Benchmark Runner provides a Tool Window that displays all benchmarks in the current solution. From there you can run one or more benchmarks and view the results.
When running a benchmark, the Benchmark Runner builds and runs the benchmark project, passing the right command line parameters. Therefore it is necessary that the Benchmark
project uses the BenchmarkSwitcher class in its main method to make sure that BenchmarkDotNet can process the command line parameters, e.g.:
static void Main(string args)
It will not work if you explicitly start a Benchmark in the Main method, e.g. with BenchmarkRunner.Run<MyClass>()
How to use it
Open the "Benchmarks" Tool Window through View -> Other Windows -> Benchmarks
Click the "Refresh" button in the Benchmarks Window
Select a node in the tree representing the benchmarks you want to run. Depending on the current grouping you can select a project, namespace, category, class or method.
Run the benchmarks via the toolbar or the context menu. This will build the project and start it.
The console window with the BenchmarkDotNet output is displayed and will stay open until manually closed
Results can be viewed by enabling the result section from the toolbar ("Results Right" or "Results Bottom").
You can also open the results folder by selecting the folder icon in the toolbar.
If you need additional diagnosers or command line parameters, you can specify them in the options dialog.
You can navigate to the code of a benchmark class or method by double clicking it in the tree or selecting "Go To Code" in the context menu.
Limitations / Known Issues:
If you add/remove/rename benchmark classes/methods in the code, the changes won't be visible in the tree view immediately. You need to click "Refresh" to update the tree.
If benchmarks of more than one class are run together, the logfile content won't show up in the log tab.
You can only select one node in the tree view at once