Data-Dependence Graph

entry entry

Z > 1 B1 Z > 1 B1

X=1 X = 2 B4 X=1 X = 2 B4
B2 Z>2 B2 Z>2

Z=X–3 Z=X–3
B3 Y = X + 1 B5 X=4 B3 Y = X + 1 B5 X=4

B6 Z = X + 7 B6 Z = X + 7

exit exit
 Data-Dependence Graph

entry entry

Z > 1 B1 Z > 1 B1

X=1 X = 2 B4 X=1 X = 2 B4
B2 Z>2 X B2 Z>2
X
X
Z=X–3 Z=X–3
B3 Y = X + 1 B5 X=4 B3 Y = X + 1 B5 X=4

X
B6 Z = X + 7 B6 Z = X + 7

exit exit
 Data-Dependence Graph

entry entry
Z
Z
Z > 1 B1 Z > 1 B1

Z Y
X=1 X = 2 B4 X=1 X = 2 B4
B2 Z>2 X B2 Z>2
X
X X
Z=X–3 Z=X–3
B3 Y = X + 1 B5 X=4 B3 Y = X + 1 B5 X=4

X X
Y B6 Z = X + 7 B6 Z = X + 7

Z
exit exit
Control-Dependence Analysis
Intuition
• A statement S1 is
control dependent on a entry
statement S2 if the
outcome of S2 directly
determines whether S1 Z > 1 B1
may not be reached T F
• What are the control
dependences for each X=1 X = 2 B4
statement in the CFG on B2 Z > 2 F
the right?
T Z=X–3
B3 Y = X + 1 B5 X = 4

B6 Z = X + 7

exit
Control-Dependence Analysis
Intuition
• A statement S1 is
control dependent on a entry
statement S2 if the
outcome of S2 directly
determines whether S1 Z > 1 B1
may not be reached T F
• What are the control
dependences for each X=1 X = 2 B4
statement in the CFG on B2 Z > 2 F
the right?
• entry, B1, exit – T Z=X–3
entering code B3 Y = X + 1 B5 X = 4
• B2 – B1T
• B3 – B2T
• B4 – B1F
B6 Z = X + 7
• B5 – B2F, B1F
• B6 – B2F, B1F
exit
Control-Dependence Analysis
Definition
Let G be a CFG, with X and entry
Y nodes in G. Y is
control-dependent on
edge (u->v) if
Z > 1 B1
1. Y post-dominates v and T F
2. If Y != u, then Y does
not post-dominate u X=1 X = 2 B4
B2 Z > 2 F
T Z=X–3
B3 Y = X + 1 B5 X = 4

B6 Z = X + 7

exit
Computing CD Using FOW
Augment the CFG
by adding a node AugCFG En

Start with edge

T
(Start, entry) T
1
F
labeled “T” and 2 4
edge (Start, exit) Start
T
F
labeled “F”; call
3 5
this AugCFG F

Ex
Computing CD Using FOW
Construct the
postdominator AugCFG En

tree for AugCFG

T 1
T F
2 4
Start F
T
3 5
F

Ex
Computing CD Using FOW
Construct the
postdominator Pdom Ex
Tree
tree for AugCFG
Start 2 3 6 1

5 En

4
 Computing CD Using FOW
Consider set S of edges (m, n)
in AugCFG such that n does
not postdominate m AugCFG En
For AugCFG S consists of:
T 1
T F

Pdom 2 4
Ex
Start F
Tree T
3 5
Start 2 3 6 1 F

5 En 6

4 Ex
 Computing CD Using FOW
Consider set S of edges (m, n)
in AugCFG such that n does
not postdominate m AugCFG En
For AugCFG S consists of:
• (Start, En) T 1
• (1,2) T F
• (1,4) 2 4
• (2,3) Start F
• (2,5)
T
3 5
F

Ex
Computing CD Using FOW
Consider, for each edge
(A,B) in S, those
nodes in the Pdom Pdom Ex
tree from B to least Tree
common ancestor L
of A and B Start 2 3 6 1
• Including L if L is A
• Excluding L if L is not A 5 En

4
Computing CD Using FOW
Consider, for each edge
(A,B) in S, those
nodes in the Pdom Pdom Ex
tree from B to least Tree
common ancestor L
of A and B Start 2 3 6 1
• Including L if L is A
• Excluding L if L is not A 5 En

Edge L Nodes
Start, En 4
1, 2
1, 4
2, 3
2, 5
Computing CD Using FOW
Consider, for each edge
(A,B) in S, those
nodes in the Pdom Pdom Ex
tree from B to least Tree
common ancestor L
of A and B Start 2 3 6 1
• Including L if L is A
• Excluding L if L is not A 5 En

Edge L Nodes
Start, En Ex En, 1 4
1, 2 Ex 2
1, 4 Ex 4, 5, 6
2, 3 Ex 3 All identified nodes are
2, 5 Ex 5, 6 control dependent on A
Computing CD Using FOW
Consider, for each edge
(A,B) in S, those
nodes in the Pdom Pdom Ex
tree from B to least Tree
common ancestor L
of A and B Start 2 3 6 1
• Including L if L is A
• Excluding L if L is not A 5 En

Edge L Nodes CD on
Start, En Ex En, 1 Start, T 4
1, 2 Ex 2 1, T
1, 4 Ex 4, 5, 6 1, F
2, 3 Ex 3 2, T All identified nodes are
2, 5 Ex 5, 6 2, F control dependent on A
Computing CD Using FOW
Given (A,B), how can
you find the set of
nodes control Pdom Ex
dependent on A Tree
easily?
Start 2 3 6 1
Why is the set we obtain
correct?
5 En

Edge L Nodes CD on
Start, En Ex En, 1 Start, T 4
1, 2 Ex 2 1, T
1, 4 Ex 4, 5, 6 1, F
2, 3 Ex 3 2, T
2, 5 Ex 5, 6 2, F
 Computing CD Using FOW

Edge L Nodes CD on Create CDG

Start, En Ex En, 1 Start, T
Start
T 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 F
T F
2, 5 Ex 5, 6 2, F 3 F 6 5 4
Program-Dependence Graph
(PDG)
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
A PDG contains nodes representing
statements in P, edges
representing control dependence
between nodes, and edges
representing data dependence
between nodes
Program-Dependence Graph
(PDG)
1. read (n) Compute the PDG for
2. if (n < 0) the
program on the left
3. n=0
4. i := 1
5. sum := 0
6. product := 1
7. while i <= n do
8. sum := sum + i
9. product := product *
i
10. i := i + 1
11. write (sum)
12. write (product)
Program-Dependence Graph
(PDG)
1. read a Compute the PDG for
2. if (a <= 0) { the
3. print a program on the left
4. return
5. }
6. i = a
7. while (i > 0) {
8. a=a+i
9. i=i-1
10. }
11. print a