A multicast protocol in multistage interconnection networks

Efficient collective communication among processor nodes is critical to the performance of massively parallel systems. A system-level multicast service in which the same message is delivered from a source node to an arbitrary number of destination nodes, is fundamental in supporting collective communication primitives including the application-level broadcast, reduction, and barrier synchronization. This paper addresses a hardware supported multicast in wormhole-routed multistage networks whose switches have multicast forwarding capability. In this protocol, the source node does not need to send the message to every destination, and any receiving node does not need to forward the message. The proposed multicast protocol can significantly reduce network traffic and transmission time. Moreover, this paper compares the broadcast performance of this protocol with that of two other multicast algorithms