基于比较诊断模型的超立方网络诊断算法
发布时间：2018-08-28 15:15
【摘要】：一个有效的诊断算法对多处理器系统而言极其重要。在多处理器系统中,识别所有故障节点的能力称为诊断系统的诊断度。在比较模型下,诊断的执行是通过一个比较器处理器,给与之相邻的一对处理器发送相同的输入信号,并比较两者间的响应状态。为了提高超立方网络的诊断度,提出了一种新型的基于比较模型的超立方故障诊断算法,其利用超立方网络节点连接的特性生成一个拓扑图ES(k;n),最终得出一个3位二进制的诊断症候集,从而确定系统故障节点。该算法的诊断度最优能达到4n,大于传统超立方的诊断度n。
[Abstract]:An effective diagnosis algorithm is very important for multiprocessor systems. In multiprocessor systems, the ability to identify all fault nodes is called the diagnostic degree of the diagnostic system. In the comparison model, the diagnosis is performed by sending the same input signal to the adjacent pair of processors through a comparator processor, and comparing the response state between the two. In order to improve the diagnosis degree of hypercube network, a new hypercube fault diagnosis algorithm based on comparative model is proposed. A topological graph ES (KKN) is generated by using the characteristic of hypercube network node connection. Finally, a 3-bit binary diagnostic symptom set is obtained to determine the fault node of the system. The optimal diagnostic degree of the algorithm is 4 ns, which is larger than that of the traditional hypercube.
【作者单位】： 广西大学计算机与电子信息学院;
【基金】：国家自然科学基金项目:新型互连网络的嵌入性与容错性研究(61364002)资助
【分类号】：TP301.6;TP332
本文编号：2209776
[Abstract]:An effective diagnosis algorithm is very important for multiprocessor systems. In multiprocessor systems, the ability to identify all fault nodes is called the diagnostic degree of the diagnostic system. In the comparison model, the diagnosis is performed by sending the same input signal to the adjacent pair of processors through a comparator processor, and comparing the response state between the two. In order to improve the diagnosis degree of hypercube network, a new hypercube fault diagnosis algorithm based on comparative model is proposed. A topological graph ES (KKN) is generated by using the characteristic of hypercube network node connection. Finally, a 3-bit binary diagnostic symptom set is obtained to determine the fault node of the system. The optimal diagnostic degree of the algorithm is 4 ns, which is larger than that of the traditional hypercube.
【作者单位】： 广西大学计算机与电子信息学院;
【基金】：国家自然科学基金项目:新型互连网络的嵌入性与容错性研究(61364002)资助
【分类号】：TP301.6;TP332
【相似文献】
相关期刊论文 前1条
1 唐万梅，陈迁槐，，杨晓帆;用神经网络求解P－t可诊断模型下的计算机系统诊断问题[J];重庆大学学报(自然科学版);1994年02期
本文编号：2209776
本文链接：https://www.wllwen.com/kejilunwen/jisuanjikexuelunwen/2209776.html
