Package | Description |
---|---|
org.jikesrvm.compilers.opt.ssa |
SSA implementation (disabled because of bugs, see entry in issue tracker).
|
Modifier and Type | Field and Description |
---|---|
private ArrayList<GVCongruenceClass> |
GlobalValueNumberState.B
ArrayList of GVCongruenceClass, indexed by value number.
|
private Stack<GVCongruenceClass> |
GlobalValueNumberState.workList
Stack used for a work list.
|
Modifier and Type | Method and Description |
---|---|
(package private) GVCongruenceClass |
GlobalValueNumberState.congruenceClass(Object name) |
private GVCongruenceClass |
GlobalValueNumberState.createCongruenceClass(Object label)
Given a label, return a new congruence class for that label.
|
private GVCongruenceClass |
GlobalValueNumberState.findOrCreateCongruenceClass(Object label,
HashMap<Object,GVCongruenceClass> labelMap)
Given a label, return the congruence class for that label.
|
Modifier and Type | Method and Description |
---|---|
private void |
GlobalValueNumberState.addDependentClassesToWorklist(GVCongruenceClass c)
Assuming congruence class c has changed: find all other classes
that might be affected, and add them to the worklist
|
private boolean |
GlobalValueNumberState.checkCongruence(ValueGraphVertex v,
GVCongruenceClass c)
Does the current state of the algorithm optimistically assume
that a vertex v is congruent to the vertices in a congruence
class?
|
private void |
GlobalValueNumberState.partitionClass(GVCongruenceClass partition)
Partition a congruence class.
|
Modifier and Type | Method and Description |
---|---|
private int |
GlobalValueNumberState.findCongruenceMatch(ArrayList<GVCongruenceClass> vector,
ValueGraphVertex v)
Does vertex v belong to any congruence class in a vector?
|
private GVCongruenceClass |
GlobalValueNumberState.findOrCreateCongruenceClass(Object label,
HashMap<Object,GVCongruenceClass> labelMap)
Given a label, return the congruence class for that label.
|