Making apps blazing fast with Baseline Profiles
summary
Claims that embedding Baseline Profiles into an APK yields near-instant optimized installs, using AGP’s Profgen to convert baseline.prof into compact DEX metadata, and Play’s DEX metadata or the Profile Installer to apply it at install or first run. Workflow uses Jetpack Macrobenchmark and BaselineProfileRule to generate rules and measure effects. Results show measurable wins — a Now in Android sample cuts ~60 ms to first frame and Maps reports ~30% faster cold launch — baseline profiles ship in libraries like Compose and work.
AI-written summary · about
chapters (14)
- 0:00Introduction to Baseline Profiles
- 1:09App compilation and launch process
- 3:56Evolution of app optimizations
- 8:12Cloud profiles and their limitations
- 11:21Baseline profiles explained
- 13:28Declaring baseline profile rules
- 15:55Using baseline profiles in Play
- 17:20Profile Installer for older platforms
- 19:03Generating baseline profile rules
- 21:26Measuring performance improvements
- 23:32Macrobenchmark under the hood
- 24:58Case studies: Maps and Play
- 26:42Why baseline profiles matter now
- 27:24Tools and resources
related talks
20mMaking apps blazing fast with Baseline Profiles
34mMaking apps blazing fast with Baseline Profiles
40mMy Journey: Adding Baseline Profiles to a Production-Based Android App: Success or Failure?
39mAndroid App Performance in a Nutshell
36mActionable App Performance Measurement
41mModern App Performance
05mWorkshop: Inspect and improve app performance
38mLow effort, High impact Android performance wins