Scientists continually require additional processing power, as this enables them to compute larger problem sizes, use more complex models and algorithms, and solve problems previously thought computationally impractical. General-purpose computation on graphics processing units (GPGPU) can help in this regard, as there is great potential in using graphics processors to accelerate many scientific models and algorithms. However, some problems are considerably harder to accelerate than others, and it may be challenging for those new to GPGPU to ascertain the difficulty of accelerating a particular problem or seek appropriate optimisation guidance. Through what was learned in the acceleration of a hydrological uncertainty ensemble model, large numbers of k-difference string comparisons, and a radix sort, problem attributes have been identified that can assist in the evaluation of the difficulty in accelerating a problem using GPUs. The identified attributes are inherent parallelism, branch divergence, problem size, required computational parallelism, memory access pattern regularity, data transfer overhead, and thread cooperation. Using these attributes as difficulty indicators, an initial problem difficulty classification framework has been created that aids in GPU acceleration difficulty evaluation. This framework further facilitates directed guidance on suggested optimisations and required knowledge based on problem classification, which has been demonstrated for the aforementioned accelerated problems. It is anticipated that this framework, or a derivative thereof, will prove to be a useful resource for new or novice GPGPU developers in the evaluation of potential problems for GPU acceleration.

It is an acknowledged fact that many novice programmers experience difficulty in the process of learning to program. One of the contributing factors to this difficulty is the Program Development Environment (PDE). Professional-PDEs are those developed specifically for professional programmers, but are often used by educational institutions in the instruction of programming. It has long been accepted that such environments are inappropriate in the instruction of programming due to unnecessary complexity and lack of support for novice programmers in the learning process. Numerous pedagogical-PDEs supporting the mechanics of programming have been developed in response to this. A review of literature, however, indicates that very limited empirical studies comparing pedagogical-PDEs and professional-PDEs have been conducted. The current study investigates whether there are measurable benefits to using a pedagogical-PDE supporting the mechanics of programming in the instruction of programming instead of a professional-PDE. A comparative study of this nature requires a representative pedagogical-PDE and representative professional-PDE be compared with one another. The first part of the current study determines a set of requirements that a pedagogical- PDE should adhere to based on literature. A set of representative features for a pedagogical-PDE is derived by examining the features of existing PDEs in conjunction with the set of requirements. Based on these features, a pedagogical-PDE, known as SimplifIDE, is developed that implements the representative set of features and that meets are the requirements for a pedagogical-PDE. The second part of the current study is the specification and administration of an empirical experiment in which SimplifIDE and Borland© DelphiTM are compared with one another. A holistic approach in determining the differences between the PDEs is taken and three main areas are examined, namely academic performance, perceptions and programming behavior.

This study examined the appropriateness of introducing computer programming to kindergarten children. Three issues were explored in the research: 1. the programming capabilities of kindergarten children using a single keystroke program 2. suitable teaching techniques and learning environments for introducing programming 3. the benefits of programming at the kindergarten level. The subjects for the study were 40 kindergarten students from a surburban community in British Columbia, Canada. All students used the single keystroke program, DELTA DRAWING. Two teaching techniques were used—a structured method and a guided discovery method. Quantitative data were collected by administering five skills tests (skills relating to programming) as pretests and postests to both groups. A programming posttest was also given. Qualitative data were obtained by recording detailed observation reports for each of the 22 lessons (11 for each group), conducting an interview with each child at the end of the study and distributing a parent questionnaire. It can be concluded that it is appropriate to introduce computer programming to kindergarten students. The children in this study showed they are capable of programming. All students mastered some programming commands to instruct the "turtle" to move on the screen. DELTA DRAWING was determined to be a suitable means to introduce programming to kindergarten children. A combination of a structured teaching method and a guided discovery method is recommended for introducing a single keystroke program. It was observed that students in a guided discovery learning environment are more enthusiastic and motivated than students in a structured environment. Students need time to explore and make discoveries, but some structure is necessary to teach specific commands and procedures which may otherwise not be discovered. Social interaction should be encouraged while children use the computer, however most kindergarten children prefer to work on their own computer. There was no significant difference between the two groups on all but one of the five skills tests for both the pretests and the posttests. On the Programming Test the two groups did not perform significantly different. It can also be concluded that learning to program promotes cognitive development in certain areas. On all but one of the five skills test both the Structured Group and the Guided Discovery Group scored significantly better on the posttest than on the pretest. Lesson observation reports, student interviews and responses on parent questionnaires suggested that the computer experience was positive and rewarding for the kindergarten students. / Education, Faculty of / Graduate

No description available.

This thesis presents and evaluates a new parallel algorithm that computes all single cosets in the double coset M₂₄ P M₂₄, where P is a permutation on n points of a certain cycle structure, and M₂₄ is the Mathieu group related to a Steiner system S(5, 8, 24) as its automorphism group. The purpose of this work is not to replace the existing algorithms, but rather to explore a possibility to extend calculations of single cosets beyond the limits encountered when using currently available methods.

Thesis: M.S., Massachusetts Institute of Technology, Department of Mechanical Engineering, 1976 / MICROFICHE COPY AVAILABLE IN ARCHIVES AND ENGINEERING. At head of title: T-624. / Includes bibliographical references. / by Sergio N. Simunovic S. / M.S. / M.S. Massachusetts Institute of Technology, Department of Mechanical Engineering

We propose a scaleable multi-user benchmark called XMach-1 (XML Data Management benchmark) for evaluating the performance of XML data management systems. It is based on a web application and considers different types of XML data, in particular text documents, schema-less data and structured data. We specify the structure of the benchmark database and the generation of its contents. Furthermore, we define a mix of XML queries and update operations for which system performance is determined. The primary performance metric, Xqps, measures the query throughput of a system under response time constraints. We will use XMach-1 to evaluate both native XML data management systems and XML-enabled relational DBMS.

We present a new approach for accelerating the execution of XPath expressions using parameterized materialized XPath views (PXV). While the approach is generic we show how it can be utilized in an XML extension for relational database systems. Furthermore we discuss an algorithm for automatically determining the best PXV candidates to materialize based on a given workload. We evaluate our approach and show the superiority of our cost based algorithm for determining PXV candidates over frequent pattern based algorithms.

We propose the idea of deep probabilistic programming, a synthesis of advances for systems at the intersection of probabilistic modeling and deep learning. Such systems enable the development of new probabilistic models and inference algorithms that would otherwise be impossible: enabling unprecedented scales to billions of parameters, distributed and mixed precision environments, and AI accelerators; integration with neural architectures for modeling massive and high-dimensional datasets; and the use of computation graphs for automatic differentiation and arbitrary manipulation of probabilistic programs for flexible inference and model criticism. After describing deep probabilistic programming, we discuss applications in novel variational inference algorithms and deep probabilistic models. First, we introduce the variational Gaussian process (VGP), a Bayesian nonparametric variational family, which adapts its shape to match complex posterior distributions. The VGP generates approximate posterior samples by generating latent inputs and warping them through random non-linear mappings; the distribution over random mappings is learned during inference, enabling the transformed outputs to adapt to varying complexity of the true posterior. Second, we introduce hierarchical implicit models (HIMs). HIMs combine the idea of implicit densities with hierarchical Bayesian modeling, thereby defining models via simulators of data with rich hidden structure.

The implementation of collaboration and the use of graphic organizers in the teaching of programming and writing in the elementary grades have proven to be effective instructional strategies. There is evidence that shows the students who are motivated to program and perform well in this content area are not necessarily representative of the students who are motivated to write. Since the organizational skills required in the two content areas are similar, there may be an opportunity to motivate students who engage in computer programming to become more motivated in writing. As a result, the purpose of this study was to investigate the change in the dimensions of motivation which are: challenge, choice, enjoyment, and interest of fifth-grade students to engage in an expository writing activity after being taught to develop computer programs with the same teaching strategies used in the writing activity. The results of this study suggest that the teaching of computer programming was not effective with the intention of motivating the masses of fifth-grade students to write. However, there appears to be supporting evidence that teaching computer programming to fifth-grade students may help some individual students who are not initially motivated to write.