Performance engineers use a specialized vocabulary. Phrases like profile performance and identify bottlenecks appear in optimization reports, SLO reviews, and technical post-mortems. This quiz tests the most common ones.
0 / 5 completed
1 / 5
Before optimizing, developers should ___ the application's performance using dedicated tools.
Profile performance is the technical collocation for using a profiler tool to collect runtime data on CPU, memory, and execution time. 'Measure' is generic. 'Analyse' comes after profiling. 'Test' is broader. You cannot optimize effectively without first profiling — it's the first step in the optimization cycle.
2 / 5
The profiling report helped engineers ___ the bottlenecks causing slow response times.
Identify bottlenecks is the standard collocation in performance work. It means to pinpoint exactly where the system slows down. 'Find' is informal. 'Locate' implies a physical position. 'Detect' is used for issues or anomalies. Identify collocates naturally with bottlenecks in engineering documentation.
3 / 5
Adding caching layers to the API helped ___ latency by 40%.
Reduce latency is the canonical performance collocation. While 'lower', 'cut', and 'decrease' are technically correct, reduce latency is what you will find in engineering articles, SLO documents, and technical blogs. It is the expected, neutral, professional phrasing.
4 / 5
After each change, the team ran load tests to ___ the results against the baseline.
Benchmark results is the fixed collocation meaning to measure performance against a known standard or baseline. 'Compare' is a step within benchmarking. 'Measure' is too generic. 'Test' doesn't imply comparison. Benchmark is the noun and verb used specifically for standardized performance comparison.
5 / 5
The team's goal was to ___ throughput by processing requests in parallel.
Optimize throughput is the technical collocation meaning to tune a system so it handles the maximum number of operations efficiently. 'Increase' and 'boost' imply raw growth. 'Improve' is vague. Optimize specifically implies intelligent tuning within constraints, which is what parallelism achieves.