Com o advento da computação em nuvem, não é mais necessário ao usuário investir grandes quantidades de recursos financeiros em equipamentos computacionais. Ao invés disto, é possível adquirir recursos de processamento, armazenamento ou mesmo sistemas completos por demanda, usando um dos diversos serviços disponibilizados por provedores de nuvem como a Amazon, o Google, a Microsoft, e a própria USP. Isso permite um controle maior dos gastos operacionais, reduzindo custos em diversos casos. Por exemplo, usuários de computação de alto desempenho podem se beneficiar desse modelo usando um grande número de recursos durante curtos períodos de tempo, ao invés de adquirir um aglomerado computacional de alto custo inicial. Nosso trabalho analisa a viabilidade de execução de aplicações de alto desempenho, comparando o desempenho de aplicações de alto desempenho em infraestruturas com comportamento conhecido com a nuvem pública oferecida pelo Google. Em especial, focamos em diferentes configurações de paralelismo com comunicação interna entre processos no mesmo nó, chamado de intra-nós, e comunicação externa entre processos em diferentes nós, chamado de entre-nós. Nosso caso de estudo para esse trabalho foi o NAS Parallel Benchmarks, um benchmark bastante popular para a análise de desempenho de sistemas paralelos e de alto desempenho. Utilizamos aplicações com implementações puramente MPI (para as comunicações intra e entre-nós) e implementações mistas onde as comunicações internas foram feitas utilizando OpenMP (comunicação intra-nós) e as comunicações externas foram feitas usando o MPI (comunicação entre-nós). / With the advent of cloud computing, it is no longer necessary to invest large amounts of money on computing resources. Instead, it is possible to obtain processing or storage resources, and even complete systems, on demand, using one of the several available services from cloud providers like Amazon, Google, Microsoft, and USP. Cloud computing allows greater control of operating expenses, reducing costs in many cases. For example, high-performance computing users can benefit from this model using a large number of resources for short periods of time, instead of acquiring a computer cluster with high initial cost. Our study examines the feasibility of running high-performance applications, comparing the performance of high-performance applications in a known infrastructure compared to the public cloud offering from Google. In particular, we focus on various parallel configurations with internal communication between processes on the same node, called intra-node, and external communication between processes on different nodes, called inter-nodes. Our case study for this work was the NAS Parallel Benchmarks, a popular benchmark for performance analysis of parallel systems and high performance computing. We tested applications with MPI-only implementations (for intra and inter-node communications) and mixed implementations where internal communications were made using OpenMP (intra-node communications) and external communications were made using the MPI (inter-node communications).

Os problemas de engenharia atualmente têm aumentado a sua ordem de grandeza, por conta de diversos fatores. A modelagem em ambiente computacional dos mesmos esbarra em limitações, como grandes quantidades de tempo de processamento gastos com diversas simulações da modelagem e a pouca quantidade de memória disponível para alocar propriamente os problemas. A resolução de grandes sistemas de equações lineares, comumente abordado nos problemas atuais de engenharia, necessita da exploração das duas situações mencionadas anteriormente. A subárea computacional que permite explorar a redução do tempo e a possibilidade de alocação na memória de tais problemas é chamada de computação de alto desempenho. O objetivo deste trabalho é ilustrar o uso de softwares de resolução de sistemas de equações lineares, chamados de solvers, projetados para os ambientes computacionais de alto desempenho, testando-os e avaliando-os em um conjunto de matrizes conhecido, bem como abordar os detalhes computacionais envolvidos em tais procedimentos. / Engineering problems today have increased their order of magnitude, due to several factors. Modeling these problems with computers brings up certain limitations, as the amount of processing time needed for several simulations and the lack of available memory to properly allocate them. The resolution of large systems of linear equations, commonly discussed in current engineering problems, needs the exploration of the two situations mentioned above. The subarea that allows exploring the computational time reduction and the possibility of allocating memory in such problems is called high performance computing. The aim of this paper is to illustrate the use of software to solve systems of linear equations, called solvers, designed for high performance computing environments, to test and evaluate them for a set of matrices as well as to address the computational details involved in such procedures.

There is now more data being created than ever before and this data can be any form of data, textual, multimedia, spatial etc. To process this data, several big data processing platforms have been developed including Hadoop, based on the MapReduce model and LexisNexis’ HPCC systems. In this thesis we evaluate the HPCC Systems framework with a special interest in multimedia data analysis and propose a framework for multimedia data processing. It is important to note that multimedia data encompasses a wide variety of data including but not limited to image data, video data, audio data and even textual data. While developing a unified framework for such wide variety of data, we have to consider computational complexity in dealing with the data. Preliminary results show that HPCC can potentially reduce the computational complexity significantly. / Includes bibliography. / Thesis (M.S.)--Florida Atlantic University, 2017. / FAU Electronic Theses and Dissertations Collection

Natural matte extraction is a difficult and generally unsolved problem. Generating a matte from a nonuniform background traditionally requires a tediously hand drawn matte. This thesis studies recent methods requiring the user to place only modest scribbles identifying the foreground and the background. This research demonstrates a new GPU-based implementation of the recently introduced Fuzzy- Matte algorithm. Interactive matte extraction was achieved on a CUDA enabled G80 graphics processor. Experimental results demonstrate improved performance over the previous CPU based version. In depth analysis of experimental data from the GPU and the CPU implementations are provided. The design challenges of porting a variant of Dijkstra's shortest distance algorithm to a parallel processor are considered. / by Joel Gibson. / Thesis (M.S.C.S.)--Florida Atlantic University, 2008. / Includes bibliography. / Electronic reproduction. Boca Raton, Fla., 2008. Mode of access: World Wide Web.

Video identification or copy detection is a challenging problem and is becoming increasingly important with the popularity of online video services. The problem addressed in this thesis is the identification of a given video clip in a given set of videos. For a given query video, the system returns all the instance of the video in the data set. This identification system uses video signatures based on video tomography. A robust and low complexity video signature is designed and implemented. The nature of the signature makes it independent to the most commonly video transformations. The signatures are generated for video shots and not individual frames, resulting in a compact signature of 64 bytes per video shot. The signatures are matched using simple Euclidean distance metric. The results show that videos can be identified with 100% recall and over 93% precision. The experiments included several transformations on videos. / by Gustavo A. Leon. / Thesis (M.S.C.S.)--Florida Atlantic University, 2008. / Includes bibliography. / Electronic reproduction. Boca Raton, Fla., 2008. Mode of access: World Wide Web.

O crescente desenvolvimento de aplicações de alto desempenho é uma realidade presente nos dias atuais. Entretanto, a diversidade de arquiteturas de máquinas, incluindo monoprocessadores e multiprocessadores, clusters com ou sem máquina front-end, variedade de sistemas operacionais e implementações da biblioteca MPI tem aumentado cada dia mais. Tendo em vista este cenário, bibliotecas que proporcionem a integração de diversas implementações MPI, sistemas operacionais e arquiteturas de máquinas são necessárias. Esta tese apresenta o HyMPI, uma implementação da biblioteca MPI voltada para integração, em um mesmo ambiente distribuído de alto desempenho, nós com diferentes arquiteturas, clusters com ou sem máquina front-end, sistemas operacionais e implementações MPI. HyMPI oferece um conjunto de primitivas compatíveis com a especificação MPI, incluindo comunicação ponto a ponto, operações coletivas, inicio e termino, além de outras primitivas utilitárias. / The increasing develpment of high performance applications is a reality on current days. However, the diversity of computer architectures, including mono and multiprocessor machines, clusters with or without front-end node, the variety of operating systems and MPI implementations has growth increasingly. Focused on this scenario, programming libraries that allows integration of several MPI implementations, operating systems and computer architectures are needed. This thesis introduces HyMPI, a MPI implementation aiming integratino, on a distributed high performance system nodes with different architectures, clusters with or without front-end machine, operating systems and MPI implementations. HyMPI offers a set of primitives based on MPI specification, including point-to-point communication, collective operations, startup and finalization and some other utility functions.

Leong Monk Ping. / Thesis (Ph.D.)--Chinese University of Hong Kong, 2001. / Includes bibliographical references (p. 255-278). / Electronic reproduction. Hong Kong : Chinese University of Hong Kong, [2012] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Electronic reproduction. Ann Arbor, MI : ProQuest Information and Learning Company, [200-] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Mode of access: World Wide Web. / Abstracts in English and Chinese.

Designs of human-computer systems intended for time-critical multitasking can benefit from an understanding of the human factors that support or limit multitasking performance and a detailed account of the human-machine interactions that unfold in a given task environment. An integrated, computational cognitive model can test and provide such an understanding of the human factors related to multitasking and reveal the dynamic interactions that occur in the task at the level of hundreds of milliseconds. This dissertation provides such a detailed computation model of human multitasking, built for a time-critical, multimodal dual task experiment and validated by the eye tracking data collected from the experiment. This dissertation also develops new approaches to conducting cognitive modeling, which enable efficient and systematical exploration of multitasking strategies, as well as principled model comparisons. The dual task experiment captures many key aspects of real-world multitasking scenarios such as driving. In the experiment, the participant interleaved two tasks: one requires tracking a constantly-moving target with a joystick, and the other requires keying-in responses to objects moving across a radar display. Peripheral visibility and auditory conditions of the experiment were manipulated to assess the influence of peripheral visual information and auditory information on multitasking performance. Detailed eye tracking data were collected, and this dissertation presents a detailed analysis of this set of data, which provides the bases for model development and validation. The cognitive model presented in this dissertation, built based on the Executive Processes-Interactive Control cognitive architecture, accurately accounted for the eye movement data and other behavioral data of each participant using systematic explorations of task strategies and parameters configured for each individual participant. A parallelized cognitive modeling system was developed to accommodate the much increased computational demand of strategy exploration and individualized model building. New model comparison techniques were proposed to determine which strategy best accounts for the empirical data. Payoff analyses were applied, and they revealed people’s tendency to locally optimize task performance based on task payoff as well as instantaneous feedback. The results point to new approaches for building a priori models that predict multitasking performance. This dissertation includes previously published coauthored material.

The price of computers has dropped drastically over the past years enabling many households to have at least one computer. At the same time, the performance of computers has skyrocketed, far surpassing what a typical user needs, and most of the computational power of personal computers is wasted. Volunteer computing projects attempt to use this wasted computational power in order to solve problems that would otherwise be computationally infeasible. Some of these problems include medical applications like searching for cures for AIDS and cancer. However, the number of volunteer computing projects is increasing rapidly, requiring improvements in the field of volunteer computing to enable the increasing number of volunteer projects to continue making significant progress. This dissertation examines two ways to increase the productivity of volunteer computing: using the volunteered CPU cycles more effectively and exploring ways to increase the amount of CPU cycles that are donated. Each of the existing volunteer computing projects uses one of two task retrieval policies to enable the volunteered computers participating in projects to retrieve work. This dissertation compares the amount of work completed by the volunteered computers participating in projects based on which of the two task retrieval techniques the project employs. Additional task retrieval policies are also proposed and evaluated. The most commonly used task retrieval policy is shown to be less effective than both the less frequently used policy and a proposed policy. The potential that video game consoles have to be used for volunteer computing is explored, as well as the potential benefits of constructing different types of volunteer computing clients, rather than the most popular client implementation: the screensaver. In addition to examining methods of increasing the productivity of volunteer computing, 140 traces of computer usage detailing when computers are available to participate in volunteer computing is collected and made publicly available. Volunteer computing project-specific information that can be used in researching how to improve volunteer computing is collected and combined into the first summary of which we are aware.

Understanding the performance of applications on modern multi- and manycore platforms is a difficult task and involves complex measurement, analysis, and modeling. The Roofline model is used to assess an application's performance on a given architecture. Not much work has been done with the Roofline model using real measurements. Because it can be a very useful tool for understanding application performance on a given architecture, in this thesis we demonstrate the use of architectural roofline data with measured data for analyzing the performance of different benchmarks. We first explain how to use different toolkits to measure the performance of a program. Next, these data are used to generate the roofline plots, based on which we can decide how can we make the application more efficient and remove bottlenecks. Our results show that this can be a powerful tool for analyzing performance of applications over different architectures and different code versions.