Parallel and Distributed Computing FRQ Practice

Responses receive AI feedback against a rubric after submission. Scores are practice estimates.

Question 1 of 5

A mapping job has 3 minutes of sequential tile loading and 9 independent tiles that each take 4 minutes on one core.

(a) Compute the one-core total time. (b) Compute the time with 3 cores processing tiles after loading. (c) Compute the speedup from (a) to (b). Show arithmetic.

Question 2 of 5

(a) Define parallel computing. (b) Define distributed computing. (c) Give one reason distributed computing can solve problems a single computer cannot.

Question 3 of 5

A video render has 10 minutes of sequential encoding setup. The remaining work is 30 identical independent frames of 2 minutes each on one GPU.

(a) Which change reduces time: doubling GPUs for frames, or doubling the sequential setup length? Justify with times. (b) Explain why the sequential portion still limits speedup. (c) If communication among 15 GPUs adds 6 minutes, explain why machine count alone is not efficiency.

Question 4 of 5

Sequential time for a census cleanup is 120 minutes. A parallel run finishes in 30 minutes.

(a) Calculate speedup. (b) Explain what speedup compares. (c) Explain whether speedup of 4 guarantees zero communication cost.

Question 5 of 5

A class grades 20 independent quizzes. Startup instructions take 4 minutes once. Each quiz takes 3 minutes for one grader.

(a) Find one-grader time. (b) Find time with 5 graders after startup. (c) State which parts are sequential versus parallel.