Diiracting Trees

Shared counters are among the most basic coordination structures in multiprocessor computation , with applications ranging from barrier synchronization to concurrent-data-structure design. This paper introduces diiracting trees, novel distributed-parallel structures for shared counting and load balancing. Diiracting trees combine a randomized coordination method together with a combinatorial data structure, to yield a logarithmic depth counter that improves on the log 2 depth of counting networks, and overcomes the resiliency drawbacks of combining trees. Empirical evidence , collected on a simulated distributed shared-memory machine and several simulated message passing architectures, shows that diiracting trees scale better and are more robust then both combining trees and counting networks, currently the most eeective known methods for implementing concurrent counters. Diiracting trees have already been used to implement highly eecient pro-ducer/consumer queues, and we believe diiraction will prove to be an eeective alternative paradigm to combining and queue-locking in the design of many concurrent data structures.