SET-MRTS:一种多处理器实时系统可调度性实验平台

发布时间：2018-06-05 20:09

本文选题：实时嵌入式系统 + 多核系统　；参考：《计算机应用》2017年05期

【摘要】：近年来,随着实时调度研究的快速发展,可调度性实验的复杂性随之增加,然而,由于缺乏标准化、模块化的可调度性实验工具,研究者往往需要耗费大量时间进行实验;此外,由于实验源码不能公开获得,使得实验结果难以验证,实验代码难以重用与扩展。针对可调度性实验重复工作量大、难以验证的问题,提出一种可调度性实验基础框架。该框架通过随机分布产生任务系统集合,并测试其可调度性,基于该框架设计并实现了一个新的可调度性实验开源平台——SET-MRTS。该平台采用模块化架构设计了任务模块、处理器模块、共享资源模块、算法库、配置解析模块以及输出模块。实验结果表明,SET-MRTS支持单/多处理器实时调度算法和实时同步协议分析,能够正确地进行可调度性对比实验,输出直观的实验结果,并且支持算法库的扩充,与算法库中已实现的算法进行对比实验,具有良好的兼容性与可扩展性。SET-MRTS是第一个支持完整实验流程,包括算法实现、参数配置、结果统计、图表绘制等的可调度性实验开源平台。
[Abstract]:In recent years, with the rapid development of real-time scheduling research, the complexity of schedulability experiments increases. However, due to the lack of standardization and modularization of schedulability experimental tools, researchers often spend a lot of time on experiments. In addition, because the experimental source code can not be obtained openly, it is difficult to verify the experimental results and to reuse and extend the experimental code. In order to solve the problem that schedulability experiment is difficult to verify, a schedulability experimental framework is proposed. The framework generates the task system set by random distribution and tests its schedulability. Based on this framework, a new open source schedulability experiment platform, SET-MRTSs, is designed and implemented. The platform uses modular architecture to design task module, processor module, shared resource module, algorithm library, configuration analysis module and output module. The experimental results show that SET-MRTS supports single / multiprocessor real-time scheduling algorithm and real-time synchronization protocol analysis, can correctly carry out schedulability contrast experiments, output intuitive experimental results, and support the expansion of algorithm library. Compared with the algorithm implemented in the algorithm library, it has good compatibility and expansibility. SET-MRTS is the first to support the complete experimental flow, including algorithm implementation, parameter configuration, result statistics, etc. Chart drawing and other schedulability experiment open source platform.
【作者单位】：电子科技大学信息与软件工程学院;云南省电子工业研究所;
【基金】：国家科技支撑计划项目(2012BAH44F02)~~
【分类号】：TP332

【相似文献】