STOC '92 Proceedings of the twenty-fourth annual ACM symposium on Theory of computing
Journal of the ACM (JACM)
Coins, weights and contention in balancing networks
PODC '94 Proceedings of the thirteenth annual ACM symposium on Principles of distributed computing
Contention in counting networks
PODC '94 Proceedings of the thirteenth annual ACM symposium on Principles of distributed computing
Small-depth counting networks and related topics
Small-depth counting networks and related topics
A combinatorial treatment of balancing networks
Journal of the ACM (JACM)
ACM Transactions on Computer Systems (TOCS)
Contention in shared memory algorithms
Journal of the ACM (JACM)
Notes on Sorting and Counting Networks (Extended Abstract)
WDAG '93 Proceedings of the 7th International Workshop on Distributed Algorithms
Distributed Computing
Linearizable counting networks
Distributed Computing
Counting networks with arbitrary fan-out
Distributed Computing
ICDCS '05 Proceedings of the 25th IEEE International Conference on Distributed Computing Systems
Journal of Parallel and Distributed Computing - Special issue: 18th International parallel and distributed processing symposium
The cost of concurrent, low-contention Read&Modify&Write
Theoretical Computer Science - Foundations of software science and computation structures
The impact of randomization in smoothing networks
Proceedings of the twenty-seventh ACM symposium on Principles of distributed computing
Sorting networks and their applications
AFIPS '68 (Spring) Proceedings of the April 30--May 2, 1968, spring joint computer conference
Hi-index | 5.23 |
We present a novel counting network construction, where the number of input wires w is smaller than or equal to the number of output wires t. The depth of our network is @Q(lg^2w), which depends only on w. In contrast, the amortized contention of the network depends on the number of concurrent processes n and the parameters w and t. This offers more flexibility than all previously known networks, with the same number w of input and output wires, whose contention depends only on two parameters, w and n. In case nwlgw, by choosing twlgw the contention of our network is O(nlgw/w), which improves by a logarithmic factor of w over all previously known networks with w wires.