基于数字IP的NoC低功耗方法研究

为了设计实现高性能的片上系统SoC,针对基于分层星型连接集成数字IP核的片上网络,提出了低振幅信号发送、基于Mux-Tree的轮转法调度程序,部分激活的交叉单元和串行链路编码等不同的低功耗方法,并分别在每一个开放系统互连层得到应用实现,实验数据证明获得了功耗最经济的片上网络.     

访存与用户行为敏感的MPSoC应用映射

应用映射是MPSoC设计中的关键问题,针对多应用负载的MPSoC,提出一种访存与用户行为敏感的动态映射策略,该策略根据应用的数据访问特征区分热点与非热点应用,并对用户行为进行建模,根据用户行为模型,进一步在运行时区分关键与非关键应用.对每个进入系统的应用,按照应用的热点及关键性分类动态选择在线映射算法,让热点应用围绕存储器布局,非热点应用尽量避免占用存储器附近的资源;对关键应用,最小化应用内通信开销和链路竞争,对非关键应用,最小化应用间通信开销和链路竞争.实验表明,与单纯考虑访存或用户行为的映射策略相比,本文策略能够降低系统整体的通信能耗.     

A variable frequency link for a power-aware network-on-chip (NoC)

Although the technology scaling has enabled designers to integrate a large number of processors onto a single chip realizing chip multi-processor (CMP), problems arising from technology scaling have made power reduction an important design issue. Since interconnection networks dissipate a significant portion of the total system power budget, it is desirable to consider interconnection network's power efficiency when designing CMP. In this paper, we present a variable frequency link for a power-aware interconnection network using the clock boosting mechanism, and apply a dynamic frequency scaling (DFS) policy, that judiciously adjusts link frequency based on link utilization parameter. Experimental result shows that history-based DFS successfully adjusts link frequency to track actual link utilization over time, demonstrating the feasibility of the proposed link as a power-aware interconnection network for system-on-chip (SoC).     

片上网络拓朴优化:在离散平面上布局与布线

微系统芯片(System-on-Chip,SoC)发展到今天,集成密度指数增长和芯片面积的急剧膨胀使得全局连线的延时上升,可靠性下降,成为集成电路的设计瓶颈.片上网络(Network-on-Chip,NoC)是解决整个芯片上数据有效传输的结构之一,以片上网络为基础通信架构的微系统芯片称为片上网上系统芯片(System-on-Network-on-Chip,SoNoC).微系统芯片内通信模式兼有随机性和确定性,应该根据特定应用的通信特征设计片上网络.本文在确定SoNoC设计流程的基础上,根据SoNoC的通信特征,选择了合适的离散平面结构,对SoNoC的运算及控制等模块进行布局、对模块间的通信依赖关系进行布线,发展出FRoD(Floor-plan and Routing on Discrete Plane)算法,以自动生成片上网络的拓扑结构.该算法定义了离散平面的一般表示方法,并在四种典型的离散平面上使用不同规模的随机系统完成了系列实验.为了处理系统和网络之间的耦合关系,逐点分裂的布局算法可以逐步学习和适应系统的通信需求,同时优化系统的执行时间和通信能量,在运行随机任务流图的模拟系统上与随机布局结果相比可以节省30%左右的通信能量,20%左右的系统通信时间.串行、并行和串并混合的布线算法使用最短路径把通信关系分布在离散平面的通道上,使不同的通信关系尽量复用网络通道,与全连接网络相比可以节省10%到30%的面积代价.     

An adaptive congestion-aware routing algorithm based on network load for wireless routers in wireless network-on-chip

Wireless Network-on-Chip (WiNoC) is regarded as one of the promising alternative approaches for sorting out the issues of latency and power consumption in the conventional Network-on-Chip (NoC). Despite the additional bandwidth of wireless channels on a chip, wireless routers (WRs) are prone to congestion in WiNoC due to the limited number of wireless channels on a chip and the shared use of these channels among all the cores. In this paper, an adaptive congestion-aware routing algorithm consistent with traffic load is proposed for solving the congestion problem of WRs. The proposed algorithm selects source-destination pairs with the longest wired hop distance for using wireless channels. The number of selected packets is determined based on the wireless channel bandwidth and the network traffic load. Simulation results show up to 65% latency improvement, 16% wired/wireless link utilization improvement and a saturation throughput increase of approximately 11%.     

Thermal-aware test scheduling using network-on-chip under multiple clock rates

The increasing trend in the number of cores on a single chip has led to scalability and bandwidth issues in bus-based communication. Network-on-chip (NoC) techniques have emerged as a solution that provides a much needed flexibility and scalability in the era of multi-cores. This article presents an optimal integer linear programming (ILP) formulation and a simulated annealing (SA) solution to thermal and power-aware test scheduling of cores in an NoC-based SoC using multiple clock rates. The methods have been implemented and results on various benchmarks are presented.     

片上网络的动态虚通道分配策略

针对片上网络（NoC）的传统的静态虚通道分配不能很好适应非平衡的业务负荷问题,本论文提出了NoC动态虚通道分配策略。在静态虚通道分配基础上,动态分配虚通道通过实时监测节点端口的包流量的方向,决定分配给该端口的虚通道数目。动态虚通道资源可以在所有端口间共享,并根据通信业务需求动态调度。在二维meshNoC上的仿真表明,动态虚通道分配策略不仅节约了存储器资源,而且对NoC传输延时有一定的改善。     

一种基于BFT型拓扑结构片上网络低费用测试方法

随着集成电路制作工艺的进步,多核与众核系统是片上系统的发展趋势.传统的二维网格(2D-mesh)型拓扑结构通信效率低、功耗高和时延长等缺点变得越来越明显.本文首先分析对比了几种常用拓扑结构在多核与众核情况下的性能,进而采用布线复杂度较低、性能较好的蝴蝶型胖树(BFT)拓扑结构来解决片上系统的设计和测试问题.随后,本文针...     

优化NoC性能的存储和通信压缩

文中针对NoC体系结构,提出了两种数据压缩技术,被称为高速缓存压缩和网络接口控制(NIC)内的压缩.性能实测结果指示压缩能够使NoC设计在较低的网络延迟、较低的功耗和改进应用性能等方面获得优势.     

片上网络:新一代的片上系统结构

片上系统是使用共享或专用总线作为芯片的通信资源.由于这些总线具有一定的限制,因此扩展性较差,不能满足发展需求.在这种情况下,目前的片内互连结构将成为多核芯片的发展瓶颈.文章介绍了一种新型的片上体系结构(片上网络)来解决未来片上系统中总线所带来的不足.片上网络作为一种新的片上体系结构,可以解决片上系统设计中所带来的各种挑...     

HIBI Communication Network for System-on-Chip

This paper presents a communication network targeted for complex system-on-chip (SoC) and network-on-chip (NoC) designs. The Heterogeneous IP Block Interconnection (HIBI) aims at maximum efficiency and minimum energy per transmitted bit combined with quality-of-service (QoS) in transfers. Other features include support for hierarchical topologies with several clock domains, flexible scalability, and runtime reconfiguration of network parameters. HIBI is intended for integrating coarse-grain components such as intellectual property (IP) blocks that have size of thousands of gates.HIBI has been implemented in VHDL and SystemC and synthesized on several CMOS technologies and on FPGA. A 32-bit wrapper requires 5400 gates and runs with 315 MHz on 0.18 μ m technology which shows that only minimal area overhead is paid for the advanced features. The area and frequency results are well comparable to other NoC proposals.Furthermore, data transfers are shown to approach the maximum theoretical performance for protocol efficiency. HIBI network is accompanied with a design framework with tools for optimizing the system through automated design space exploration. Erno Salminen Tampere University of Technology (TUT), Finland.Currently he is working towards his PhD degree in the Institute of Digital and Computer Systems (DCS) at TUT. His main research interests are digital systems design and communication issues in SoCs. Tero KangasTampere University of Technology (TUT), Finland.Since 1999 he has been working as a research scientist in the Institute of Digital and Computer Systems (DCS) at TUT. Currently he is working towards his PhD degree and his main research topics are system architectures and SoC design methodologies in multimedia applications. Timo D. H?m?al?ainen Tampere University of Technology (TUT), Finland. He was nominated to full professor at TUT/Institute of Digital and Computer Systems in 2001. He heads the DACI research group that focuses on three main research areas: wireless sensor networks, high-performance multi-DSP and hardware based video encoding, and design flow tools for heterogeneous MP-SoC platforms. Jouni Riihi?mki Tampere University of Technology (TUT), Finland. Currently he is working as a senior design engineer at Nokia Technlogy Platforms. He is also working towards his PhD degree. His research interests include SoC design and verification methodologies. Vesa Lahtinen received his M.Sc. and Ph.D. from TUT in 1998 and 2004, respectively. In TUT, his main research areas were system-on-chips and their interconnects. Currently, Dr. Lahtinen is a Senior Research Engineer in the Computing Architectures Laboratory of Nokia Research Center (NRC) concentrating on architecture modeling and, specifically, memory architectures. Kimmo Kuusilinna Tampere University of Technology (TUT), Finland. His main research interests include system-level design and verification, on-chip interconnections, and parallel memories. Currently he is working as a senior research engineer at the Nokia Research Center.     

ABR业务流量控制的一种智能机制

本文提出了一种ＡＢＲ业务模糊流量控制机制,仿真结果表明它有效地避免了网络阻塞,在网络的吞吐量上要高于强比例速率控制算法（ＥＰＲＣＡ）。该机制与连接业务的特性无关,且不需要改变有关ＡＢＲ闭环反馈的流量控制结构,这为算法的实际应用提供了前提。     

A load balancing bufferless deflection router for network-on-chip

The bufferless router emerges as an interesting option for cost-efficient in network-on-chip (NoC) design. However, the bufferless router only works well under low network load because deflection more easily occurs as the injection rate increases. In this paper, we propose a load balancing bufferless deflection router (LBBDR) for NoC that relieves the effect of deflection in bufferless NoC. The proposed LBBDR employs a balance toggle identifier in the source router to control the initial routing direction of X or Y for a flit in the network. Based on this mechanism, the flit is routed according to XY or YX routing in the network afterward. When two or more flits contend the same one desired output port a priority policy called nearer-first is used to address output ports allocation contention. Simulation results show that the proposed LBBDR yields an improvement of routing performance over the reported bufferless routing in the flit deflection rate, average packet latency and throughput by up to 13%, 10% and 6% respectively. The layout area and power consumption compared with the reported schemes are 12% and 7% less respectively.     

NoC系统设计的研究

片上网络研究涉及从物理设计到体系结构、系统应用、设计方法和工具等诸多方面.文中从系统结构的角度总结了片上网络设计的一些主要研究内容和NoC技术研究发展方向.     

基于内建自测技术的Mesh结构NoC无虚通道容错路由算法

 在Zhang's算法绕行思想的基础上,提出了一种2D-Mesh结构片上网络无虚通道容错路由算法,用于解决多故障节点情况下片上网络的无虚通道容错路由问题.算法利用内建自测试机制获取故障区域的位置信息,通过优化绕行策略来均衡故障区域周围链路的负载并减少部分数据的绕行距离.针对8×8的2D-Mesh网络的仿真表明,与Chen's算法相比,在故障区域大小为2×2,网络时延为70 cycles的情况下,随着故障区域位置的变化所提算法可提高1.2%到4.8%的网络注入率.且随着故障区域面积的扩大,所提算法在减少通信时延,提高网络吞吐量方面的作用更为明显.     

A low overhead load balancing router for network-on-chip

The design of a router in a network-on-chip (NoC) system has an important impact on some performance criteria. In this paper, we propose a low overhead load balancing router (LOLBR) for 2D mesh NoC to enhance routing performance criteria with low hardware overhead. The proposed LOLBR employs a balance toggle identifier to control the initial routing direction of X or Y for flit injection. The simplified demultiplexers and multiplexers are used to handle output ports allocation and contention, which provide a guarantee of deadlock avoidance. Simulation results show that the proposed LOLBR yields an improvement of routing performance over the reported routing schemes in average packet latency by 26.5%. The layout area and power consumption of the network compared with the reported routing schemes are 15.3% and 11.6% less respectively.     

一种片上网络路由测试方法研究

对片上网络路由器的结构进行了分析,建立了相应的故障模型.针对此故障模型结合内建自测试,提出了一种基于量子遗传算法的测试矢量传递路径寻优方法.该算法具有收敛速度快,精度高等优点.最后通过对测试故障覆盖率和测试时间进行分析表明这种测试方法具有较高的故障覆盖率、较少的测试时间.     

谱分析方法的混合电路BIST

SoC芯片内对于混合信号电路测试有着举足轻重的作用.本文介绍了一种通过谱密度分析方法的混合电路内建自测试.此方法通过使用噪声源与比较器数字量化得到被测信号的频谱特性.它的主要特点是电路简单、抗干扰性能强和多点插入多路并行采集,不需要多位AD转换器和多路选择开关.此方法基本上是全数字式的,采用一位量化,数据处理速度快,能满足给定条件下的实时处理要求;并可利用系统内已有的资源,适应于SoC环境.本文给出了系统实现的详细结构和一个测试锁相环电路的测试仿真实例,验证了谱分析方法的测试有效性.     

基于遗传算法的NoC互连测试研究

针对互连测试难题的分析,提出一种基于遗传算法的NoC互连测试方案。该方案采用NoC重用测试机制的方法,在功耗限制条件下,选取合适的测试端口和最短测试路径,同时根据互连测试中实际存在的问题,对算法进行适当改进,建立基于遗传算法的NoC互连测试模型,旨在获取最优矢量集的同时,测试代价更小。当NoC的规模达到一定程度时,采用划分测试方法,缩短测试路径,降低测试时间,提高测试效率。以SoCIN结构电路为仿真平台,分别对不同规模的NoC进行实验仿真。实验结果表明,遗传算法能快速有效地收敛到最优解,在测试运行代数及测试生成时间上取得了良好的测试效果。     

Hybrid one-cycle control scheme for fault-tolerant modular multilevel rectifiers

This paper customises the classic one-cycle control (OCC) scheme for modular multilevel rectifiers (MMRs) and overcomes the inherent defect of the OCC. To be specific, a hybrid one-cycle control scheme is proposed combining the OCC and the virtual loop mapping (VLM). First, on the basis of the classic OCC and the volt-second equivalence principle, the relationships between the MMR arm equivalent duty cycle and the sub-module (SM) duty cycle are derived. Then, by making use of VLM methods, the switching frequency of the SM is reduced and the dynamic capacitor voltage balance is obtained as well. Further, to achieve the single-line-to-ground fault-tolerance capability and eliminate the second harmonic ripple in the DC voltage, the constant power control is presented using the negative-sequence voltage compensation. Specifically, the whole control scheme only needs one proportional integration controller, which greatly reduces the complexity of system control and the system cost. The validity of the proposed scheme is verified through simulation and experimental studies.