2.1.5 - Thinking Concurrently

Question 1

What is concurrent processing?

Answer 1

• Carrying out more than one process at a time (overlapping)
• Where one process does not have to finish before the other starts
• Each process is given a slice of processor time
• Different processes can be executed by different cores in parallel

Question 2

What are the benefits of concurrent processing?

Answer 2

• More efficient use of the processor
• Longer running tasks do not delay short running tasks
• User is able to interact with the computer while other tasks are running

Question 3

What are the trade-offs of concurrent processing?

Answer 3

• Locking will be needed meaning more complex programming (prevents two processes from updating the same thing)
• Not all the processes will be able to be done in parallel (parallelizable).
• Cannot be done when the current stage depends on the previous stage.
• X processors don’t mean it will run in 1/xth of the time of one processor.

Question 4

How might tree traversal use concurrent processing?

Answer 4

Apply different searches simultaneously: perform breadth-first and depth-first searches simultaneously.

Searching a tree: rather than going down 1 path, go down 2+

Question 5

How might a linear search use concurrent processing?

Answer 5

Can have multiple processors searching different areas at the same time.

Question 6

How might a binary search use concurrent processing?

Answer 6

It cannot - it is not parallelizable.

The next move depends on the outcome of checking the mid-point.

This algorithm does not benefit from using multiple processors.

Question 7

How might a cinema booking system use concurrent processing?

Answer 7

Multiple orders happening at the same time.

Record locking used to stop double bookings