ptolemy.domains.sdf.optimize

Class OptimalScheduleFinder._State

java.lang.Object

ptolemy.domains.sdf.optimize.OptimalScheduleFinder._State

Enclosing class:

OptimalScheduleFinder

protected static class OptimalScheduleFinder._State
extends java.lang.Object

State models a global state of the SDF graph and remembers the actor that was fired to reach it. It consists of two integer array containing the channel content and the actor states (remaining number of firings to complete an iteration) respectively. Moreover, it memorizes the firing to reach the state, which actor fired and whether the firing was exclusive. A state is typically first cloned from its predecessor state and then an actor firing is executed on it. It also maintains a link to its predecessor state. This way the path to reach this state can be reconstructed. It also maintain an optimization 'value' associated with the path leading to this state. This value is used for optimization.

Field Summary

Fields

Modifier and Type	Field and Description
(package private) int[]	actorContent Array to store all actor content, remaining number of firings.
(package private) int[]	channelContent Array to store all channel content.
(package private) OptimalScheduleFinder._Actor	firingActor The actor that was fired to reach this state.
(package private) boolean	firingExclusive true if the firing to reach the state was exclusive.
(package private) OptimalScheduleFinder._State	previousState Link to the previous state.
(package private) int	value Value to be optimized, smaller is better.

Constructor Summary

Constructors

Constructor and Description
_State(OptimalScheduleFinder._State thePreviousState) Construct an instance of State, with a reference to a previous state.

Method Summary

Modifier and Type	Method and Description
(package private) OptimalScheduleFinder._State	clone(OptimalScheduleFinder._State state) Create new state by cloning state 'state' and marking s as its predecessor state.
(package private) int	getFiringsToCompletion() Determine the number of remaining firings.
(package private) boolean	isEndState() Test whether this is a valid end state, i.e. whether all actors have completed their required number of firings.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

firingExclusive

boolean firingExclusive

true if the firing to reach the state was exclusive.

firingActor

OptimalScheduleFinder._Actor firingActor

The actor that was fired to reach this state. remains nil for the initial state.

previousState

OptimalScheduleFinder._State previousState

Link to the previous state.

channelContent

int[] channelContent

Array to store all channel content. Channels and input ports within those channels are given their unique index into this array.

actorContent

int[] actorContent

Array to store all actor content, remaining number of firings. Actors are given their unique index into this array.

value

int value

Value to be optimized, smaller is better.

Constructor Detail

_State

_State(OptimalScheduleFinder._State thePreviousState)

Construct an instance of State, with a reference to a previous state.

Parameters:

thePreviousState - link to previous state

Method Detail

clone

OptimalScheduleFinder._State clone(OptimalScheduleFinder._State state)

Create new state by cloning state 'state' and marking s as its predecessor state.

Parameters:

state - state to clone

Returns:

the new state

isEndState

boolean isEndState()

Test whether this is a valid end state, i.e. whether all actors have completed their required number of firings.

Returns:

true is end state

getFiringsToCompletion

int getFiringsToCompletion()

Determine the number of remaining firings.

Returns:

number of remaining firings