Question 1
i) Draw a diagram to compare an algorithm run in a parallel pipeline, an array processor and a single processor using up to 4 processors.
ii) Calculate the time unit and the number of processors needed for each type of systems in (i) using up to 20 processors and the size of data is n= 1000
iii) Explain why synchronisation between different processes belonging to the same or different programs is an important task in parallel computing. Give an example to support your answer.
Question 2
a) Discuss the advantage and disadvantage of using single and multi-bus to implement for parallel computer system.
Support your answer with diagrams.
b) Explain the two classes of Data-flow machines.
c) Consider the following sequence :
I1 :V3=V1*V2
I2 :V4=V3+V6
I3 :V0=V3+V7
I4 :S2=V1*S3
I5 :S4=V4*S3
Explain how it is possible to run these equations in parallel.
Question 3
a) The performance of a parallel computer is measured through the speedup and efficiency. Explain speedup and efficiency of a parallel computer. Give a full example to support your answer
b) Processor manufacturers often characterise their processors as Reduced Instruction Set Computer "RISC" and Complex Instruction Set Computers "CISC" design. Compare RISC to CISC.
c) Suppose you wanted to calculate the average of three numbers A, B and C.
i) Explain briefly how you could do that in CISC
ii) Explain briefly how you could do that in RISC