Evaluation of Bus Based Interconnect Mechanisms in Clustered VLIW Architectures

Gangwar, Anup; Balakrishnan, M.; Panda, Preeti; Kumar, Anshul
December 2007
International Journal of Parallel Programming;Dec2007, Vol. 35 Issue 6, p507
Academic Journal
With new sophisticated compiler technology, it is possible to schedule distant instructions efficiently. As a consequence, the amount of exploitable instruction level parallelism (ILP) in applications has gone up considerably. However, monolithic register file VLIW architectures present scalability problems due to a centralized register file which is far slower than the functional units (FU). Clustered VLIW architectures, with a subset of FUs connected to any RF provide an attractive solution to address this issue. Recent studies with a wide variety of inter-cluster interconnection mechanisms have reported substantial gains in performance (number of cycles) over the most studied RF-to-RF type interconnections. However, these studies have compared only one or two design points in the RF-to-RF interconnects design space. In this paper, we extend the previous reported work. We consider both multi-cycle and pipelined buses. To obtain realistic bus latencies, we synthesized the various architectures and calculated post-layout clock periods. The results demonstrate that while there is less that 10% variation in interconnect area, the bus based architectures are slower by as much as 400%. Also, neither multi-cycle or pipelined buses nor increasing the number of buses itself is able to achieve performance comparable to point-to-point type interconnects.


Related Articles

  • EMBEDDED SYSTEMS.  // EDN;9/1/2005, Vol. 50 Issue 18, p98 

    This article presents two new embedded systems. With a point-to-point mesh topology that requires no switch cards, the two-slot Advanced-Telecom Computing Architecture (ATCA) backplane from Elma Bustronic features a 10-layer stripline design for routing through signal-integrity studies. ATCA...

  • Successful SoCs. Wong, William // Electronic Design;2/4/2002, Vol. 50 Issue 3, p57 

    Focuses on developments related to system-on-a-chip (SoC) central processing unit (CPU) architectures as of February 4, 2002. Factors limiting the popularity of Java processors in the embedded CPU market; Real-time Java standards' impact on construction of Java real-time operating systems.

  • Fiber for Everyone. Devereaux-Weber, David // Communications Technology;Sep2006, Vol. 23 Issue 9, p58 

    A letter is presented in response to a question for readers asked in "CT's Pipeline" about how deep fiber needs to go.

  • Getting advanced networks off the ground. Kim, Gary // America's Network;7/1/94, Vol. 98 Issue 13, p28 

    Focuses on the tendency of providers of advanced services to disregard interoperability in their desire to establish new services architecture quickly. Comments from industry experts; Impairment of seamless communications by the cacophony of operating systems choices; Protection against...

  • Future of the 3745 challenges IBM and its customers. Passmore, David // Business Communications Review;Apr94, Vol. 24 Issue 4, p14 

    Features the 3745 primary network modal processor for enterprise SNA networks. Representation of IBM's source of networking revenue; Long term viability of the 3745; Internal evaluations of evolutionary paths; Discussion of potential options with its customers.

  • Developing an architecture. Passmore, David // Business Communications Review;Mar1997, Vol. 27 Issue 3, p18 

    Focuses on the development of network architecture while focusing on the templates which can be applied to particular network environments. Identification of the most important element of an architecture; Listing of technology positions which are needed to address major issues; Presentation of...

  • New developments in CompactPCI. Pavlat, Joe // Electronic Design;11/24/97, Vol. 45 Issue 26, p314 

    Focuses on the computer network CompactPCI. Features of the network; Functions of the network; Description of the network.

  • Here comes the PCI buses. Alderman, Ray // Electronic Design;11/24/97, Vol. 45 Issue 26, p315 

    Focuses on the computer network architecture, PCI bus. Features of the bus; Description of the bus.

  • Latency is the problem. Alderman, Ray // Electronic Design;11/24/97, Vol. 45 Issue 26, p340 

    States that latency is the main problem with all the innovations in computer architecture. Information on latency; Highlights of latency problems and how to deal with them.


Read the Article


Sorry, but this item is not currently available from your library.

Try another library?
Sign out of this library

Other Topics