Download Exam Questions for Bachelor of Engineering in Electronic and Computer Engineering and more Exams Computer Science in PDF only on Docsity!
Cork Institute of Technology
Bachelor of Engineering (Honours) in Electronic Engineering – Award
(NQF – Level 8)
January 2007
Computer Engineering
(Time: 3 Hours)
Read instructions carefully
Section A : Answer any TWO questions
Section B: Answer any TWO questions
Use separate answer books for Sections A and B
All questions carry equal marks
Examiners: Prof. G. Hurley Dr. S. Foley Dr. D. Pesch Mr. F. O’Reilly
Section A
Answer two (2) questions from this section.
Q1 (a) List three fundamental principles of a Von Neumann machine architecture. Explain the benefits they bring and how/where these have appeared in the modern computer/programming model. [6 marks]
(b) Explain how pipelining can create difficulties in memory access in a Von Neumann machine and how this can be tackled effectively. How is this solution implemented in modern micro- processors, (give examples)? [6 marks]
(c) You have been asked to devise an algorithm to compute the following calculation across floating point vectors C and D each with 200,000 elements. You have available a Private Memory system with 100 processor nodes.
2
200 , 000
1
2
y ( C [ x ]) ( D [ 200 , 000 x ])
x
= Develop an algorithm using pseudo code or pseudo C code which could execute on one of these processor nodes. Describe the partitioning of the problem, the data distribution over the system and provide the code to achieve the communication between the nodes to calculate the final result. [13 marks]
[Total: 25 marks]
Q2 (a) Using diagrams describe how a high-end server type machine, with 4 commodity 32 bit processors (32 bit registers but 36 bit addresses for memory) (e.g. Sun UltraSparc or Pentium Xeon), would typically be organised, for good flexibility and performance. It will have 8-16 GB RAM. State and justify the memory models used. If you wanted to scale this machine for super-computing applications to 128 processors with 256 GB RAM, describe what changes you might make, stating why. [8 marks]
(b) Describe using diagrams how a paged based memory management system can translate a virtual space of 2048MB (31 bit) into a physical space of 256 MB (28 bits). Take into account that the average active code/data block is 256 KB in size for this system. [10 marks]
(c) Explain briefly what a process/task is. Draw a process state diagram, explain briefly the states in it, how one process follows it and the performance costs of managing/transferring between multiple processes. [7 marks]
[Total: 25 marks]
Q3 (a) Describe briefly two (2) of the following, using diagrams where appropriate.
- Control and Data Hazards and their solutions
- Superscalar Execution, benefits and assisting techniques.
- Two High Performance interconnect networks. . [6 marks]
(b) For Massive Parallel Systems, scalability is of major importance. Using work and efficiency curves describe your understanding of it and how memory and communication networks create the boundaries of this scalability. What effects do these have on algorithm design? [6 marks]
(c) A 3 Tap FIR filter can be described using the following equation.
=
2
0
k
y n xn k h k
where x is the sample value and h the co-efficient.
Show how this FIR filter can be represented using a 2-D data dependence graph and subsequently convert this into a 1-D Systolic Array processor. Show the inputs, outputs and all necessary timing information in the finished processor. [6 marks]
Show how this would extend to a 5 tap filter. When the filter coefficients h(n) for a particular 5 tap application are calculated, it works out that odd coefficients ( h(1), h(3) ) have zero value. Show what effects this might have on the data dependence graph and derive efficiencies in the final 5 tap design. [7 marks] [Total: 25 marks]
Q5. (a) Consider a point-to-point communication link between two computers. A go-
back-N ARQ technique with a window length of N = 7 is used to provide a reliable data communication. The data rate is R = 128kbit/s, frames are of length L = 512bits with a payload of 448bits, the speed of the electrical signal is c = 2 x 10 8 m/s, and the distance between the computers is d = 2500km. Calculate the maximum Bit Error Probability Pb that can persist on the link in order to guarantee transmission of a 100kbyte file in under 20 seconds.
NOTE: In your calculation assume that acknowledgement frames are never in error and that their transmission delay can be neglected. You may find the formulae given at the end of the paper useful. [9 marks] (b) Briefly explain why the maximum sequence number can only be half as large as the length of the transmission window for the selective-repeat ARQ protocol.
[6 marks]
(c) Consider a college LAN with a large number of PCs and a single e-mail server. The server receives 150 requests for e-mail downloads per minute. Requests arrive according to a Poisson process. The server requires 0.5 seconds on average to process a request and the average transmission time on the LAN is 0.25 seconds. Both server processing time and LAN transmission time are exponentially distributed.
(i) Calculate the total time each user has to wait until she receives her e- mails from the server. (ii) By upgrading the e-mail server to a dual processor machine, the average response time can be reduced to 0.3 seconds. Assuming users wait the same time as in (i) on average, calculate the percentage increase in e- mail requests the server can handle now. [10 marks]
Q6. (a) The input to a router is a data packet stream modelled by a Poisson process
with arrival rate λ = 500 packets/sec. The length of the data packets is exponentially distributed with mean L = 1024bit. (i) Assume that the router has m = 5 output lines of data rate R = 128kbit/s which transmit packets from the head of the queue. Calculate a) the probability that arriving packets have to queue before transmission and b) the average waiting time for packets in the queue.
(ii) Assume that the 5 output lines are aggregated into a single output line of data rate R = 640kb/s. Calculate the average waiting time for packets in the queue for this case. Discuss what you observe when comparing this with the result of part (i) of the question? [10 marks]
(b) (i) Briefly list the different locations within the topology of a computer
network where routing decisions are usually made.
(ii) Briefly explain how routing without any routing tables in nodes can work solely by including the destination address in a packet. [6 marks]
(c) Using the Bellman-Ford algorithm, develop the least-cost routing table for source node 7 for the network of 8 nodes shown in Figure 1. The link costs are valid in both directions. In your answer also provide the least cost with each route between the source node and any other node. [9 marks]
3
1
5
1
7
1 1
2
4 3
Figure 1
