项目名称: 基于柏拉图立体多级裂变模型的三维片上网络拓扑结构的研究
项目编号: No.61272006
项目类型: 面上项目
立项/批准年度: 2013
项目学科: 自动化技术、计算机技术
项目作者: 张大坤
作者单位: 天津工业大学
项目金额: 61万元
中文摘要: 基尔比发明了世界上第一个集成电路,改变了整个世界,也因此获得了诺贝尔奖。随着半导体技术的发展,片上系统和片上网络相继问世,然而,受二维布局条件限制难以保证关键部件相邻以减小信号延迟,三维片上网络应运而生,已成为主流研究方向。目前三维片上网络拓扑结构主要是二维拓扑的堆叠,而直接从三维空间进行研究很有必要,具有代表性的是超立方体三维片上网络拓扑结构,但它不具有可扩展性且不易构成多维超立方。柏拉图立体历史悠久,应用广泛,当今又成为新的研究热点;研究基于柏拉图立体的三维片上网络拓扑结构具有探索性,而5个柏拉图立体的节点度不尽相同(最高5),且可扩展性差。因此,本项目提出柏拉图立体多级裂变模型(同步/异步两种,且具普适性),旨在用其研究三维片上网络拓扑结构,该拓扑结构具有规则性、低节点度(恒为3)、对剖宽度不变、空间旋转对称性等优点,并使网络扩展从"加法"转为"裂变",选题具有创新性(见查新报告)。
中文关键词: 三维片上网络;拓扑结构;柏拉图立体;多级裂变模型;性能仿真
英文摘要: Jack Kilby as the inventor of the first integrated circuit was awarded the Nobel Prize and with this invention changed the world. The world has since then witnessed the creations of System-on-Chip (SoC) and Network-on-Chip (NoC) with the development of the advanced semiconductor technology. However the signal latency constraint requirement between the key components cannot be easily fulfilled due to the 2D layout. So 3D NoC was proposed to solve the problem and has become the mainstream research trend. Currently the most popular topological structure of 3D NoCs is constructed by stacking multiple layers of 2D NoC. However the study of other new 3D topologies based on true 3D space structures is very necessary and typical among them is hypercube. But lack of extendibility and difficulty in making multi-dimensional hypercube make it less favourable. The study of Platonic solids has a long history and extensive application and has become a new hot research area including the exploring research of 3D NoC based on Platonic solids. But the five Platonic solids have different degree (with highest as 5) and poor extendibility. Therefore, we propose in this project a Platonic multi-class fussion model (including both synchronized and asynchronized) that can be applied more generally. The aim is to adopt it in the study o
英文关键词: 3D NoC;Topology structure;Platonic solids;Multi-stage fission model;Performance simulation