A Concurrent Multithreaded Scheduling Model for Solving Fibonacci Series on Multicore Architecture
A Concurrent Multithreaded Scheduling Model for Solving Fibonacci Series on Multicore Architecture
复制标题
DOI：
10.4156/ijact.vol3.issue2.4
复制
发表时间：
2011-12
期刊：
影响因子：
--
通讯作者：
Lee Sai Peck
中科院分区：
文献类型：
--
作者：
Lee Sai Peck
多核技术的出现使个人电脑的硬件设计发生了根本性的变化。这些变化表现为每个芯片的核心数量增加。事实上，这种增长不仅在硬件方面，而且在软件方面，都在规划方面施加了新的方向。这项研究是关于开发新的算法，使一类软件(分而治之)更适合多核体系结构。我们提出了一个新的算法和一个改进的算法，证明了它们在多核环境下执行并发多线程调度的能力。为了使这些算法发挥作用，我们设计了一个由这些算法控制的层次模型。通过建模工具对所设计的模型进行了成功的仿真。模型化设计解决了以前研究中的几个缺陷。此外，它提供了一种新的、高效的负载分配方法，而不是不时地产生轻微的开销来处理排序。
The emergence of multicore technology has led to essential changes in the hardware design of personal computers. These changes are represented by an increased growth of cores per chip. As a matter of fact, this growth has imposed new directions in planning not only in the hardware but also in the software side. This study is about developing new algorithms that make a class of software (Divide and Conquer) more adaptable with multicore architecture. We propose one novel algorithm and one enhanced algorithm that prove their ability to perform concurrent multithreaded scheduling for solving Fibonacci series problem on a multicore environment. To put these algorithms to work, we have designed a hierarchal model that is controlled by these algorithms. The designed model has been simulated successfully through a modelling tool. The modelled design solved several shortcomings in the previous studies. Furthermore, it provides a new and efficient way in load-distribution other than incurring a slight overhead to handle sorting from time to time.