Approximations for the general block distribution of a matrix

资源类型: 资源大小: 文档分类: 上传者:陈际华

文档信息

【作者】 Bengt Aspvall  Magnús M. Halldórsson  Fredrik Manne 

【关键词】Communication Network Image Compression Block Distribution Parallel Computing University Of Bergen 

【摘要】The general block distribution of a matrix is a rectilinear partition of the matrix into orthogonal blocks such that the maximum sum of the elements within a single block is minimized. This corresponds to partitioning the matrix onto parallel processors so as to minimize processor load while maintaining regular communication patterns. Applications of the problem include various parallel sparse matrix computations, compilers for high-performance languages, particle in cell computations, video and image compression, and simulations associated with a communication network. We analyze the performance guarantee of a natural and practical heuristic based on iterative reenement, which has previously been shown to give good empirical results. When p 2 is the number of blocks, we show that the tight performance ratio is (p p). When the matrix has rows of large cost, the details of the objective function of the algorithm are shown to be important, since a naive implementation can lead to a (p) performance ratio. Extensions to more general cost functions, higher-dimensional arrays, and randomized initial conngurations are also considered.

【刊名】Theor. Comput. Sci.

问答

我要提问