• 综合
  • 标题
  • 关键词
  • 摘要
  • 学者
  • 期刊-刊名
  • 期刊-ISSN
  • 会议名称
搜索

作者:

Zhang, Wang (Zhang, Wang.) | Wu, Wuchen (Wu, Wuchen.) (学者:吴武臣) | Zuo, Lei (Zuo, Lei.) | Peng, Xiaohong (Peng, Xiaohong.)

收录:

EI Scopus

摘要:

The Network-on-Chip (NoC) has been recognized as a paradigm to solve System-on-Chip (SoC) design challenges, due to many of the problems that will be faced by the designers of multi-billion transistor chips. The infrastructure of network tail determines system performance and cost. The virtual channel buffer depth of tail's input channel is one of the key design problems. The deep input channel buffer depth, which is at each tail in the NoC, increases the overall area of chips. At the same time, depending on the network workload, increasing the buffer size can reduce the network latency by orders of magnitude. This paper analyses the buffer depth of 2-Dimension mesh topology NoC with odd-even routing algorithm based on NoC Interconnect Routing and Application Modeling (NIRGAM) simulator. The analysis results reveal that the optimized input First-In-Fist-Out (FIFO) buffer depth of virtual channel has not relationship with network scale for 2-Dimension mesh topology NoC at Constant Bit Rate (CBR) traffic condition. The optimized input FIFO buffer depth of virtual channel is all 6 for 2X2, 3X3 and 4X4 mesh topology NoC. ©2009 IEEE.

关键词:

Computer graphics Distributed computer systems Indium compounds Integrated circuit design Mesh generation Network-on-chip Programmable logic controllers Routing algorithms Servers System-on-chip Topology

作者机构:

  • [ 1 ] [Zhang, Wang]VLSI and System Lab., Beijing University of Technology, Beijing, China
  • [ 2 ] [Wu, Wuchen]VLSI and System Lab., Beijing University of Technology, Beijing, China
  • [ 3 ] [Zuo, Lei]VLSI and System Lab., Beijing University of Technology, Beijing, China
  • [ 4 ] [Peng, Xiaohong]VLSI and System Lab., Beijing University of Technology, Beijing, China

通讯作者信息:

电子邮件地址:

查看成果更多字段

相关关键词:

相关文章:

来源 :

年份: 2009

语种: 英文

被引次数:

WoS核心集被引频次: 0

SCOPUS被引频次: 2

ESI高被引论文在榜: 0 展开所有

万方被引频次:

中文被引频次:

近30日浏览量: 2

归属院系:

在线人数/总访问数:236/2891250
地址:北京工业大学图书馆(北京市朝阳区平乐园100号 邮编:100124) 联系我们:010-67392185
版权所有:北京工业大学图书馆 站点建设与维护:北京爱琴海乐之技术有限公司