Capp 07

download Capp 07

of 22

Transcript of Capp 07

  • 7/27/2019 Capp 07

    1/22

    Dataflow Architectures

    Topic 7

  • 7/27/2019 Capp 07

    2/22

    Outline

    What is data flow technique?

    History of Data Flow Machines

    Data Flow Models Types of Data Flow Systems

    Macrodataflow

    Conclusion

  • 7/27/2019 Capp 07

    3/22

    Mechanism on Architectures

    Control flow mechanism-von Neumann

    Data driven or dataflow mechanism Demand flow mechanism

    Pattern flow mechanism

  • 7/27/2019 Capp 07

    4/22

    Data Flow Technique

    Alternative computer design to storedprogram systems.

    The Difference

    Traditional MethodsAn instruction executed when the previous instruction in a

    defined sequence has been executed

    In this model,It is executed when the operands required become

    available.

  • 7/27/2019 Capp 07

    5/22

    History

    developed in 1960s by Karp and Miller as

    graphic means of representing

    computations.

    In the early 1970s Dennis and others

    began to develop computer architectures

    based upon the dataflow computational

    model.

  • 7/27/2019 Capp 07

    6/22

    Data Flow Models

    It uses a directed graph called a data dependence graph ordataflowgraph

    + nodes

    arcs

    The general flow of the model is top to down

    Each node of a data graph can be represented as an activity template

    Add

    [ ]

    [ ]

  • 7/27/2019 Capp 07

    7/22

    Terminology

    Data, data dependency graph

    Arcs :

    flow of the data

    Nodes :

    represent operations

    Token :

    represent data and operations

    Firing:

    when data is available, operation is started,

    then result is sent to destination.

    Handshaking:

    Result is consumed

  • 7/27/2019 Capp 07

    8/22

    Data flow example

  • 7/27/2019 Capp 07

    9/22

    Switch and Merge

    to buildcondional and

    loop operations

  • 7/27/2019 Capp 07

    10/22

    Loop graph

  • 7/27/2019 Capp 07

    11/22

    Static dataflow

    Recursion is not implemented

    Each node can fired only as once instance

    Firing rules

  • 7/27/2019 Capp 07

    12/22

    Firing rules on static

    Nodes fire

    when all input tokens available

    output tokens have been consumed

    input token are then removed

    output tokens are generated

    Acknowledgement

    destination sends back signal to source

    PE when token consumed

  • 7/27/2019 Capp 07

    13/22

    Token format in static

    OPCODE OPERAND DESTINATION

    VALUE DESTINATION

    PE receives operation packes of the form

    PE sends packets of the form

  • 7/27/2019 Capp 07

    14/22

    static dataflow diagram

  • 7/27/2019 Capp 07

    15/22

    Dynamic data flow

    Firing rule

    A node fires all input tokens with the same tag

    appear

    More than one token is allowed on each arc

    Previous output need not be consumed

    before the node fire again

  • 7/27/2019 Capp 07

    16/22

    Dynamic data flow architecture

  • 7/27/2019 Capp 07

    17/22

    Disadvantages of dynamic dataflow

    Overhead involved in matching tokens is heavy

    Resourse allocation is a complicated process

    The long instunction cycle

  • 7/27/2019 Capp 07

    18/22

    Other types of dataflow mechanism

    Pure dataflow

    Macro dataflow

    Hybrid dataflow

  • 7/27/2019 Capp 07

    19/22

    Pure Dataflow machines

  • 7/27/2019 Capp 07

    20/22Hybrid data flow

  • 7/27/2019 Capp 07

    21/22

    MACRO DATA FLOW

  • 7/27/2019 Capp 07

    22/22

    Conclusion

    Dataflow has had a fairly long development

    time, with a few groups studying the

    technique without it becoming widespread in

    commercial use. Compitability with existing system inhibit the

    introduction of a radically different computer

    system requiring a different style of

    programming and different programminglanguages.