Class 4

• Basic Analyses (4)

• Assign (see Schedule for links)
• Readings
• Control/program-dependence analysis
• Static single assignment and control dependence
• Problem Set 2: due 9/1/09

Control-dependence Analysis

2
Control-dependence Analysis

1. Introduction (motivation, overview)

2. Computation of control-dependence using
FOW
3. Computation of control-dependence using
dominance frontiers (later)

Introduction
Intuition
y A statement S1 is control CFG entry
dependent on a statement S2
if the outcome of S2
determines whether S1 is Z > 1 B1
reached in the CFG T F
y What are the control
dependences for each X=1 X = 2 B4
statement in the CFG at right? B2 Z > 2 F
y B2 – B1T
y B3 – B2T T Z=X–3
y entry, B1, exit – entering B3 Y = X + 1 B5 X = 4
code
y B4 – B1F
y B5 – B2F, B1F
y B6 – B2F, B1F
B6 Z = X + 7

exit
Introduction
Intuition
y A statement S1 is control CFG entry
dependent on a statement S2
if the outcome of S2
determines whether S1 is Z > 1 B1
reached in the CFG T F
y What are the control
dependences for each X=1 X = 2 B4
statement in the CFG at right? B2 Z > 2 F
y entry, B1, exit – entering
code T Z=X–3
y B2 – B1T
B3 Y = X + 1 B5 X = 4
y B3 – B2T
y B4 – B1F
y B5 – B2F, B1F
y B6 – B2F, B1F
B6 Z = X + 7

exit

Introduction
Definition (formal)
1. Let G be a CFG, with X and Y CFG entry
nodes in G. Y is control-
dependent on X iff
1. There exists a directed path Z > 1 B1
P from X to Y with any Z in P T F
(excluding X and Y)
postdominated by Y and X=1 X = 2 B4
2. X is not postdominated by Y B2 Z > 2 F
Definition (informal) T Z=X–3
There are two edges out of Y
y traversing one edge always
B3 Y = X + 1 B5 X = 4
leads to X,
y traversing the other edge the
other may not lead to X B6 Z = X + 7

exit
Introduction
Definition (formal)
1. Let G be a CFG, with X and Y CFG entry
nodes in G. Y is control-
dependent on X iff
1. There exists a directed path Z > 1 B1
P from X to Y with any Z in P T F
(excluding X and Y)
postdominated by Y and X=1 X = 2 B4
2. X is not postdominated by Y B2 Z > 2 F
Definition (informal) T Z=X–3
There are two edges out of X
y traversing one edge always
B3 Y = X + 1 B5 X = 4
leads to Y,
y traversing the other edge the
other may not lead to Y B6 Z = X + 7

exit

Computing Control-dependence Using

FOW

1. Augment the CFG by

CFG En
adding a node Start with
edge (Start, entry)
labeled “T” and edge 1
T F
(Start, exit) labeled “F”;
call this AugCFG 2 4
F
T
3 5

Ex
Computing Control-dependence Using
FOW

1. Augment the CFG by

AugCFG En
adding a node Start with
edge (Start, entry)
labeled “T” and edge T 1
T F
(Start, exit) labeled “F”;
call this AugCFG 2 4
Start T F

3 5
F

Ex

Computing Control-dependence Using

FOW

2. Construct the
AugCFG En
postdominator tree for
AugCFG
T 1
T F
2 4
Start T F

3 5
F

Ex
Computing Control-dependence Using
FOW

2. Construct the
postdominator tree for Pdom Ex
AugCFG Tree

Start 2 3 6 1

5 En

Computing Control-dependence Using

FOW
3. Consider edges in AugCFG
that are labeled (i.e., those Pdom Ex
nodes on which another node Tree
might be control dependent);
call this set S 2 3 6 1
Start
4. For AugCFG S consists of
(Start, En), (1,2), (1,4), (2,3),
(2,5) (i.e., those edges (A,B) 5 En
in the AugCFG for which B is
not an ancestor of A in Pdom
4
tree)
Computing Control-dependence Using
FOW

5. Consider each edge

(A,B) in S, those nodes Pdom Ex
in the Pdom tree from B Tree
to least common
Start 2 3 6 1
ancestor L of A and B
y Including L if L is A
y Excluding L if L is not A 5 En

Computing Control-dependence Using

FOW

Pdom Ex
Edge L Nodes CD on Tree
Start, En
Start 2 3 6 1
1, 2
1, 4 5 En
2, 3
4
2, 5
Computing Control-dependence Using
FOW

Pdom Ex
Edge L Nodes CD on Tree
Start, En Ex En, 1 Start, T
Start 2 3 6 1
1, 2 Ex 2 1, T
1, 4 Ex 4, 5, 6 1, F 5 En
2, 3 Ex 3 2, T
4
2, 5 Ex 5, 6 2, F

Computing Control-dependence Using

FOW

6. Create control-
Edge L Nodes CD on dependence graph
Start, En Ex En, 1 Start, T
T
Start
T
1, 2 Ex 2 1, T
En 1
1, 4 Ex 4, 5, 6 1, F T
F
2, 3 Ex 3 2, T 2 F
T F
F
2, 5 Ex 5, 6 2, F
3 F 6 5 4
Computing Control-dependence Using
FOW

7. Add region nodes to

CDG Start
T
R1

En 1
T F

R2 R3

2 6 5 4
T F
R4 R5

Computing Control-dependence Using

FOW

7. Add region nodes to 8. Create new regions if

CDG Start necessary
Start
T T
R1 R1
En 1 1
T F En
T F
R2 R3 R2 R3

2 6 5 4 4
T F 2 R6
T F
R4 R5 R4 R5 5
6
3 3
Computing Control-dependence Using
FOW

7. Merge region nodes if 8. Create new regions if

possible Start necessary
Start
T T
R1 R1

En 1 En 1
T F T F
R2 R3 R2 R3

2 F R6 4 R6 4
T
2
T F
R4 5 R4 R5 5
6 6
3 3

Program-dependence Graph

y A program dependence graph (PDG) for a

program P is the combination of the control-
dependence graph for P and the data-
dependence graph for P
y A PDG contains nodes representing statements
in P, edges representing control dependence
between nodes, and edges representing data
dependence between nodes
Create PDG for Program

T
Start
T

En 1
T
F
2 F F
T F
3 F 6 5 4

Program-dependence Graph

En Completed in class

1
2 3
4 5
6
8
7
Ex

22
Program-dependence Graph

1. A=1
B1 2. B=2 Completed in class

B2 3. C=A+B
4. D=C-A

5. D = B * D B3
T B5
F
6. D=A+B 8. B=A+B
7. E=E+1 9. E=C-A
F
T
B4
10. A = B * D
B6 11. B = A - D