page.title=Launch-Time Performance @jd:body

In this document

  1. Launch Internals
    1. Cold start
    2. Warm start
    3. Lukewarm start
  2. Profiling Launch Performance
    1. Time to initial display
    2. Time to full display
  3. Common Issues
    1. Heavy app initialization
    2. Heavy activity initialization
    3. Themed launch screens

Users expect apps to be responsive and fast to load. An app with a slow startup time doesn’t meet this expectation, and can be disappointing to users. This sort of poor experience may cause a user to rate your app poorly on the Play store, or even abandon your app altogether.

This document provides information to help you optimize your app’s launch time. It begins by explaining the internals of the launch process. Next, it discusses how to profile startup performance. Last, it describes some common startup-time issues, and gives some hints on how to address them.

Launch Internals

App launch can take place in one of three states, each affecting how long it takes for your app to become visible to the user: cold start, warm start, and lukewarm start. In a cold start, your app starts from scratch. In the other states, the system needs to bring the app from the background to the foreground. We recommend that you always optimize based on an assumption of a cold start. Doing so can improve the performance of warm and lukewarm starts, as well.

To optimize your app for fast startup, it’s useful to understand what’s happening at the system and app levels, and how they interact, in each of these states.

Cold start

A cold start refers to an app’s starting from scratch: the system’s process has not, until this start, created the app’s process. Cold starts happen in cases such as your app’s being launched for the first time since the device booted, or since the system killed the app. This type of start presents the greatest challenge in terms of minimizing startup time, because the system and app have more work to do than in the other launch states.

At the beginning of a cold start, the system has three tasks. These tasks are:

  1. Loading and launching the app.
  2. Displaying a blank starting window for the app immediately after launch.
  3. Creating the app process.

As soon as the system creates the app process, the app process is responsible for the next stages. These stages are:

  1. Creating the app object.
  2. Launching the main thread.
  3. Creating the main activity.
  4. Inflating views.
  5. Laying out the screen.
  6. Performing the initial draw.

Once the app process has completed the first draw, the system process swaps out the currently displayed background window, replacing it with the main activity. At this point, the user can start using the app.

Figure 1 shows how the system and app processes hand off work between each other.


Figure 1. A visual representation of the important parts of a cold application launch.

Performance issues can arise during creation of the app and creation of the activity.

Application creation

When your application launches, the blank starting window remains on the screen until the system finishes drawing the app for the first time. At that point, the system process swaps out the starting window for your app, allowing the user to start interacting with the app.

If you’ve overloaded {@link android.app.Application#onCreate() Application.oncreate()} in your own app, the app starts by calling this method on your app object. Afterwards, the app spawns the main thread, also known as the UI thread, and tasks it with creating your main activity.

From this point, system- and app-level processes proceed in accordance with the app lifecycle stages.

Activity creation

After the app process creates your activity, the activity performs the following operations:

  1. Initializes values.
  2. Calls constructors.
  3. Calls the callback method, such as {@link android.app.Activity#onCreate(android.os.Bundle) Activity.onCreate()}, appropriate to the current lifecycle state of the activity.

Typically, the {@link android.app.Activity#onCreate(android.os.Bundle) onCreate()} method has the greatest impact on load time, because it performs the work with the highest overhead: loading and inflating views, and initializing the objects needed for the activity to run.

Warm start

A warm start of your application is much simpler and lower-overhead than a cold start. In a warm start, all the system does is bring your activity to the foreground. If all of your application’s activities are still resident in memory, then the app can avoid having to repeat object initialization, layout inflation, and rendering.

However, if some memory has been purged in response to memory trimming events, such as {@link android.content.ComponentCallbacks2#onTrimMemory(int) onTrimMemory()}, then those objects will need to be recreated in response to the warm start event.

A warm start displays the same on-screen behavior as a cold start scenario: The system process displays a blank screen until the app has finished rendering the activity.

Lukewarm start

A lukewarm start encompasses some subset of the operations that take place during a cold start; at the same time, it represents less overhead than a warm start. There are many potential states that could be considered lukewarm starts. For instance:

Profiling Launch Performance

In order to properly diagnose start time performance, you can track metrics that show how long it takes your application to start.

Time to initial display

From Android 4.4 (API level 19), logcat includes an output line containing a value called {@code Displayed}. This value represents the amount of time elapsed between launching the process and finishing drawing the corresponding activity on the screen. The elapsed time encompasses the following sequence of events:

  1. Launch the process.
  2. Initialize the objects.
  3. Create and initialize the activity.
  4. Inflate the layout.
  5. Draw your application for the first time.

The reported log line looks similar to the following example:

ActivityManager: Displayed com.android.myexample/.StartupTiming: +3s534ms

If you’re tracking logcat output from the command line, or in a terminal, finding the elapsed time is straightforward. To find elapsed time in Android Studio, you must disable filters in your logcat view. Disabling the filters is necessary because the system server, not the app itself, serves this log.

Once you’ve made the appropriate settings, you can easily search for the correct term to see the time. Figure 2 shows how to disable filters, and, in the second line of output from the bottom, an example of logcat output of the {@code Displayed} time.


Figure 2. Disabling filters, and finding the {@code Displayed} value in logcat.

The {@code Displayed} metric in the logcat output does not necessarily capture the amount of time until all resources are loaded and displayed: it leaves out resources that are not referenced in the layout file or that the app creates as part of object initialization. It excludes these resources because loading them is an inline process, and does not block the app’s initial display.

Time to full display

You can use the {@link android.app.Activity#reportFullyDrawn()} method to measure the elapsed time between application launch and complete display of all resources and view hierarchies. This can be valuable in cases where an app performs lazy loading. In lazy loading, an app does not block the initial drawing of the window, but instead asynchronously loads resources and updates the view hierarchy.

If, due to lazy loading, an app’s initial display does not include all resources, you might consider the completed loading and display of all resources and views as a separate metric: For example, your UI might be fully loaded, with some text drawn, but not yet display images that the app must fetch from the network.

To address this concern, you can manually call {@link android.app.Activity#reportFullyDrawn()} to let the system know that your activity is finished with its lazy loading. When you use this method, the value that logcat displays is the time elapsed since the creation of the application object, and the moment {@link android.app.Activity#reportFullyDrawn()} is called.

If you learn that your display times are slower than you’d like, you can go on to try to identify the bottlenecks in the startup process.

Identifying bottlenecks

Two good ways to look for bottlenecks are Android Studio’s Method Tracer tool and inline tracing. To learn about Method Tracer, see that tool’s documentation.

If you do not have access to the Method Tracer tool, or cannot start the tool at the correct time to gain log information, you can gain similar insight through inline tracing inside of your apps’ and activities’ {@code onCreate()} methods. To learn about inline tracing, see the reference documentation for the {@link android.os.Trace} functions, and for the Systrace tool.

Common Issues

This section discusses several issues that often affect apps’ startup performance. These issues chiefly concern initializing app and activity objects, as well as the loading of screens.

Heavy app initialization

Launch performance can suffer when your code overrides the {@code Application} object, and executes heavy work or complex logic when initializing that object. Your app may waste time during startup if your Application subclasses perform initializations that don’t need to be done yet. Some initializations may be completely unnecessary: for example, initializing state information for the main activity, when the app has actually started up in response to an intent. With an intent, the app uses only a subset of the previously initialized state data.

Other challenges during app initialization include garbage-collection events that are impactful or numerous, or disk I/O happening concurrently with initialization, further blocking the initialization process. Garbage collection is especially a consideration with the Dalvik runtime; the Art runtime performs garbage collection concurrently, minimizing that operation's impact.

Diagnosing the problem

You can use method tracing or inline tracing to try to diagnose the problem.

Method tracing

Running the Method Tracer tool reveals that the {@link android.app.Instrumentation#callApplicationOnCreate(android.app.Application) callApplicationOnCreate()} method eventually calls your {@code com.example.customApplication.onCreate} method. If the tool shows that these methods are taking a long time to finish executing, you should explore further to see what work is occurring there.

Inline tracing

Use inline tracing to investigate likely culprits including:

Solutions to the problem

Whether the problem lies with unnecessary initializations or disk I/O, the solution calls for lazy-initializing objects: initializing only those objects that are immediately needed. For example, rather than creating global static objects, instead, move to a singleton pattern, where the app initalizes objects only the first time it accesses them.

Heavy activity initialization

Activity creation often entails a lot of high-overhead work. Often, there are opportunities to optimize this work to achieve performance improvements. Such common issues include:

Diagnosing the problem

In this case, as well, both method tracing and inline tracing can prove useful.

Method tracing

When running the Method Tracer tool, the particular areas to focus on your your app’s {@link android.app.Application} subclass constructors and {@code com.example.customApplication.onCreate()} methods.

If the tool shows that these methods are taking a long time to finish executing, you should explore further to see what work is occurring there.

Inline tracing

Use inline tracing to investigate likely culprits including:

Solutions to the problem

There are many potential bottlenecks, but two common problems and remedies are as follows: