December 2013

You might also like

Download as pdf
Download as pdf
You are on page 1of 4
CONFIDENTIAL x CSIDEC 2013/CSC580 SS UNIVERSITI TEKNOLOGI MARA FINAL EXAMINATION COURSE PARALLEL PROCESSING COURSE CODE CSC580 EXAMINATION DECEMBER 2013 TIME 3 HOURS INSTRUCTIONS TO CANDIDATES 1 This question paper consists of six (6) questions. 2 ‘Answer ALL questions in the Answer Booklet. Start each answer on a new page. 3. Do not bring any material into the examination room unless permission is given by the invigitator. 4 Please check to make sure that this examination pack consists of i) the Question Paper ii) an Answer Booklet - provided by the Faculty DO NOT TURN THIS PAGE UNTIL YOU ARE TOLD TO D0 SO This examination paper consists of 4 printed pages (© Hak Cipta Universiti Teknologi MARA CONFIDENTIAL, CONFIDENTIAL 2 CSIDEC 2013/CSC580 QUESTION 1 a) Define FOUR (4) organizations of computers and memory modules according to Flynn's classification. (8 marks) b) Draw a representation of the perfect shuffle topology for parallel computers using 16 processors. (8 marks) ©) Explain the diameter and bisection width of the network drawn in 4(b). (4 marks) QUESTION 2 Briefly describe and draw THREE (3) categories share memory of a multiprocessor system. (15 marks) QUESTION 3 a) Identify and briefly explain THREE (3) choices of module granularity for parallel execution. (9 marks) b) Consider the following pseudocode: A(I) = x(1) +3 B(I + 1) = A(I) * C(I + 1) c(r + 4) = BZ) +A +1) D(Z + 2) = D(x) + D(I +1) cr + 3) = xq) Draw the dependency graph for the above statements. (9 marks) (© Hak Cipta Universiti Teknolog! MARA CONFIDENTIAL, CONFIDENTIAL 3 CSIDEC 2013/CSC580 QUESTION 4 Consider the following graph. a) Find the adjacency matrix (vertex matrices), M , for the above graph. (3 marks) b) Find the number of 3 step connection (or paths of length 3) from P2 to PS. (3 marks) ©) Find the number of 1, 2 or 3-step connections from P3 to P2. (6 marks) QUESTION 5 a) Define TWO (2) techniques in block partitioning. (4 marks) b) Discuss the implementation of barrier in a parallel algorithm. (5 marks) ©) Explain the purpose of using UNIX system call fork () in parallel computers. (4 marks) (© Hak Cipta Universiti Teknologi MARA CONFIDENTIAL CONFIDENTIAL 4 CSIDEC 2013/¢SC580 QUESTION 6 a) Consider the following weighted graph and devise a parallel minimum spanning tree utilizing the Prim’s algorithms. (10 marks) b) Assume two n xn matrices, A and B, where n is a power of 2, are used in parallel matrix multiplication. Each matrix is divided into four square submatrices. Suppose submatrices of Aare labeled App, Azo, Age, aNd Ago, and the submatrices of B are labeled Byp, Bpz, Bap, and Bgg. Explain on how a parallel algorithm technique can be used to complete the multiplication, (12 marks) END OF QUESTION PAPER (© Hak Cipta Universiti Teknolog! MARA CONFIDENTIAL

You might also like