7.7 PARALLEL GRAPH ALGORITHMS 211
take two cycles. The last parallel block takes one comparison to see if the new
node is in the processors set and then N / p updates of closest and dis-
tance. In summary, the main processing loop has 2(N / p) + p + 1 instruc-
tions and is executed N 1 times, giving (N 1) * (2(N / p)+p + 1). This
works out to an O(N^2 / p) run time and a cost of p*O(N^2 / p), or O(N^2 ). As
we saw in the other cases, to achieve optimality, the number of processors
should be about N / lg N.
7.7.3
- Give the weighted adjacency matrix for the following four graphs, and cal-
culateA^2 ,A^4 , and A^8.
■7.7.3 EXERCISES
A B C
F G H
D E
4
11
1
2
2
2 2
5
4
4
4
4
2
2
2
3
3 3
35
3 66
a. b. A B
DE
FG
C
A B C
F G H
D E
45
4
1
22
5 24
1
1
4
2
1
3
2
2
3
5
5
2 23
c. d. A B
DE
G H
C
F