K-ary N-cube网络中的维度气泡流控与无死锁完全自适应路由

利用虚跨步切换技术中消息的依存关系只与相邻缓冲区队列相关的特点,设计了一种称为维度气泡流控（DBFC）的新型流控策略.该流控策略建立在虚跨步（VCT）切换和信约流控机制之上,通过分析端口信约值和路由信息实现点点间的流控.在无边带k-ary n -cube网络中,如果采用DBFC流控策略,即使网络中存在环相关,设计的自适应维度气泡路由（ADBR）算法仍可实现无死锁的最短距离的路由.对于以上结论,文中提供了详细的证明.最后,通过修改模拟工具RSIM的网络模拟器--NETSIM的代码,实现了DBFC流控策略和ADBR算法.模拟结果显示,ADBR算法在性能上比常用的维序路由优越,在报文延迟上有近17.5%的降低.     

Adaptive wormhole routing in k-ary n-cubes

Distributed memory multiprocessor (DMMP) systems have gained much attention because their performance can be easily scaled up by increasing the number of processor-memory modules. The k-ary n-cube is the most popular interconnection network topology currently used in DMMPs. Wormhole routing is one of the most promising switching technology and has been used in many new generation multicomputers. Wormhole routing makes the communication latency insensitive to the network diameter and reduces the size of the channel buffer of each router. The concept of virtual channels and virtual networks are widely invented for deadlock-free design. A fully adaptive wormhole routing method for k-ary n-cubes has been proposed by Linder in 1991 [10]. Unfortunately, the need of 2^{n − 1} virtual networks makes it unreasonable. In this paper, we propose a virtual network system to support an adaptive, minimal and deadlock free routing in k-ary n-cubes. It uses only four virtual networks but can get a higher degree of adaptability and higher traffic capacity. Simulation results are presented to verify the performance.     

带有条件故障边的k元2方体的圈嵌入

[k]元[n]方体已经成为分布式储存并行系统最常用的网络拓扑结构。研究带有条件故障边的[k]元2方体的圈嵌入问题,证明了在[k≥4]为偶整数的[k]元2方体中,若其故障边数不超过3且每个顶点至少与两条非故障边相关联,那么该[k]元2方体存在长度在4到[k2]间的任意偶长的无故障圈。     

k元n方体的条件强匹配排除

为了度量发生故障时k元n方体对其可匹配性的保持能力,通过剖析条件故障下使得k元n方体中不存在完美匹配或几乎完美匹配所需故障集的构造,研究了条件故障下使得k元n方体不可匹配所需的最小故障数。当k ≥ 4为偶数且n ≥ 2时,得出了k元n方体这一容错性参数的精确值并对其所有相应的最小故障集进行了刻画;当k ≥ 3为奇数且n ≥ 2时,给出了该k元n方体容错性参数的一个可达下界和一个可达上界。结果表明,选取k为奇数的k元n方体作为底层互连网络拓扑设计的并行计算机系统在条件故障下对其可匹配性有良好的保持能力;进一步地,该系统在故障数不超过2n时仍是可匹配的,要使该系统不可匹配至多需要4n-3个故障元。     

k元n方体的可靠性评估

并行计算机系统功能的实现很大程度上依赖于系统互连网络的性能。为了精确度量以k元n方体为底层拓扑结构的并行计算机系统的容错能力,研究了点故障模型下k元n方体中k元（n-1）方体子网络的可靠性。当k ≥ 3且为奇数时,分别在固定划分模式和灵活划分模式下对k元n方体中不同数目的k元（n-1）方体子网络保持无故障状态的平均失效时间进行了分析,并得出了这一子网络可靠性评估参数的计算公式。结果表明,当基于k为奇数的k元n方体构建的并行计算机系统指派子网络执行用户任务时,在点故障模型下灵活划分模式相比固定划分模式有着更好的容错能力。     

k元船方体网络的可靠性

[k]元[n]方体[Qkn]是设计大规模多处理机系统时最常用的互连网络拓扑结构之一。对于[1≤m≤n-1],设[F]是[Qkn]中的一个由非空点集[VF]和非空边集[EF]构成的故障集,满足[Qkn-F]中不存在[Qkn-m]且[VF]破坏的[Qkn-m]的集合与[EF]破坏的[Qkn-m]的集合互不包含。设[f＊（n,m）]是破坏[Qkn]中的所有子立方[Qkn-m]所需要的故障集[F]的最小基数。证明了对于奇数[k≥3],[fk（n,1）]为[k＋1],[fk（n,n-1）]为[kn-1-1＋n],[f＊（n,m）]的上下界分别为[Cm-1n-1km＋Cm-1n-2km-1]和[km]。举例说明了上界[Cm-1n-1km＋Cm-1n-2km-1]是最优的。     

k元n树互联网络的2-终点可靠性研究

k元n树是一种用于大规模并行处理的高性能互联网络的拓扑结构。该文提出计算k元n树的2-终点可靠性的递归算法,其计算复杂度为O(n)。结合市场上网络元器件可靠性的实际情况,对k元n树的2-终点可靠性进行了分析。当n趋于无穷大时,计算出k元n树的2-终点可靠性的下限。     

Fast barrier synchronization in wormhole k-ary n-cube networks with multidestination worms

This paper presents a new approach to implement fast barrier synchronization in wormhole k-ary n-cubes. The novelty lies in using multidestination messages instead of the traditional single destination messages. Two different multidestination worm types, gather and broadcasting, are introduced to implement the report and wake-up phases of barrier synchronization, respectively. Algorithms for complete and arbitrary set barrier synchronization are presented using these new worms. It is shown that complete barrier synchronization in a k-ary n-cube system with e-cube routing can be implemented with 2n communication start-ups as compared to 2nlog₂k start-ups needed with unicast-based message passing. This leads to an asymptotic improvement by a factor of log₂k. Simulation results for different system and architectural parameters indicate that the new framework can reduce barrier synchronization cost considerably compared to the unicast-based scheme. For arbitrary set barrier, an interesting trend is observed where the synchronization cost keeps on reducing beyond a certain number of participating nodes. The framework demonstrates potential for supporting fast barrier synchronization in large wormhole-routed systems.     

k元n立方网络的k圈排除问题的递归算法

为了度量以k元n立方网络为底层网络拓扑的并行计算机系统的容错能力,通过构造k元n立方网络中使得所有的k元1立方子网都发生故障的最小节点集合的方法,提出求解其k元1立方子网排除点割集的一种递归算法;证明了要使k元n立方网络中所有k元1立方子网都发生故障至少需要破坏掉kn-1个节点。结果表明,在不超过kn-1-1个节点被破坏的情况下,以k元n立方网络为底层拓扑构建的并行计算机系统中依然存在无故障的k元1立方子网。