HW11_sol - number for i . Each processor has local...

Info iconThis preview shows pages 1–2. Sign up to view the full content.

View Full Document Right Arrow Icon
Fall Semester 2008 CS300 Algorithms Solution #11 1. Comment: Each processor carries out the algorithm using its own number for i . Each processor has local variables sum , temp , and incr . Assume that all memory cells in a PRAM that do not contain input contain zero. Now assume that K is d log n e . write M[i] into Sum[i]; incr := 1; for step := 1 to K do read Sum[i] into sum; read Sum[i+incr] into temp; sum := sum+temp; incr := 2*incr; write sum into Sum[i]; end{for}. // Sum[1] will be the sum of M[i]s. if( M[i] < Sum[1]/n ) write 1 into M[i]; else write 0 into M[i]; incr := 1; for step := 1 to K do read M[i] into sum; read M[i+incr] into temp; sum := sum+temp; incr := 2*incr; write sum into M[i]; end{for}.// then M[1] will be the number of students below average. 2. Description: C = A × B Input: A , B , C : n × n boolean matrices M [1] , ··· ,M [ n 2 ] A (store row-wise) M [ n 2 + 1] , ··· ,M [2 n 2 ] B (store column-wise) Output: C M [1] , ··· M [ n 2 ] Comment and Algorithm: Each processor carries out the algorithm using its own
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Background image of page 2
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: number for i . Each processor has local variables a , b , row , col , index and tmp . Assume that all memory cells in a PRAM that do not contain input contain zero. row := d i/n 2 e ; col := d i/n e - b i/n 2 c × n ; index := i- b i/n c × n ; 1 read M[(row-1) × n + index] into a ; read M[ n 2 + (col-1) × n + index] into b ; tmp := a ∧ b ; if tmp == 1 then write 1 in M[(row-1) × n + col] . • the number of processors = n 3 • the number of steps = 2 (reads) 3. Each i th processor reads v i and w i and compare them. If v i ≥ w i , then it writes 1 into i th memory cell. Otherwise, it writes 0. This computation takes O (1) time since it can be done in parallel using n processors. Then, by Binary Fan-in technique, the sum of the values in each memory cell can be computed in O (log n ) time. If the sum is larger than b n/ 2 c , V is stronger than W . 2...
View Full Document

Page1 / 2

HW11_sol - number for i . Each processor has local...

This preview shows document pages 1 - 2. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online