A few weeks ago I talked about static anonymous functions and how they can help to limit the number of heap allocations when using lambdas. A colleague contacted me after this post with the question how to detect those allocations.
Great question! Let me share you some ways on how to do this.
Let me first give you a general answer and let’s then dive in 2 specific tools.
To discover excessive allocations when using lambdas you can use any memory profiler tool and look for allocations of *__DisplayClass* or various variants of Action* and Func*.
With this information, you already know what to look for.
Visual Studio Performance profiler
A first option to help you is the Performance Profiler in Visual Studio.
- Change the Build type to Release.
- Now go to Debug –> Performance Profiler…
- There are multiple profiling targets available, but we want to use the .NET Object Allocation tracking option so select this check box.
- Click the Start button to run the tool.
- After closing the profiled application or clicking on Stop collection we can view all allocations on the Allocation tab
More information: Analyze memory usage for .NET objects - Visual Studio (Windows) | Microsoft Learn
Roslyn Clr Heap Allocation Analyzer
Another option is the Roslyn based C# heap allocation diagnostic analyzer that can detect explicit and many implicit allocations like boxing, display classes a.k.a closures, implicit delegate creations, etc.
It can be installed directly in your project as a NuGet package:
dotnet add package ClrHeapAllocationAnalyzer
As with any analyzer, it gives you inline hints:
and warnings:
If you want to see it in action, have a look at the following video:
Remark: If you are using JetBrains Rider, you can achieve the same thing using the Heap Allocation Viewer plugin.