Skip to content

python-session-04-part-02-quiz.md

Latest commit

 

History

History
141 lines (101 loc) · 2.99 KB

python-session-04-part-02-quiz.md

File metadata and controls

141 lines (101 loc) · 2.99 KB
Raw

Session 4 Part 2 Quiz (Pool and Parallel Image Processing)

Question 1

What does multiprocessing.Pool mainly help with?

  • Managing SQL connections
  • Running the same function over many inputs in parallel
  • Rotating images automatically
  • Creating virtual environments

Answer: 2 Type: single Time: 40 Explanation: Pool is designed for distributing similar tasks across worker processes.

Question 2

What does pool.map(func, data) do?

  • Calls func only once
  • Applies func to each item in data
  • Sorts data in place
  • Creates output folders

Answer: 2 Type: single Time: 35 Explanation: map runs the same function over each input item.

Question 3

Why should each worker save to a unique output filename?

  • To make files larger
  • To avoid workers overwriting each other’s outputs
  • To speed up imports
  • To disable parallelism

Answer: 2 Type: single Time: 40 Explanation: Unique filenames prevent collisions when multiple workers write files.

Question 4

Which of the following is a good output naming pattern for this lab?

  • rotated.jpg for every image
  • rotated_image_{idx}.jpg
  • output.txt
  • file

Answer: 2 Type: single Time: 35 Explanation: Indexed filenames are clear and collision-safe.

Question 5

Why can this exercise benefit from parallel execution?

  • It removes all image operations
  • Images are independent tasks that can run concurrently
  • It changes JPG to PNG automatically
  • It avoids writing files

Answer: 2 Type: single Time: 40 Explanation: Each image can be downloaded and processed independently.

Question 6

What should you create before saving outputs?

  • A database table
  • Output folders (for example images/ and processed/)
  • A zip archive
  • A YAML file

Answer: 2 Type: single Time: 30 Explanation: Output directories must exist before files are written.

Question 7

Which timer is best for elapsed-time benchmarking in this lab?

  • time.timezone
  • time.sleep()
  • time.perf_counter()
  • os.getpid()

Answer: 3 Type: single Time: 30 Explanation: perf_counter() is high-resolution and suitable for timing.

Question 8

A fair serial vs parallel comparison should use:

  • Different URL sets
  • The same URL list and same processing steps
  • Fewer outputs in parallel
  • No timing output

Answer: 2 Type: single Time: 40 Explanation: Both modes must run equivalent work to compare meaningfully.

Question 9

If parallel timing is not faster in one run, the best next step is:

  • Assume multiprocessing is wrong
  • Repeat runs and compare trends
  • Delete timing code
  • Remove image rotation

Answer: 2 Type: single Time: 40 Explanation: Performance varies by network/system load; repeated runs are more reliable.

Question 10

Which statement is most accurate?

  • Pool-based parallelism is always faster for every workload
  • Parallel speed depends on workload size and system overhead
  • Serial always beats parallel
  • Worker count never matters

Answer: 2 Type: single Time: 40 Explanation: Real speedup depends on task characteristics and runtime overhead.