You're reviewing a PR that claims the search results screen got 40% faster. It looks like it mixes the author's laptop simulator numbers with QA's real device numbers. How would you set the baseline and the metrics to verify the claim?
Fix the device model and OS version, the data fixture and network conditions, and cold versus warm, measure the pre-change commit 10 or more times under those conditions as the baseline, and use the median and the top 5% tail latency instead of the average, then measure the post-change commit under exactly the same conditions. Bake those conditions into an XCTest performance test or a CI job so the next regression gets caught too.
Performance and debuggingShort answerAdvanced
The account detail screen in a banking app takes 5 seconds to first display. To avoid…
#performance
Performance and debuggingShort answerAdvanced
A photo feed collection view stutters while scrolling. How do you use Instruments to tell…
#performance
Performance and debuggingShort answerBasics
Reviews of a secondhand marketplace app keep saying "it gets slow and then closes." How…
#performance
CS fundamentals and algorithmsMultiple choiceAdvanced
You iterate over 100,000 coordinates and sum distances. Between these two versions…
#performance