A framework for scientific computing with GPUs

Dissertação para obtenção do Grau de Mestre em Engenharia Informática

Heterogeneous computing with an algorithmic skeleton framework

The Graphics Processing Unit (GPU) is present in almost every modern day personal computer. Despite its specific purpose design, they have been increasingly used for general computations with very good results. Hence, there is a growing effort from the community to seamlessly integrate this kind of devices in everyday computing. However, to fully exploit the potential of a system comprising GPUs and CPUs, these devices should be presented to the programmer as a single platform. The efficient combination of the power of CPU and GPU devices is highly dependent on each device’s characteristics, resulting in platform specific applications that cannot be ported to different systems. Also, the most efficient work balance among devices is highly dependable on the computations to be performed and respective data sizes. In this work, we propose a solution for heterogeneous environments based on the abstraction level provided by algorithmic skeletons. Our goal is to take full advantage of the power of all CPU and GPU devices present in a system, without the need for different kernel implementations nor explicit work-distribution.To that end, we extended Marrow, an algorithmic skeleton framework for multi-GPUs, to support CPU computations and efficiently balance the work-load between devices. Our approach is based on an offline training execution that identifies the ideal work balance and platform configurations for a given application and input data size. The evaluation of this work shows that the combination of CPU and GPU devices can significantly boost the performance of our benchmarks in the tested environments, when compared to GPU-only executions.

Knowledge management framework based on brain models and human physiology

The life of humans and most living beings depend on sensation and perception for the best assessment of the surrounding world. Sensorial organs acquire a variety of stimuli that are interpreted and integrated in our brain for immediate use or stored in memory for later recall. Among the reasoning aspects, a person has to decide what to do with available information. Emotions are classifiers of collected information, assigning a personal meaning to objects, events and individuals, making part of our own identity. Emotions play a decisive role in cognitive processes as reasoning, decision and memory by assigning relevance to collected information. The access to pervasive computing devices, empowered by the ability to sense and perceive the world, provides new forms of acquiring and integrating information. But prior to data assessment on its usefulness, systems must capture and ensure that data is properly managed for diverse possible goals. Portable and wearable devices are now able to gather and store information, from the environment and from our body, using cloud based services and Internet connections. Systems limitations in handling sensorial data, compared with our sensorial capabilities constitute an identified problem. Another problem is the lack of interoperability between humans and devices, as they do not properly understand human’s emotional states and human needs. Addressing those problems is a motivation for the present research work. The mission hereby assumed is to include sensorial and physiological data into a Framework that will be able to manage collected data towards human cognitive functions, supported by a new data model. By learning from selected human functional and behavioural models and reasoning over collected data, the Framework aims at providing evaluation on a person’s emotional state, for empowering human centric applications, along with the capability of storing episodic information on a person’s life with physiologic indicators on emotional states to be used by new generation applications.

A middleware for service oriented computing in dynamic environments

Dissertação apresentada na Faculdade de Ciências e Tecnologia da Universidade Nova de Lisboa para a obtenção do Grau de Mestre em Engenharia Informática

Parallelization of web processing services on cloud computing: A case study of Geostatistical Methods

Dissertation submitted in partial fulfillment of the requirements for the Degree of Master of Science in Geospatial Technologies.

Algorithmic skeleton framework for the orchestration of GPU computations

Dissertação para obtenção do Grau de Mestre em Engenharia Informática

Cloud computing adoption - Determinants of cloud computing adoption by firms

Dissertação apresentada como requisito parcial para obtenção do grau de Mestre em Estatística e Gestão de Informação

Personalization platform for multimodal ubiquitous computing applications

Dissertação para obtenção do Grau de Mestre em Engenharia Informática

A new MDA-SOA based framework for intercloud interoperability

Cloud computing has been one of the most important topics in Information Technology which aims to assure scalable and reliable on-demand services over the Internet. The expansion of the application scope of cloud services would require cooperation between clouds from different providers that have heterogeneous functionalities. This collaboration between different cloud vendors can provide better Quality of Services (QoS) at the lower price. However, current cloud systems have been developed without concerns of seamless cloud interconnection, and actually they do not support intercloud interoperability to enable collaboration between cloud service providers. Hence, the PhD work is motivated to address interoperability issue between cloud providers as a challenging research objective. This thesis proposes a new framework which supports inter-cloud interoperability in a heterogeneous computing resource cloud environment with the goal of dispatching the workload to the most effective clouds available at runtime. Analysing different methodologies that have been applied to resolve various problem scenarios related to interoperability lead us to exploit Model Driven Architecture (MDA) and Service Oriented Architecture (SOA) methods as appropriate approaches for our inter-cloud framework. Moreover, since distributing the operations in a cloud-based environment is a nondeterministic polynomial time (NP-complete) problem, a Genetic Algorithm (GA) based job scheduler proposed as a part of interoperability framework, offering workload migration with the best performance at the least cost. A new Agent Based Simulation (ABS) approach is proposed to model the inter-cloud environment with three types of agents: Cloud Subscriber agent, Cloud Provider agent, and Job agent. The ABS model is proposed to evaluate the proposed framework.

Towards an algorithmic skeleton framework for programming the Intel R Xeon PhiTM processor

The Intel R Xeon PhiTM is the first processor based on Intel’s MIC (Many Integrated Cores) architecture. It is a co-processor specially tailored for data-parallel computations, whose basic architectural design is similar to the ones of GPUs (Graphics Processing Units), leveraging the use of many integrated low computational cores to perform parallel computations. The main novelty of the MIC architecture, relatively to GPUs, is its compatibility with the Intel x86 architecture. This enables the use of many of the tools commonly available for the parallel programming of x86-based architectures, which may lead to a smaller learning curve. However, programming the Xeon Phi still entails aspects intrinsic to accelerator-based computing, in general, and to the MIC architecture, in particular. In this thesis we advocate the use of algorithmic skeletons for programming the Xeon Phi. Algorithmic skeletons abstract the complexity inherent to parallel programming, hiding details such as resource management, parallel decomposition, inter-execution flow communication, thus removing these concerns from the programmer’s mind. In this context, the goal of the thesis is to lay the foundations for the development of a simple but powerful and efficient skeleton framework for the programming of the Xeon Phi processor. For this purpose we build upon Marrow, an existing framework for the orchestration of OpenCLTM computations in multi-GPU and CPU environments. We extend Marrow to execute both OpenCL and C++ parallel computations on the Xeon Phi. We evaluate the newly developed framework, several well-known benchmarks, like Saxpy and N-Body, will be used to compare, not only its performance to the existing framework when executing on the co-processor, but also to assess the performance on the Xeon Phi versus a multi-GPU environment.

DYNAMOD – A dynamic agent based modelling framework for digital businesses

Digital Businesses have become a major driver for economic growth and have seen an explosion of new startups. At the same time, it also includes mature enterprises that have become global giants in a relatively short period of time. Digital Businesses have unique characteristics that make the running and management of a Digital Business much different from traditional offline businesses. Digital businesses respond to online users who are highly interconnected and networked. This enables a rapid flow of word of mouth, at a pace far greater than ever envisioned when dealing with traditional products and services. The relatively low cost of incremental user addition has led to a variety of innovation in pricing of digital products, including various forms of free and freemium pricing models. This thesis explores the unique characteristics and complexities of Digital Businesses and its implications on the design of Digital Business Models and Revenue Models. The thesis proposes an Agent Based Modeling Framework that can be used to develop Simulation Models that simulate the complex dynamics of Digital Businesses and the user interactions between users of a digital product. Such Simulation models can be used for a variety of purposes such as simple forecasting, analysing the impact of market disturbances, analysing the impact of changes in pricing models and optimising the pricing for maximum revenue generation or a balance between growth in usage and revenue generation. These models can be developed for a mature enterprise with a large historical record of user growth rate as well as for early stage enterprises without much historical data. Through three case studies, the thesis demonstrates the applicability of the Framework and its potential applications.

Molecular mobility of n-ethylene glycol dimethacrylate glass formers upon free radical polymerization

Quando um líquido evita a cristalização durante o arrefecimento, diz-se que entra no estado sobrearrefecido. Se a temperatura continuar a diminuir, o consequente aumento da viscosidade reflecte-se na mobilidade molecular de tal maneira que os tempos característicos se tornam da mesma ordem de grandeza que os tempos acessíveis experimentalmente. Se o arrefecimento continuar, o líquido altamente viscoso acaba por vitrificar, i.e. entra no estado vítreo onde apenas os movimentos locais são permitidos. Os monómeros da família n -etileno glicol dimetacrilato ( n -EGDMA, para n = 1 até 4, que constituem o objecto deste estudo, facilmente evitam a cristalização, sendo pois bons candidatos para estudar a mobilidade molecular nos estados sobrearrefecido e vítreo. A Espectroscopia de Relaxação Dieléctrica (DRS) foi a técnica escolhida para obter informação detalhada sobre a sua dinâmica molecular (Capítulos 1 e 2). A primeira parte deste trabalho consistiu na caracterização dieléctrica dos processos de relaxação existentes acima e abaixo da temperatura de transição vítrea (g T ), a qual aumenta com o aumento do peso molecular (w M ), sendo este resultado confirmado por Calorimetria Diferencial de Varrimento (DSC). No que respeita ao processo cooperativo a , associado à transição vítrea, e ao processo secundário b, observa-se uma dependência com w M , enquanto que o outro processo secundário, g , aparenta ser independente deste factor (Capítulo 3). Nos capítulos seguintes, foram levadas a cabo diferentes estratégias com o objectivo de clarificar os mecanismos que estão na origem destas duas relaxações secundárias (b e g ), assim como conhecer a sua respectiva relação com a relaxação principal (a ). Do estudo, em tempo real, da polimerização isotérmica via radicais livres do TrEGDMA por Calorimetria de Varrimento Diferencial com Modulação de Temperatura (TMDSC), levado a cabo a temperaturas abaixo da g T do polímero final, concluem-se entre outros, dois importantes aspectos: i) que a vitrificação do polímero em formação conduz a graus de conversão relativamente baixos, e ii) que o monómero que está por reagir é expulso da rede polimérica que se forma, dando lugar a uma clara separação de fases (Capítulo 4). Com base nesta informação, o passo seguinte foi estudar separadamente a polimerização isotérmica do di-, tri- e tetra-EGDMA, dando especial atenção às alterações de mobilidade do monómero ainda por reagir. Com as restrições impostas pela formação de ligações químicas, as relaxações a e b detectadas no monómero tendem a desaparecer no novo polímero formado, enquanto que a relaxação g se mantém quase inalterada. Os diferentes comportamentos que aparecem durante a polimerização permitiram a atribuição da origem molecular dos processos secundários: o processo g foi associado ao movimento twisting das unidades etileno glicol, enquanto que a rotação dos grupos carboxilo foi relacionada com a relaxação b (Capítulo 5). No que respeita ao próprio polímero, um processo de relaxação adicional foi detectado, pol b , no poly-DEGDMA, poly-TrEGDMA e poly-TeEGDMA, com características similares ao encontrado nos poli(metacrilato de n -alquilo). Este processo foi confirmado e bem caracterizado aquando do estudo da copolimerização do TrEGDMA com acrilato de metilo (MA) para diferentes composições (Capítulo 6). Para finalizar, o EGDMA, o elemento mais pequeno da família de monómeros estudada, além de vitrificar apresenta uma marcada tendência para cristalizar quer a partir do estado líquido ou do estado vítreo. Durante a cristalização, a formação de uma fase rígida afecta principalmente o processo a , cuja intensidade diminui sem no entanto se observarem modificações significativas na dependência do tempo de relaxação característico com a temperatura. Por outro lado, o processo secundário b torna-se melhor definido e mais estreito, o que pode ser interpretado em termos de uma maior homogeneidade dos micro-ambientes associados aos movimentos locais(Capítulo 7).

Developing persuasive interfaces

MSCC Dissertation in Computer Engineering

Adaptação da framework i* para linhas de produtos

Dissertação de Mestrado em Engenharia Informática

Uma linguagem de domínio específico para a framework i*

Dissertação apresentada para obtenção do Grau de Mestre em Informática, pela Universidade Nova de Lisboa, Faculdade de Ciências e Tecnologia