任务最坏执行时间分析与任务调度检测仿真工具的实现

发布时间：2018-05-25 05:12

本文选题：嵌入式实时系统 + 最坏执行时间　；参考：《电子科技大学》2012年硕士论文

【摘要】：随着社会的发展，计算机的应用已经渗透到了人们经济生活中的各个方面，特别是嵌入式系统的应用正在不断的从军工设备向企业设备以及普通的家用设备转移。这种发展趋势使得嵌入式软件的复杂度以及规模都在不断增长，从而导致了开发时间和费用的不断增长，寻找到一种新的能够提高开发效率的嵌入式软件开发模式是当前研究的重点。由于在嵌入式软件开发中的绝大部分错误是在需求分析的早期阶段引入的，这些错误将随着开发的深入逐渐放大，而且这些错误发现得越晚，对其进行修改所需付出的代价也越大。所以在嵌入式软件开发的早期阶段就发现问题，对于节省嵌入式软件开发时间起着至关重要的作用。本文就致力于在嵌入式实时系统开发的早期阶段，检测系统内任务模型的可调度性。为了检测嵌入式实时系统内任务模型的可调度性，首先应该得到任务的最坏执行时间信息，只有在得到任务最坏执行时间前提下，才能分析任务的可调度性。所以在本文的前一部分，重点讨论了如何采用静态分析的方法来计算任务的最坏执行时间（Worst Case Execution Time，WCET）。采用此方法可以避免传统动态测量方法的各种弊端，如消耗时间长，，结果不准确等，使得到的结果更加安全和准确。在得到任务的最坏执行时间后，就可以开展对任务的可调度性检测工作了，在论文的后半部分，详细介绍了一款嵌入式任务调度检测仿真工具（SchedulingCheckSimulate Tool，SCST）的使用以及开发过程。该工具可以对嵌入式系统模型进行比较准确和快速的可调度性检测，而且能够以甘特图的形式对系统内任务的执行过程进行仿真。本人在工具开发过程中主要负责系统模型和任务模型的建立以及调度算法的设计。本文最后采用静态分析的方法对一些基准程序进行了分析，并用所得到的结果与真实值做对比，验证了分析的精确性；然后用SCST检测了几个任务模型的用例，并与国外同类型软件TIMES作分析结果对比，验证了此工具的高效性和准确性。最后总结了在课题研究中所遇到的问题和困难，指出了分析方法和工具的不足之处以及可以改进的地方。
[Abstract]:With the development of society, the application of computer has penetrated into every aspect of people's economic life, especially the application of embedded system is constantly transferring from military equipment to enterprise equipment and ordinary household equipment. This trend leads to the increasing complexity and scale of embedded software, which leads to the increasing development time and cost. It is the focus of current research to find a new embedded software development model which can improve the development efficiency. Since most of the errors in embedded software development are introduced in the early stages of requirements analysis, these errors will be magnified as the development progresses, and the later these errors are discovered, The more expensive it is to modify it. Therefore, finding problems in the early stage of embedded software development plays an important role in saving embedded software development time. This paper focuses on detecting the schedulability of task model in the early stage of embedded real-time system development. In order to detect the schedulability of the task model in embedded real-time system, the worst-case execution time information of the task should be obtained first. Only when the worst-case execution time is obtained, can the schedulability of the task be analyzed. So in the first part of this paper, we mainly discuss how to use static analysis method to calculate the worst execution time of the task. This method can avoid the disadvantages of traditional dynamic measurement methods, such as long time consumption, inaccurate results and so on, which makes the results more secure and accurate. After getting the worst execution time of the task, the schedulability detection of the task can be carried out. In the second half of the paper, the use and development process of a simulation tool for embedded task scheduling detection, named scheduling CheckSimulate tool SCST, is introduced in detail. The tool can detect the schedulability of embedded system model accurately and quickly, and can simulate the task execution process in the form of Gantt diagram. I am mainly responsible for the establishment of system model and task model and the design of scheduling algorithm in the process of tool development. In the end, the static analysis method is used to analyze some benchmark programs, and the results obtained are compared with the real values to verify the accuracy of the analysis, and then the use cases of several task models are tested with SCST. Compared with the foreign software TIMES, the efficiency and accuracy of the tool are verified. Finally, the problems and difficulties encountered in the research are summarized, and the shortcomings of the analytical methods and tools are pointed out as well as the areas for improvement.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TP368.1

【参考文献】