Fast Hints Checkpoint

Estimation, measurement, memory layout, and avoiding unnecessary work.


Get every question correct to pass this checkpoint. You can review feedback, revise your answers, and submit again.

1. A 100 ms request spends 10 ms formatting. Formatting becomes twice as fast; other costs stay fixed. New total?

Select one answer.

2. A batch API reduces per-item overhead but raises p99 waiting time. What should determine whether to keep it?

Select one answer.

3. A microbenchmark improves while production slows. What evidence is most useful next?

Select one answer.

4. A task takes 80 ms, of which 20 ms is a completely removable cost. What is the best total time if other work is unchanged? Answer with the integer number of milliseconds only.
5. Four workers spend most time waiting for one lock. Which experiment tests the suspected bottleneck?

Select one answer.

6. A pool retains very large buffers after rare requests. Which comparison is needed before claiming an improvement?

Select one answer.

7. EncodeKeys sorts the dictionary and adjusts every ID. Reconstruction still succeeds. Is that sufficient for correctness?

Select one answer.

Join Discord