Communication-Aware Processor Allocation for Supercomputers

摘要

We give processor-allocation algorithms for grid architectures, where the objective is to select processors from a set of available processors to minimize the average number of communication hops. The associated clustering problem is as follows: Given n points in R~d, find a size-k subset with minimum average pairwise L_1 distance. We present a natural approximation algorithm and show that it is a 7/4-approximation for 2D grids. In d dimensions, the approximation guarantee is 2 — 1/(2d), which is tight. We also give a polynomial-time approximation scheme (PTAS) for constant dimension d and report on experimental results.