Study on Broadcasting Algorithm for Parallel Computers

主要研究成果如下：(1)针对链路故障的故障超立方体网络，提出了一种新的全对全个性化通信算法。该算法通过扩展全对全广播算法，在故障节点最多为n/2的情况下实现容错。(2)讨论了多级互联网络中的容错组播算法，以构建高效的通信算法。在这项工作中，根据所需的路由步骤评估了总共九种分馏化方案。最后，我们推荐了两种具有相同求值的容错算法。(3)我们提出了一种故障超立方体网络的一对一路由算法，并表明通过在路由中引入初步知识可以找到相当数量的未探索的最短路径。

Our main results are as follows :(1)We proposed a new all-to-all personalized communication algorithm for faulty hypercube networks with link faults. This algorithm is obtained by extending an all-to-all broadcasting algorithm fault-tolerant in the presence of up to n/2 faulty nodes.(2)We presented a discussion on fault-tolerant multicasting algorithms in multistage interconnection networks for constructing efficient communication algorithm. In this work, total nine fractionalizecl schemes are evaluated in terms of required routing steps. Finally, we recommended two fault-tolerant algorithms with the same evaluation.(3)We presented a one-to-one routing algorithm for faulty hypercube networks, and showed that a considerable number of unexplored shortest paths can be found by introducing a preliminary knowledge into the routing.