Performance Paradox: Navigating the Complexities of Mobile App Optimisation
summary
Walks through concrete Android tests: uses Android Studio profiler and macrobenchmark rules on real devices, captures heap dumps to find two memory leaks (fixed version shows zero leaks), measures startup median ~560 ms, finds 95th‑percentile frame overrun 51 ms and CPU frame duration 95th 43 ms, and generates a baseline profile (ahead‑of‑time compiled) that claims ~30–40% startup improvement in real projects; demos run successfully with live metrics and reproducible scripts.
AI-written summary · about
chapters (7)
- 0:00Mobile performance challenges
- 7:56Android performance metrics
- 13:11Startup and power metrics demo
- 20:37Frame timing and jank
- 23:29Memory leaks detection
- 31:42Baseline profiling for production
- 39:57Q&A
related talks
39mAndroid App Performance in a Nutshell
38mLow effort, High impact Android performance wins
41mModern App Performance
36mActionable App Performance Measurement
05mWorkshop: Inspect and improve app performance
35mIn Search of Performance: the Subtle Art of Android Profiling
35mThe Impact of Micro-Optimizations on Large-Scale Apps
30mLearn from developers: Lyft and Google Maps share their experiences optimizing app performance