The method was realized in powerful cluster MPI computing environment using master-slave model. Strategies of load relocation on different nodes and fault-tolerance were adopted.
本算法在高性能计算机集群MPI环境中并行实现时，采用主从模式设计程序，合理调配各节点的计算负载，并应用容错处理手段，达到了较高的并行效率。
The method was realized in powerful cluster MPI computing environment using master-slave model. Strategies of load relocation on different nodes and fault-tolerance were adopted.
本算法在高性能计算机集群MPI环境中并行实现时，采用主从模式设计程序，合理调配各节点的计算负载，并应用容错处理手段，达到了较高的并行效率。
应用推荐