Theoretical and experimental approaches for the hypercube-embedding problem
The hypercube-embedding problem, a restricted version of the general mapping problem, is the problem of mapping a set of communicating processes to a hypercube multiprocessor. The goal is to find a mapping that minimizes the length of the paths between communicating processes so that communication overhead is minimized. Unfortunately, the hypercube-embedding problem has been shown to be NP-hard, even for trees. This thesis studies both experimental and theoretical issues for the hypercube-embedding problem. Many heuristics have been proposed for hypercube embedding. For experimental studies, a versatile test bed is established for the evaluation of heuristics. Extensive experiments were performed for a wide range of hypercube-embedding heuristics chosen from the literature. Overall, ten different heuristics are evaluated. It is shown that two of the new heuristic proposed are particularly successful in comparison with other heuristics. For theoretical studies, a simple linear-time heuristic is presented which embeds arbitrary binary trees to hypercubes with expansion 1 and average dilation no more than 2.
- Research Organization:
- North Carolina State Univ., Raleigh, NC (United States)
- OSTI ID:
- 5431083
- Country of Publication:
- United States
- Language:
- English
Similar Records
On the efficient simulation of networks by hypercube machines
Distributed multi-destination routing in hypercube multiprocessors