Do you need the Jetpack Macrobenchmark library?
summary
Jetpack Macrobenchmark measures real app startup and scrolling with startup and frame-timing metrics, requires explicit iteration counts, and needs a single consistent physical device for repeatable results. It produces systrace-based traces for per-frame timing and supports hot/warm/cold startup and compilation modes (none, speed-profile, speed). Benchmarks detect regressions over time but are slow to author and run; CI setup at Allegro proves nightly runs on master work better than per-PR checks.
AI-written summary · about
chapters (13)
- 0:00Introduction to Jetpack Macrobenchmark
- 1:15What is Jetpack Macrobenchmark
- 3:01Jetpack Micro vs Macro Benchmarks
- 5:13Advantages of Micro Benchmarks
- 8:01Debug vs Release Performance
- 11:13Measuring App Startup Performance
- 17:56Startup Benchmark Parameters
- 20:26Testing Scroll Performance
- 24:42When to Use Benchmarks
- 29:38Benchmark Alternatives
- 33:08Real-world Benchmark Experience
- 39:49Choosing the Right Tool
- 41:13Q&A Session
related talks
10mMeasuring Jank and Startup with Macrobenchmark
40mMonitor App performance using Jetpack Benchmark
25mJetpack Benchmark and where to find them
26mJetpack Benchmarks and Where to Find Them
33mAndroid Benchmarking and other stories
44mPractical guide to performance Benchmarks
29mAndroid Benchmarking and other stories
20mMeasure What Matters with Jetpack Benchmark (Lightning)