If you can't read please download the document
Dominator tree
Embed Size (px)
Citation preview
- 1. D(5) = entry; D(1) = exit; D(4) = P(4)UD(4) = D(5)UD(4) =
{5,4}; D(3) = P(3)UD(3) = D(5)UD(3) = {5,3}; D(2) = P(2)UD(2) =
D(3)UD(2) = {5,3,2}; @ Pre(3) D(1) = P(1)UD(1) = D(5)UD(1) = {5,1};
D(2) = P(2)UD(3) = D(1)UD(2) = {5,,3,2}&{5,1}UD(2) @ Pre(1) =
{5,2}
2. Post dominator tree pre order{3,4,5,2,1}
post_order{1,4,5,2,3} Dominator tree pre order{5,1,2,3,4}post order
{1,2,3,4,5} Feedback cut @ 1 5 1 3 2 4 5 1 3 3 2 4 2 3. A Simple,
Fast Dominance Algorithm 4. 1 2 3 6 5 1 2 3 4 6 5 1 2 3 4 6 5 1 2 3
4 5.
- Thus, the total cost per iteration is O(N + E D)
- D is the size ofthe largest Dom set. 6. E is taken over the
entire traversal 7. N is traversing the graph to compute the
reverse postorder sequence
