An Efficient Mean Field Annealing Formulation for Mapping Unstructured Domains to Hypercubes
暂无分享,去创建一个
We propose an efficient MFA formulation for mapping unstructured domains to hypercube-connected distributed-memory architectures. In the general MFA formulation, N×P spin variables are maintained and an individual MFA iteration requires Θ(d avg P + P2) time for the mapping of a sparse domain graph with N vertices and average vertex degree of d avg to a parallel architecture with P processors. The proposed hypercube-specific MFA formulation asymptotically reduces the number of spin variables and the computational complexity of an individual MFA iteration to Nlg2P and Θ(d avg lg2P+Plg2P), respectively, by exploiting the topological properties of hypercubes.
[1] Steven J. Plimpton,et al. Massively parallel methods for engineering and science problems , 1994, CACM.
[2] Tevfik Bultan,et al. A New Mapping Heuristic Based on Mean Field Annealing , 1992, J. Parallel Distributed Comput..
[3] John G. Lewis,et al. Sparse matrix test problems , 1982, SGNM.
[4] M. H. Schultz,et al. Topological properties of hypercubes , 1988, IEEE Trans. Computers.