Dfg & sg ppt (1)
- 1. DATA FLOW and SEQUENCING GRAPH (Synthesis and Optimization of Digital Circuits) Submitted to Submitted by Mrs. Nivedita Shruti Shreya Apoorva
- 2. Data Flow Graph Graphical representation of the "flow" of data through an information system. Represents a data dependencies between a number of operations. A data-flow graph G(V, E) is a directed graph whose vertex set V = {vi; i = 1,2, . . . , n,) is in one-to-one correspondence with the set of tasks. Assumes that the values are preserved during their lifetime, i.e., that some sort of storage (registers in case of synchronous implementations) is provided.
- 3. Example Consider the following model fragment describing a set of computations XI = x + dx: ul = u-(3*x*u*dx)-(3*y*dx): yl = y+u*dx; c = xi < a; Four assignments → a set of 11 simple operations 11 vertices are labeled by numeric identifiers Input = {x, y, u, dx, a, 3) Output = {XI, yl, ul, c}
- 5. Reasons for existence of Data Flow Graph Increasing demand for higher computation speed Ability to specify algorithms For this reason Data Flow Graphs have been used successfully in the simulation of hardware like computer systems. Advantages of Data Flow Graphs Compactness and general amenability to direct interpretation. The algorithms are controlled by the arrival of data at the nodes.
- 7. A hierarchical control/data flow graph Control flow primitives modelled through hierarchy Data-flow & serialization dependencies modelled by graphs Supports a model call Hierarchy of directed graphs represented by Gs(V,E)
- 8. Generic element in the hierarchy An extended data flow graph having 2 kinds of vertices Operations Links Example of a Sequencing Graph
- 9. 2 submodels Submodel(a,z){x=a*b; y=x*c; z=a+b} Submodel(m,n){p=m+n; q=m*n} Vertex a,4 Models the call to submodel Links the two graph entities Example of a Hierarchical Sequencing Graph
- 10. Some vertices are links to other sequencing graph entities representing Model call Branching Iteration Constructs Model call Vertex is a pointer to another sequencing graph entity at a lower level in the hierarchy
- 11. Can be modelled by Branching Clause Branching Bodies Set of tasks selected according to the value of branching clause Execution is mutually exclusive Branching is modelled by associating a sequencing graph entity with each branching body and a link vertex with the branching clause
- 12. Code sequence : x=a*b; y=x*c; z=a+b; if (z≥0) {p=m+n; q=m*n} Branching clause z≥0 is calculated at a,4 TRUE: Required operations take place FALSE: No-operations Example of a Hierarchical Sequencing Graph with a Branching Construct
- 13. Are modelled by Iteration Clause Iteration Body/Loop A set of tasks repeated as long as the iteration clause is true The adjacent circuit performs 3 tasks: Reading input data Iterating the evaluation of a set of statements Writing the result to a port Vertex labeled LOOP evaluates the iteration clause c Example of a Hierarchical Sequencing Graph with an Iteration Construct
- 14. Marking denotes the state of the corresponding operation which can be Waiting for execution Executing Having completed execution Firing an operation means starting its execution Semantics of the model : an operation can be fired as soon as all of its direct predecessors have completed execution
- 15. Attributes such as area or delay cost can be attributed to the vertices and edges of a sequencing graph model Delay can be Data dependent Data independent
- 16. Operations with such delay are Data-dependent delay branching Data-dependent delay iteration Can be of two types Bounded Unbounded
- 17. Can be estimated before synthesis Can be characterised by its overall delay called Latency Graphs with bounded delays are called bounded-latency graphs Else they are called unbounded-latency graphs
- 18. Used to model circuit behaviour at architectural level Can be cast as directed bipartite graphs Vertices correspond to Places (States) Transitions (Operations) Petri net model is more general and powerful Sets framework to state properties of general concurrent circuits