Android App Performance in a Nutshell
summary
Shows concrete wins and tools: shipping baseline profiles cuts runtime and startup about 30% (Google Maps, Somato) and cutting startup by 77% is reported for a car‑sharing app; combining baseline profiles with dex layout optimizations (AGP 8.2+) yields roughly 60% faster cold start. Recommends using the macrobenchmark library and UI Automator to generate profiles, run tests on physical devices and release/profilable builds, add a profilable tag to capture real traces, and use Perfetto to verify that jitting and frame overruns actually disappear.
AI-written summary · about
chapters (10)
- 0:00Introduction to app performance
- 1:37Android quality guidelines
- 4:06User-focused performance issues
- 8:17Real-world performance measurement
- 12:47Improving app startup time
- 18:49Baseline profiles explained
- 25:32Benchmarking metrics and tools
- 27:40DEX layout optimizations
- 30:08Debugging with system tracing
- 37:57Next steps and resources
related talks
41mModern App Performance
36mActionable App Performance Measurement
42mPerformance Paradox: Navigating the Complexities of Mobile App Optimisation
30mLearn from developers: Lyft and Google Maps share their experiences optimizing app performance
38mLow effort, High impact Android performance wins
45mDeep dive into Modern Android Performance
05mWorkshop: Inspect and improve app performance
29mWhat's new in app performance