By default, the Profiler window displays information about Unity’s native code. It uses built-in profiler markers to organize and divide up the performance data it collects.

You can add profiler markers to your own code to make timings spent in these areas visible in the Profiler window with the

ProfilerMarker

API.

You can also add Profiler counters to your code to collect data for metrics in your application, and use custom profiler modules to display this data. This is useful if you want to track performance changes in your application. Adding profiler counters to your code speeds up the investigation of performance issues because you can use the information from your counters in conjunction with Unity’s built-in counters and instrumentation data.

To add counters and markers with metadata to your code, you need to use the Unity Profiling Core package.

com.unity.profiling.core

You can use the

ProfilerMarker

API to mark up resource-intensive script code blocks and make them visible in the Unity Profiler, or use the

ProfilerMarker<TP1>

API in the Unity Profiling Core package to add extra metadata to each sample it produces.

This can be useful because the built-in Unity Profiler doesn’t profile all method calls. The alternative is to use deep profiling, but this causes a large overhead that significantly slows down your application execution and skews the results. Using is a more efficient way of marking up your code.

represents a named Profiler handle and is the most efficient way of profiling your code. You can use it in any of your application's C# code.

Profiler markers have no overhead when Unity deploys them in a non-development build, so you can mark up as many samples as you like.

Unity marks and methods with

ConditionalAttribute

and conditionally compiles them, so they have zero execution overhead in non-development builds. While methods are not entirely compiled out in non-development builds, they are conditionally compiled to just return null, making their overhead negligible. The profiler marker’s field is present in release builds and takes up memory equivalent to its

IntPtr

, which is 8 bytes.

The

ProfilerRecorder

API also uses profiler markers to capture performance data. During development in the Unity Editor and in Development Players, you can use the API to get a performance overview of the different parts of your code and to identify performance issues.

To display custom metrics in the Unity profiler, use the API in the Unity Profiling Core package. The Profiler can display data from

ProfilerCounter

or

ProfilerCounterValue

.

Unity groups Profiler counters into categories based on the type of work the counters profile, for example, Rendering, Scripting, or Animation. You can assign a custom Profiler counter to any of Unity's profiling categories. For a full list of available Profiler categories, refer to

ProfilerCategory

.

You can use the or API to track metrics of your application and make them visible in the Unity Profiler or in other code. If you’re an Asset Store package developer, you can add Profiler counters to your code to help other developers understand important performance characteristics of your system, and they can use this information for optimization or budgeting tooling.

The following diagram displays a high level overview of the Profiler counters data flow:

The

ProfilerRecorder

API retrieves Profiler counter data in your application code, and the

RawFrameDataView

or

HierarchyFrameDataView

APIs get the Profiler counter data in the Editor code. You can visualize this counter data in the Profiler window by configuring a custom Profiler module in the Module Editor.

and support the following types: