期刊
THEORETICAL COMPUTER SCIENCE
卷 905, 期 -, 页码 69-86出版社
ELSEVIER
DOI: 10.1016/j.tcs.2021.12.016
关键词
Graph embedding; Congestion; Wirelength; Spined cube
In this paper, the importance of graph embedding in the field of data science is discussed, along with the characteristics and improvements of network-based architectures like binary cube and spined cube. Furthermore, an algorithm for embedding the spined cube into a grid and computing the minimum wirelength is proposed using the edge congestion technique.
In the field of data science, graph embedding is a vital tool to simulate, visualize and design a parallel architecture. The binary cube is one of the most desirable and reliable network-based architecture with high performance, yet many variations have been proposed on the links of the cube to enhance its computational power. The spined cube is a new efficient variant of the binary cube with a smaller diameter and preserves the advantageous properties of the original cube. Grid embedding is accustomed to study simulation capacity of a parallel and distributed computer system and conjointly designs its VLSI layout. In this paper, we propose an algorithm to embed the spined cube into a grid and compute the minimum wirelength over all possible embeddings using the edge congestion technique. (C) 2021 Elsevier B.V. All rights reserved.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据