Create a Performance Advisor report for your application

Connect to Android device and collect frame data

Now that you’ve seen a Performance Advisor example report , you can use it to capture data from your own application.

Performance Advisor runs on a Streamline capture file, so the first step is to take a capture with Streamline. Streamline must capture extra frame data from the device, which Performance Advisor needs to generate a report. To capture the extra frame data, you must first run the provided Python script, streamline_me.py.

This script does the following:

• Temporarily installs a daemon application on your device, called gatord, which Streamline uses to collect counter data.
• Temporarily installs the OpenGL ES or Vulkan layer library file on your device, which is needed to collect frame data.
• Allows you to specify options for the capture, such as whether to collect screenshots when the FPS drops below a certain threshold.

To connect to the Android device and capture frame data:

1. Open a terminal or command prompt, navigate to the Arm Performance Studio install directory and locate the streamline_me.py script:

       
   
           
           
   cd <installation_directory>/streamline/bin/android
   
       
   

2. Run the script, enabling frame boundaries, with:

       
   
           
           
   python3 streamline_me.py --lwi-mode=counters 
   
       
   

3. The script returns a numbered list of the Android package names for the debuggable applications that are installed on your device. Enter the number of the application you want to profile.

       
   
           
           
   Searching for devices:
   RZ8MC03VVEW / SM-A505FN found
   
   Select a device:
   Auto-selected RZ8MC03VVEW / SM-A505FN
   
   Searching for debuggable packages:
   5 debuggable packages found         
   
   Select a debuggable packages:
    1) com.Arm.DarkArms
    2) com.UnityTechnologies.BoatAttack
    3) com.arm.malideveloper.openglessdk.occlusionculling
    4) com.arm.pa.paretrace
    5) com.sample.texturedteapot
    0) Exit script
   
       
   

   The script identifies the GPU in the device, installs the daemon application and layer library, then waits for you to complete the capture in Streamline.

For full instructions, see the Get started with Performance Advisor Tutorial .

Capture data with Streamline

1. Open Streamline and select the device and application on the Start tab.

2. Select Start capture to start capturing profile data from the target. Enter a name and location for the capture file that Streamline creates.

3. The application starts automatically on the device. Interact with the application as required.

4. When you’ve collected enough data, select Stop capture.

5. Return to your terminal, and press ENTER to terminate the streamline_me.py script.

Generate an HTML performance report

1. In the terminal window, navigate to the location where you stored the Streamline capture file (.apc).

2. Run Streamline’s streamline-cli command with the -pa option on the Streamline capture file to generate the report. The default name is report.html.

       
   
           
           
   streamline-cli -pa <options> my_capture.apc
   
       
   

   For a list of available options, see The Streamline-cli -pa command in Arm documentation, or run the following command:

       
   
           
           
   streamline-cli -pa -h
   
       
   

   To pass a list of options in a separate file to streamline-cli -pa, use:

       
   
           
           
   streamline-cli -pa <capture.apc> "@<options-file>"
   
       
   

Generate a JSON performance report

This feature is particularly useful when used within a CI workflow .

1. In the terminal window, navigate to the location where you stored the Streamline capture file (.apc).

2. Run Streamline’s streamline-cli command with the -pa and --type=json options on the Streamline capture file to generate the report named report.json:

       
   
           
           
   streamline-cli -pa --type=json:report.json <other_options> my_capture.apc
   
       
   

Specify performance budgets

You can specify a performance budget that is reflected in the Performance Advisor report. For more information, see Setting performance budgets in Arm documentation.

What you’ve accomplished and what’s next

You’ve now generated JSON and HTML Performance Advisor reports for your application.

Next, you’ll perform frame-based analysis on your application using Frame Advisor.