Baseline Profiles overview

Baseline Profiles improve code execution speed by about 30% from the first launch by avoiding interpretation and just-in-time (JIT) compilation steps for included code paths.

By shipping a Baseline Profile in an app or library, Android Runtime (ART) can optimize specified code paths through Ahead-of-Time (AOT) compilation, providing performance enhancements for every new user and every app update. This Profile Guided Optimization (PGO) lets apps optimize startup, reduce interaction jank, and improve overall runtime performance for users from the first launch.

These performance enhancements directly result in improved business metrics such as user retention, transactions, and ratings. You can read more about how performance impacts business metrics in stories from Josh, Lyft, TikTok, and Zomato.

Benefits of Baseline Profiles

Baseline Profiles make all user interactions—such as app startup, navigating between screens, or scrolling through content—smoother from the first time they run. By increasing the speed and responsiveness of an app, Baseline Profiles can lead to more daily active users and a higher average return visit rate.

Baseline Profiles help guide optimization beyond app startup by providing common user interactions that improve app runtime from the first launch. Guided AOT compilation doesn't rely on user devices and can be done once per release on a development machine instead of a mobile device. By shipping releases with a Baseline Profile, app optimizations become available much faster than by relying on Cloud Profiles alone.

When not using a Baseline Profile, all app code is either JIT-compiled in memory after being interpreted, or written to an odex file in the background when the device is idle. After installing or updating an app, users have a suboptimal experience from the first time they run it until new code paths are optimized. Many apps measure performance boosts of about 30% after optimizing.

Startup profiles

Startup profiles are similar to Baseline Profiles, but the difference is they are used at compile time rather than for on-device optimization. A startup profile is used to optimize the layout of the DEX file to improve startup times. The code identified in the startup profile is put into the primary classes.dex file, and other code is put into separate DEX files. This improves startup times by reducing the number of page faults during app startup. To learn more about how startup profiles and DEX layout optimizations can improve app startup times, see DEX layout optimizations and startup profiles.

Get started

To start optimizing performance in your existing app, see Create Baseline Profiles.

The dependency chain provides stable and developmental release versions. To generate and install a Baseline Profile, use the following supported versions or higher of Android Gradle plugin, Macrobenchmark library, and Profile Installer. These dependencies are required at different times and work together as a toolchain to enable an optimal Baseline Profile.

  • Android Gradle plugin:
  • Macrobenchmark library: androidx.benchmark:benchmark-macro-junit4:1.1.1
  • Profile Installer: androidx.profileinstaller:profileinstaller:1.3.1

Profile generation example

The following is an example class to create a Baseline Profile for app startup, as well as several navigation and scroll events using the recommended Macrobenchmark library:

class BaselineProfileGenerator {
    val baselineProfileRule = BaselineProfileRule()

    fun appStartupAndUserJourneys() {
        baselineProfileRule.collect(packageName = PACKAGE_NAME) {
            // App startup journey.

            device.findObject(By.text("COMPOSE LAZYLIST")).clickAndWait(Until.newWindow(), 1_000)
            device.findObject(By.res("myLazyColumn")).also {

You can see this code in full context and more detail as part of our performance samples on GitHub.

What to include

When using Baseline Profiles in an app, you can include app startup code and common user interactions like navigation between screens or scrolling. You can also gather entire flows such as registration, login, or payment. Any user journeys that you deem critical can benefit from Baseline Profiles by improving their runtime performance.

If you are experimenting with different approaches to improve performance, consider including Baseline Profiles for both arms of your experiment. By doing this, you can make your results easier to interpret by ensuring all of your users are consistently running compiled code.

Libraries can provide their own Baseline Profiles and ship them with releases to improve app performance. For example, see the Use a Baseline Profile section in Jetpack Compose performance.

How Baseline Profiles work

While developing your app or library, consider defining Baseline Profiles to cover common user interactions where rendering time or latency are important. Here's how they work:

  1. Human-readable profile rules are generated for your app and compiled into binary form in the app. You can find them in assets/dexopt/ You can then upload the AAB to Google Play as usual.

  2. Google Play processes the profile and ships it directly to users along with the APK. During installation, ART performs AOT compilation of the methods in the profile, resulting in those methods executing faster. If the profile contains methods used in app launch or during frame rendering, the user might experience faster launch times and reduced jank.

  3. This flow cooperates with Cloud Profiles aggregation to fine-tune performance based on actual usage of the app over time.

Figure 1. This diagram demonstrates the Baseline Profile workflow from upload through end-user delivery, and how that workflow relates to Cloud Profiles.

Cloud Profiles

Cloud Profiles offer an additional form of PGO—aggregated by Google Play Store and distributed for install time compilation—together with Baseline Profiles.

While Cloud Profiles are driven by real-world user interactions with the app, they take several days to weeks after an update to be distributed, limiting their availability. Until profiles are fully distributed, app performance is suboptimal for users of new or updated apps. Further, Cloud Profiles only support Android devices running Android 9 (API level 29) or higher, and only scale well for apps that have a sufficiently large user base.

Compilation behavior across Android versions

Android Platform versions use different app compilation approaches, each with a corresponding performance tradeoff. Baseline Profiles improve upon the previous compilation methods by providing a profile for all installs.

Android version Compilation method Optimization approach
5 up to 6 (API level 21 up to 23) Full AOT The entire app is optimized during install, resulting in long wait times to use the app, increased RAM and disk space usage, and longer times to load code from disk, potentially increasing cold startup times.
7 up to 8.1 (API level 24 up to 27) Partial AOT (Baseline Profile) Baseline Profiles are installed by androidx.profileinstaller on the first run when the app module defines this dependency. ART can improve this further by adding additional profile rules during the app's use, and compiling them when the device is idle. This optimizes for disk space and time to load code from the disk, thereby reducing wait time for the app.
9 (API level 28) and higher Partial AOT (Baseline + Cloud Profile) Play uses Baseline Profiles during app installs to optimize the APK and Cloud profiles—if available. After installation, ART profiles are uploaded to Play, aggregated, and then provided as Cloud Profiles to other users when they install or update the app.

Solutions for possible issues

The following are possible issues and solutions, or issues for which there are ongoing developments for workarounds:

  • Baseline profiles aren't packaged correctly when building the APK from an app bundle. To resolve this issue, apply or higher (issue) .

  • Baseline profiles are only correctly packaged for the primary classes.dex file. This affects apps with more than one .dex file. To resolve this issue, apply or higher.

  • Resetting ART profile caches isn't allowed on user (non-rooted) builds. To work around this, androidx.benchmark:benchmark-macro-junit4:1.1.0 includes a fix that reinstalls the app during the benchmark (issue) .

  • Android Studio Profilers don't install Baseline Profiles when profiling the app (issue.

  • Non-Gradle-build systems—such as Bazel or Buck—don't support compiling Baseline Profiles into output APKs.

  • Non-Google-Play-Store app distribution channels might not support using Baseline Profiles at installation. Users of apps installed through these channels don't see the benefits until background dexopt runs—which is likely overnight.

  • The build compiler accepts only one baseline-prof.txt in the src/main folder and doesn't reflect files in different flavors or build types. This is being actively improved.

  • Battery optimizations can interfere with profile installation. To help ensure that your profiles are installed effectively, disable any battery optimizations in your benchmark devices.

  • Performance improvements might differ between benchmarks and production. This happens because local benchmarks measure performance with Baseline Profiles enabled or disabled. In a production app, the measurement is incremental when adding a new part of the app to a Baseline Profile, where parts are already profiled through contributing libraries.

Additional resources