Modern App Performance
summary
Overview of concrete fixes and measured gains: partners shave app startup by about 30% (Somato reports 30% faster cold start and 600,000 extra orders), Google Maps sees ~30% startup and 2.5% search improvement using baseline profiles. Workflow uses microbenchmark and a profilable tag to generate and include baseline profiles in APKs, prefer release builds on real devices, compare compilation modes (none vs partial vs full), and monitor regressions with field and CI-backed local metrics.
AI-written summary · about
chapters (11)
- 0:00Introduction to modern performance
- 1:09Performance pillars and user focus
- 7:04Key app performance metrics
- 11:59Inspecting app performance
- 19:23Automating performance inspection
- 22:10Improving performance with baseline profiles
- 28:47Measuring baseline profile impact
- 29:30Monitoring app performance
- 33:01Field monitoring with Android vitals
- 35:49Local monitoring and CI benchmarks
- 38:15Closing and resources
related talks
39mAndroid App Performance in a Nutshell
45mDeep dive into Modern Android Performance
36mActionable App Performance Measurement
30mLearn from developers: Lyft and Google Maps share their experiences optimizing app performance
42mPerformance Paradox: Navigating the Complexities of Mobile App Optimisation
38mLow effort, High impact Android performance wins
05mWorkshop: Inspect and improve app performance
28mMaking apps blazing fast with Baseline Profiles