Multistage bus network (MBN): an interconnection network for cache coherent multiprocessors

نویسندگان

  • Laxmi N. Bhuyan
  • Ashwini K. Nanda
چکیده

Single bus multiprocessor systems do not scale well due to the limited bandwidth of the bus. Hierarchical bus interconnections are scalable, but unfortunately the top level bus becomes a bottleneck for larger systems. In this paper we present a novel interconnection network called the Multistage Bus Network(MBN). The MBN consists of multiple stages of buses and it preserves the bandwidth properties of a conventional Multistage Interconnection Network(M1N). At the same time the MBN also retains the economic design and fast broadcast advantages of the bus based systems. We show that the MBN performs almost as good as a MIN and is much more cost effective than the MIN. We also design and study a multistage snooping cache protocol for the MBN. 1 Introduction The efficiency of a large multiprocessor system depends to a great extent on the underlying interconnection network. When the interconnection network becomes more complex, the memory access time increases and severely degrades the system performance. A large amount of research has been done on the design of efficient interconnection networks for such multiproces-sors. Among various interconnection networks, the hierarchical buses[l2, 131 and Multistage Interconnection Networks(MINs)[3, 8, 101 show a good promise for use in future. These networks usually operate under synchronous packet switched environments[2,3]. The processor submits a memory request and waits till it is satisfied before submitting another request. In a strictly hierarchical system[l2, 131, there are a number of buses connected in the form of a tree between the processors and the memories. When the system size grows, the buses in the upper levels, especially the top level bus, form the system bottleneck. Thus, the scalability of a hierarchical bus system becomes limited by the bandwidth of the top level bus. However, the simplicity of the bus based designs and the availability of a fast broadcasting mechanism are factors that make these interconnections attractive. Moreover, bus based networks efficiently support cache coherence protocols due to their snooping capabilities[l, 5, 121. On the other hand, the MINs offer uniform band-width across all stages of the network. The bandwidth of the network increases proportionally with the increase in system size, making the MIN a highly scal-able interconnection. Several commercial machines and research projects are based on such MINs[2, 6, 111. The switches in a MIN are made up of small crossbar switches. When the system size grows, bigger switches can be used t o keep the number …

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Analysis of Interconnection Networks for Cache Coherent Multiprocessors with Scientific Applications

Interconnection networks, such as, shared bus and multistage interconnection networks (MINs) are very suitable for the design of shared memory multiprocessors. The existing analytical models of these networks are based on unrealistic synthetic workload for simplicity of the analyses. Also, they consider the networks in isolation without incorporating other architectural details of a multiproces...

متن کامل

The MINC (Multistage Interconnection Network with Cache Control Mechanism) Chip

Although bus connected multiprocessors have been widely used as high-end workstations or servers, the number of connected processors is strictly limited by the maximum bandwidth of the shared bus. Instead of them, a switch connected multiprocessor which uses a crossbar or Multistage Interconnection Networks(MINs) for connecting processors and memory modules is a hopeful candidate. However, in s...

متن کامل

Execution Based Evaluation of Multistage Interconnection Networks for Cache-Coherent Multiprocessors

In this paper, performance of multistage interconnection network with wormhole routing and packet switching has been evaluated for cache-coherent shared-memory multiprocessors. The traac in cache-coherent systems is characterized by traac bursts, one-to-many and many-to-one traac, and small xed length messages. The evaluation is based on execution-driven simulation using various applications. T...

متن کامل

Performance of Multistage Bus Networks for a Distributed Shared Memory Multiprocessor

A Multistage Bus Network (MBN) is proposed in this paper to overcome some of the shortcomings of the conventional multistage interconnection networks (MINs), single bus and hierarchical bus interconnection networks. The MBN consists of multiple stages of buses connected in a manner similar to the MINs and has the same bandwidth at each stage. A switch in an MBN is similar to that in a MIN switc...

متن کامل

Impact of Switch Design on the Application Performance of Cache-Coherent Multiprocessors

In this paper, the effect of switch design on the application performance of cache-coherent non-uniform memory access (CC-NUMA) multiprocessors is studied in detail. Wormhole routing and cut-through switching are evaluated for these shared-memory multiprocessors that employ multistage interconnection network (MIN) and full map directory-based cache coherence protocol. The switch design also con...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 1991