Efficient implementation of barrier synchronization in wormhole-routed hypercube multicomputers
Hong Xu, Philip K. McKinley, Lionel Ming-shuan Ni · 2003
Practical and efficient implementations of barrier synchronization for wormhole-routed hypercube multicomputers are presented. Both broadcast and multicast barrier synchronization are considered. For systems that do not support hardware broadcast or multicast, a software U-cube tree is proposed. This method generalizes to n-dimensional meshes. Performance measurements for several barrier synchronization techniques implemented on a 64-node nCUBE-2 are given.>