{[ promptMessage ]}

Bookmark it

{[ promptMessage ]}

scan0018

# scan0018 - SDSU Fall 2006 CS—662 Parallel Algorithms...

This preview shows page 1. Sign up to view the full content.

This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: SDSU Fall 2006 CS—662 Parallel Algorithms PRACTICE EXAM #1 (100 points; problems are equally weighted) PROBLEM #1 Indicate whether each statement below is TRUE or FALSE by circling the appropriate letter F If a problem can be solved in} Q(log N) time by a PRAM with if BREW processors, then the same problem can be solved in 0(N log N) time by a PRAM with N common CRCW processors. _ ./ (/ T i; F The largest element in any unsorted array of N nonnegative integers can be “ determined in O( 1 ) time by a PRAM with N2 common CRCW processors. T {F I The largest element in any unsorted array of N2 nonnegative integers can be determined in O(1 NN) time by a PRAM with N common CRCW 0g processors and 0(N2) auxiliary memory. ‘ :T F If a problem can be solved in 0( log log N) time by a PRAM with N BREW ' processors, then the same problem can be solved in 0( log log N ) time by a PRAM with N common CRCW processors. T , F The largest element in an unsorted array of N nonnegative integers, each of which is strictly less than N 3 , can be determined in 0(1) time by a PRAM with N common CRCW processors. " T: F N nonnegative integers can be summed in 0(log N) time by a PRAM with N EREW processors. \. 1T F N nonnegative integers can be summed in 0(10g N) time by a PRAM with N log N EREW processors. \ { {I F N nonnegative integers can be summed in 0(10g N) time by a PRAM with N CREW processors. log N T; F The second—largest element in any unsorted array of N nonnegative integers can be determined in 0(loglog N) time using a PRAM with N common CRCW processors. T F .‘ if sequential algorithm A requires execution of 1200 instructions, 600 of which ' M be performed sequentially (so that 50% of algorithm A's code cannot be parallelized)= then according to Amdahl‘s Law the maximum speedup possible from parallelizing algorithm A across 16 processors is therefore 8 . ...
View Full Document

{[ snackBarMessage ]}

Ask a homework question - tutors are online