (respectivement statiques). Récemment, une stratégie appelée HeteroPrio a été proposée, qui s’appuie sur les affinités entre les tâches et les ressources pour un petit ensemble de tâches différentes s’exécutant sur deux types de ressources. Nous avons étendu cette stratégie d’ordonnancement pour des graphes de tâches généraux pour deux types de ressources puis pour plus de deux types. De manière complémentaire, nous avons également démontré des facteurs d’approximation et des pires cas pour HeteroPrio dans le cas d’un ensemble de tâches indépendantes sur différents types de plates-formes. / Due to massive computation power of accelerators such as GPU, Xeon phi, multicore machines equipped with accelerators are becoming popular in High Performance Computing (HPC). The added complexity led to the development of different task-based runtime systems, which allow computations to be expressed as graphs of tasks and rely on runtime systems to schedule those tasks among all resources of the platform. The real challenge is to design efficient schedulers for such runtimes to make effective utilization of all resources. Developing good schedulers, even for a single hybrid node, and analyzing them can thus have a strong impact on the performance of current HPC systems. We consider the problem of scheduling dense linear algebra applications on fully hybrid platforms made of CPUs and GPUs. The relative performance of CPU and GPU highly depends on the sub-routine. For instance, GPUs are much more efficient to process matrix-matrix multiplications than matrix factorizations. In this thesis, we analyze the performance of static and dynamic scheduling strategies and we propose a set of intermediate strategies, by adding static (resp. dynamic) features into dynamic (resp. static) strategies. A resource centric dynamic scheduler, HeteroPrio, which is based on affinity between tasks and resources, has been proposed recently for a set of small independent tasks on two types of resources. We extend and analyze this scheduler for general task graphs first on two types of resources and then on more than two types of resources. Additionally, we provide approximation ratios and worst case examples of HeteroPrio for a set of independent tasks on different platform sizes.

L'objet de cette thèse est d'explorer le croisement entre la didactique de l'espagnol langue vivante étrangère pour son enseignement au collège en France en 4° LV2 et la pratique théâtrale. Il s'agit de développer une dramadidactique, dans l'optique très pragmatique de son utilisation en cours par un enseignant, afin de développer particulièrement la compétence langagière qu'est l'interaction orale. Le résultat concret de ce travail est la production d'une méthode nommée DramAula. L'objectif principal de cette dramadidactique est de permettre de mieux enseigner l'espagnol aux élèves dans la motivation et l'action grâce aux éléments dynamisants des activités que le genre dramatique apporte afin d'obtenir plus de confiance en soi, d'autonomie et de créativité dans l'apprentissage de la langue espagnole par les élèves à l'oral en interaction. Cette didactique s'intègre dans la perspective actionnelle présentée dans le Cadre Européen commun de référence pour les langues. Elle entend développer des éléments qui améliorent et enrichissent l'enseignement actuel. / This thesis aims to analyse Spanish didactic and theatrical performance. Spanish didactic is analysed in the context of learning a second language in school with Year 8 children (4eme). The aim is to develop a theatrical didactic to be used by teacher to improve verbal communication in school. This research allows the creation of a new methodology called DramAula. The main objective of this work is to improve the teaching of a second language in school by developing confidence, autonomy and creativity in children with the help of theatrical activities. Theatrical activity is thought to increase motivation and verbal exchanges in children and create a better learning environment. This didactic method is in accordance with the European program to improve language skills learning.

The present study examines and classifies learning activities found in teaching materials of Spanish as a foreign language produced for a Swedish context with the aim of determining to what extent they conform to the functional, action-oriented language view advocated by the Swedish curriculum. Both the distribution of learning activities within receptive, productive and interactive skills, the activity types within each skill, and the theoretical underpinnings of the activities have been studied in four popular coursebooks and two websites for teaching Spanish as a foreign language. The activities analyzed have shown a dominance of the written language even in oral activities. The results have also shown a bias towards more controlled, non-communicative activities with an explicit focus on forms at the expense of freer communicative activities with a primary focus on meaning, known to support language acquisition.