You are on page 1of 3

CONFIDENTIAL CS/APR2011/CSC580

UNIVERSITI TEKNOLOGI MARA


FINAL EXAMINATION

COURSE PARALLEL PROCESSING


COURSE CODE CSC580
EXAMINATION APRIL 2011
TIME 3 HOURS

INSTRUCTIONS TO CANDIDATES

1. This question paper consists of five (5) 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
invigilator.

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 DO SO


This examination paper consists of 3 printed pages
© Hak Cipta Universiti Teknologi MARA CONFIDENTIAL
CONFIDENTIAL 2 CS/APR2011/CSC580

QUESTION 1

a) Describe Flynn's taxonomy of parallel processing architectures and draw the diagram
for each category.
(8 marks)

b) Differentiate Single Instruction Multiple Data (SIMD) and Multiple Instruction Multiple
Data (MIMD) architectures in parallel computers. Give TWO (2) differences for each.
(8 marks)

c) Describe TWO (2) important classes of parallel computers.


(4 marks)

QUESTION 2

a) Show the representation of a perfect shuffle interconnection network for 16 processors.


Use solid lines to represent the exchange links and arrow lines to represent the
processors connections.
(12 marks)

b) Describe the routing algorithm for Omega topology with 8 processors. Draw appropriate
diagram to show a possible path.
(8 marks)

QUESTION 3

a) Briefly explain the TWO (2) major differences between message passing and shared
memory computers.
(6 marks)

b) Define a system call that creates a new process. What is the characteristic of the newly
created process?
(4 marks)

c) Create a function in C named "Matrix Multiply". It takes two (n x n) matrices and


computes the product of two matrices. The amount of work may be divided among a
number of processes. Your function should take a number of processes as its argument
and create that number of processes using f ork() UNIX system call. Each child process
should take its own portion of the computation and display the appropriate message.
(10 marks)

© Hak Cipta Universiti Teknologi MARA CONFIDENTIAL


CONFIDENTIAL 3 CS/APR2011/CSC580

QUESTION 4

a) Describe the FOUR (4) phases in designing parallel algorithms.

(8 marks)

b) Define the following terms:

i. Data parallelism
ii. Functional parallelism
(6 marks)
c) Differentiate between p a r construct and f o r a i i construct.
(6 marks)

QUESTION 5

a) There are THREE (3) types of partitioning in block matrix computation. Describe each of
them with appropriate diagram.
(10 marks)

b) Find the minimum spanning tree of the following tree by using Prim algorithm.
A 2 D

F
(10 marks)

END OF QUESTION PAPER

© Hak Cipta Universiti Teknologi MARA CONFIDENTIAL

You might also like