Making apps blazing fast with Baseline Profiles
summary
Baseline Profiles claim an easy 30–40% cold-start win and often deliver it in an afternoon by shipping a compact profile (authoritative examples: Maps sees ~30% startup speedup, Play ~40%). Developers add a human-readable baseline-profile.txt (AGP runs profgen to merge and compile it), embed DEX metadata so installs produce ODEXes, and use profile installer plus the macrobenchmark rule to bootstrap and measure cold-start improvements locally.
AI-written summary · about
chapters (8)
- 0:00Introduction to Baseline Profiles
- 1:14App Compilation Basics
- 8:16Profile-Guided Optimization
- 15:40Introducing Baseline Profiles
- 18:05Creating Baseline Profiles
- 22:39Testing and Validation
- 25:05Case Studies and Results
- 32:48Key Takeaways and Resources
related talks
28mMaking apps blazing fast with Baseline Profiles
20mMaking apps blazing fast with Baseline Profiles
39mAndroid App Performance in a Nutshell
36mActionable App Performance Measurement
40mMy Journey: Adding Baseline Profiles to a Production-Based Android App: Success or Failure?
41mModern App Performance
05mWorkshop: Inspect and improve app performance
36mWin on startup time! Improving app launch performance