From local to global: an analysis of nearest neighbor balancing on hypercube
SIGMETRICS '88 Proceedings of the 1988 ACM SIGMETRICS conference on Measurement and modeling of computer systems
Dynamic load balancing for distributed memory multiprocessors
Journal of Parallel and Distributed Computing
Load balancing and Poisson equation in a graph
Concurrency: Practice and Experience
Load Balancing in Parallel Computers: Theory and Practice
Load Balancing in Parallel Computers: Theory and Practice
Engineering Diffusive Load Balancing Algorithms Using Experiments
IRREGULAR '97 Proceedings of the 4th International Symposium on Solving Irregularly Structured Problems in Parallel
Convergence of the diffusion method for weighted torus graphs using Fourier analysis
Theoretical Computer Science
Optimum diffusion for load balancing in mesh networks
EuroPar'10 Proceedings of the 16th international Euro-Par conference on Parallel processing: Part I
Hi-index | 0.00 |
The convergence analysis of the Extrapolated Diffusion (EDF) was developed in [7] and [8] for the weighted torus and mesh graphs, respectively using the set $\mathcal{N}_1(i)$ of nearest neighbors of a node i in the graph. In the present work we propose a Diffusion scheme which employs the set $\mathcal{N} _1(i)\cup \mathcal{N}_2(i)$, where $\mathcal{N}_2(i)$ denotes the four neighbors of node i with path length two (see Figure 1) in order to increase the convergence rate. We study the convergence analysis of the new Diffusion scheme with nine neighbors (NEDF) for weighted torus graphs. In particular, we find closed form formulae for the optimum values of the edge weights and the extrapolation parameter. A 60% increase in the convergence rate of NEDF compared to the conventional EDF method is shown analytically and numerically.