An efficient long-lived adaptive collect algorithm

  • Authors:

  • Burkhard Englert

  • Affiliations:

  • Dept. of Comp. Engr. & Comp. Science, California State University Long Beach, Long Beach, CA

  • Venue:
  • DISC'05 Proceedings of the 19th international conference on Distributed Computing
  • Year:
  • 2005

Quantified Score

Hi-index 0.00

Visualization

Abstract

We present a new long-lived, efficient, adaptive collect algorithm. Namely, our algorithm adapts to $\mathcal{K}$-contention – it has the property that if during an operation the interval contention k exceeds a predetermined constant $\mathcal{K}$ the step complexity is O(N). If, it falls below $\mathcal{K}$, the processors executions will eventually have adaptive step complexity of O(k3). Moreover, for $\mathcal{K}$ such that $\mathcal{K}^3 \leq N$ our algorithm requires only O(N2) shared memory registers.