<p>Critical software requires formal and rigorous inspection to achieve the required quality, and good documentation provides a solid basis for inspection. The Document Driven Inspection approach takes advantage of precise and complete documentation to serve as a mathematically rigorous and effective technique to review software in a disciplined way. However, it is often the case that precise and complete documents are not available to the inspectors. In these cases, the Document Driven Inspection approach is still useful as illustrated by our case study. As far as we are aware, this is the first application of this approach to an object-oriented critical software system. In this thesis, we investigate several state-of-the-art techniques in software inspection, and also some new techniques that focus on inspection of object-oriented design and code. The Document Driven Inspection approach proposed by Parnas is introduced and further analyzed as applied to an object-oriented design, especially those with incomplete documentation. We illustrate the application of this approach as applied to an object-oriented case study, by producing the complete and rigorous documents that can then act as a basis for further inspection. These documents also illustrate how to document an object-oriented design using tabular expressions.</p> / Master of Science (MSc)

<p>The multicore revolution in chip design has fundamentally altered the demands placed on developers. Thread-level parallelism is critical to optimizing software performance on multicore chips. However thread-level parallelism presents challenges with respect to optimization, safety and program representation. Program models and compiler technologies must act as a bridge from applications to efficient hardware usage.</p> <p>Coconut (COde CONstructing User Tool) is an ongoing project at McMaster to develop a platform for experimenting with novel ideas in reliable and high performance code generation, currently targeting the Cell/B.E.. The Coconut Multicore Framework uses a virtual machine abstraction layer to model multicore layer parallelism on the Cell/B.E.. The abstraction creates a correspondence between ILP and multicore layers of parallelism. The abstraction also allows us to perform efficient static analysis of virtual machine programs; with this ability we have developed a tool to automatically check for parallel bugs in linear time with respect to the atomic virtual machine instructions.</p> <p>In this thesis we will discuss the creation of a performance simulation tool developed to simulate the execution of our virtual machine instructions on a Cell/B.E.. The tool has scalability to future many-core architectures, due to its linearly bounded runtime complexity. The tool allows for Coconut developers to contrast the performance of different scheduling algorithms. It provides meaningful feedback as to optimization opportunities by identifying data transfer latencies which cause execution to stall. The design and performance testing results of the performance simulation tool are presented.</p> / Master of Science (MS)

Thesis (M.S.)--University of Nevada, Reno, 2005. / "December, 2005." Includes bibliographical references (leaves 73-77). Online version available on the World Wide Web.

Thesis (Ph. D.)--University of Texas at Austin, 2000. / Vita. Includes bibliographical references (leaves 131-144). Available also in a digital version from Dissertation Abstracts.

<p>The traditional "stack grows from the top, heap grows from the bottom" memory layout allows a single-threaded process to make use of all available address space. This layout is not ideal when multiple threads of execution need to share one address space, for memory exhaustion is no longer signified by the heap meeting the stack. In the commonly used multi-threaded memory layout where each thread has its "worst case" stack memory exclusively reserved, a process may prematurely run out of memory when one thread's stack collides with another's, even if there is unused address space elsewhere. This problem is exacerbated as the number of threads in a process increases since there is less stack space available per thread.<br />In this thesis, alternative stack mechanisms that attempt to alleviate this problem are reviewed, and a new stack mechanism is put forward that utilizes the MMU to detect stack overflow. An experimental compiler implementing a subset of the C language is used to implement promising stack mechanisms, and a suite of test programs are used to compare their performance and scalability under varying usage patterns.</p> / Master of Science (MS)

<p><em>Repetition is the reality and seriousness of life.</em><br /><em> - Soren Kierkegaard</em><br />The study of repetitions exhibits roots in many modern sciences - sinusoidal waves in physics (smooth repetitive oscillations such as the electromagnetic spectrum), highly repetitive DNA in biology (tandem repeats, satellite DNA), regularities of ciphertexts in cryptography and the periodicity of sounds and sequences in music. A string on a given alphabet ∑ provides the simplest common representation of this underlying property. A <strong><em>repetition</em></strong> defined on a string consists of two or more adjacent identical substrings (e.g. abab or aaaa).<br />A particular problem regarding repetitions is to count the number of different repetitions in a string. Conventional approaches execute in ϴ(<em>n</em>log<em>n</em>) time (Crochemore, 1981; Apostolico and Preparata, 1983; Main and Lorentz, 1984) and employ computationally heavy preprocessing. An ϴ(<em>n</em>) time algorithm introduced in 2000 (Kolpakov and Kucherov, 2000) prevailed over its slower predecessors by succinctly encoding all repetitions as runs. A <em>run</em> is a maximally periodic (nonextendible) substring. For example, the string <em>abaabaabb</em> encodes 3 runs - (<em>aba</em>)²(<em>ab</em>), <em>aa</em> (twice) and <em>bb</em>. The first of these identifies three repetitions - (<em>aba</em>)², (<em>baa</em>)² and (<em>aab</em>)² In the early part of this thesis, we survey current algorithms for computing all repetitions.</p> <p>Brute force is the essential drawback of previous attempts for detecting repetitions, despite evidence and proof that their occurrence in strings is sparse (Puglisi and Simpson, 2008). By establishing combinatorial constraints to predict the expected sparsity of runs, extant preprocessing may be reformatted to exclude redundant computations. In (Fan <em>et al.</em>, 2006), it was shown that if two runs begin at the same position <em>i</em>, consequently no runs begin at some neighbouring position<em> i</em>+<em>k</em>. This is the fundamental idea behind our combinatorial work, in which we provide well substantiated conjectures, some of which are supported by proofs, implying that three neighbouring squares in a string force a trivial breakdown of the substring beginning at position <em>i</em> into repetitions of a small period.<br /><br /></p> / Master of Science (MS)

<p>With so many software-related failures happening these days, there is an increasing demand for software quality. Rigorous development approaches, which apply mathematical techniques to the design and implementation, should be getting more consideration as one of the solutions to software reliability.</p> <p>Pre/postcondition approaches and relational approaches are two groups of influential rigorous techniques. Both of them use classical mathematical concepts to describe and simplify programming objects. To further propel the application of these approaches, their relative strengths and limitations in terms of practicability and accessibility need to be identified and elaborated.</p> <p>In this thesis, we conduct a comparative study between the pre/postcondition approaches, proposed by Floyd, Hoare, Dijkstra and Baber, and the relational approaches, proposed by Mills and Parnas. We investigate aspects related to their mathematical models. Their abilities of specifying different termination behaviours, dealing with non-determinism, distinguishing between specifications and descriptions, etc. are discussed. Some practical issues, such as considerations on common programming constructs, side effects, verification procedures, etc. are reviewed. The comparison criteria are grouped into two categories - theory and practice. Under each criterion, we illustrate and evaluate the strength or weakness of each approach. Suggestions regarding the applications of these approaches are also presented.</p> / Master of Science (MS)

<p>Many algorithms, particularly in the area of image processing, are expensive to develop and computationally resource intensive. We illustrate the advantages of symbolic code generation using an example - closed-loop visual target recognition and tracking in extreme lighting conditions. We quantify the effect of symbolic code generation methods on code efficiency, and explain how these methods allowed us to reduce the development time as well as improve reliability. Working directly with symbolic models improves software quality by reducing transcription errors, and enabled us to rapidly prototype different models for the visual tracking application, where the need to evaluate trackers in their real-time context precludes the effective use of scripting languages. We describe the model in detail, including formulations as an optimization problem; explain the challenges in solving the model; present our method of building the solvers; and summarize the impact on the performance of our methods.</p> / Master of Applied Science (MASc)

The diploma thesis deals with the topic of cloud computing and its appropriate deployment in small and medium business. The objective of this thesis is to analyze the offer of Google products and to demonstrate measurable benefits of this solution. Introductory part of the thesis summarizes the history of IT and it is looking for the connection between history of IT and the term cloud computing. This term is described in detail with selected cloud computing characteristics. The main part of the thesis deals with analysis of Google services, technologies and platforms, which could be used in enterprise. The matters are viewed not only economically, but also from the technological and security point of view. In practical implementations are demonstrated measurable benefits of deployment to the Google cloud platform. Finally, the thesis provides the summary of platform's advantages and disadvantages with a description of practical advices for small and medium businesses

Thesis (Ph. D., Information and Computer Science)--University of California, Irvine, 1997. / Includes bibliographical references.