Real-time Scheduling on Heterogeneous Multiprocessors

随着实时计算机应用系统变得越来越大和复杂，在多处理器平台上而不是在单处理器平台上实现这样的应用变得势在必行。 此外，通常情况下，实现平台不是完全同质的，而是由几种不同类型的处理器和其他资源组成。 这个项目将不同种类的处理器和其他资源。 这个项目将研究基本的任务理论问题，和务实的实现问题，出现时，周期性的任务系统将在这样的异构多资源平台上实现。 这项研究将建立在以前的研究，由两名研究人员，在实时系统中的公平调度。 本研究的理论结果将通过应用于北卡罗来纳州大学目前正在开发的一些实施项目进行验证。

As real-time computer application systems become larger and more complex, it is becoming imperative that such applications be implemented on multi-processor platforms rather than on uniprocessor ones. Furthermore it is often the case that the implementation platforms are not completely homogeneous, but are comprised of several different kinds of processors and other resources. This project will different kinds of processors and other resources. This project will study both the fundamental scheduling-theoretic questions, and the pragmatic implementation issues, that arise when periodic task systems are to be implemented on such heterogeneous multi-resource platforms. This research will build upon previous research, performed by both investigators, on fair scheduling in real-time system. The theoretical findings resulting from this research will be validated by applying them to some implementation projects currently under development at the University of North Carolina.