首页> 外文会议>International Symposium on Embedded Multicore/Many-core Systems-on-Chip >Adaptive VC Organization and Arbitration for Efficient NoC Design
【24h】

Adaptive VC Organization and Arbitration for Efficient NoC Design

机译:高效的NoC设计的自适应VC组织和仲裁

获取原文

摘要

Network-on-Chip (NoC) has emerged as one of the main communication structure suitable for the interconnection of processing and other IP cores of a system-on-chip (SoC). An NoC typically utilizes virtual channels (VCs) to improve wormhole routing among the SoC cores by enabling multiple data packets to share a communication channel and to avoid deadlocks. Dynamically allocation multi queues based VC organization has higher buffer utilization but it also has some problems such as complexity, setup limitation, etc. Arbitration is also an important part of NoC routers. Past arbitration techniques have some problems that are related to lower speed, fairness, and router pipelining. We present a novel NoC router (RDQ-IRR-v2) architecture that incorporates some solutions to all of these problems. The router design utilizes adaptive VC buffering and index based arbitration techniques. The experiments and simulations indicate the efficiency of our proposed NoC router in terms of power consumption, chip area, frequency of router as well as latency and throughput related performance metrics in different NoC configurations and for various traffic patterns.
机译:片上网络(NoC)已经成为适用于处理系统和片上系统(SoC)的其他IP内核互连的主要通信结构之一。 NoC通常利用虚拟通道(VC)通过使多个数据包共享一个通信通道并避免死锁来改善SoC内核之间的虫洞路由。基于动态分配的多队列VC组织具有较高的缓冲区利用率,但同时也存在诸如复杂性,设置限制等问题。仲裁也是NoC路由器的重要组成部分。过去的仲裁技术存在一些与较低速度,公平性和路由器流水线相关的问题。我们提出了一种新颖的NoC路由器(RDQ-IRR-v2)架构,该架构结合了针对所有这些问题的一些解决方案。路由器设计利用自适应VC缓冲和基于索引的仲裁技术。实验和仿真表明,在不同的NoC配置和各种流量模式下,我们提出的NoC路由器在功耗,芯片面积,路由器频率以及与延迟和吞吐量相关的性能指标方面均具有很高的效率。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号