MHLab.Benchmark

A single-file utility to measure and profile some metrics from your code execution (execution time, average execution time, garbage collections).

Why?

Profiling/measuring how a piece of code performs is always a good thing. It helps to optimize resources usage and it saves costs!

I know that the great Benchmark.NET already exists and you really should use it for precise and in-depth profiling, but sometimes I just wanted to test a little snippet of code in a single line, without too many settings.

How to use it

It is really simple to use this utility. Just add it as reference from NuGet or include Benchmarker.cs file into your project.

Then just take a look at the following snippets.

Prepare your benchmark parameters:

var parameters = new BenchmarkParameters()
{
    // How many times your action will be invoked.
    BenchmarkIterations = 100_000,

    // Determines if the warmup will be performed.
    PerformWarmup = true,

    // The initialization action that will be invoked before the warmup.
    InitializeAction = null,

    // How many times your action will be invoked without collecting results.
    WarmupIterations = 1000
};

Then you are ready to collect metrics from your code:

private static void ActionToTest()
{
    // Your code to profile here
}

BenchmarkResult result = Benchmarker.Start(ActionToTest, parameters);

Or, if you have particular needs and the classic linear loop performed by this utility is not enough for you, simply call StartScope to define the benchmarking logic by yourself:

BenchmarkResult result = new BenchmarkResult("MyCode-1");

using (var benchmark = Benchmarker.StartScope(result))
{
    // Your code to profile here
}

// You can now use the result here, it has been populated with metrics.

Also, take in mind that BenchmarkResult has a ToString override to show results easily:

Console.WriteLine(result.ToString());

/*
Output:

Execution Time (TimeSpan): {result.Elapsed}
Execution Time (ms): {result.ElapsedMilliseconds}
Execution Ticks: {result.ElapsedTicks}
Average Execution Time (ms): {result.AverageMilliseconds}
Average Execution Ticks: {result.AverageTicks}
Garbage Collections (0): {result.GarbageCollections0Count}
Garbage Collections (1): {result.GarbageCollections1Count}
Garbage Collections (2): {result.GarbageCollections2Count}
*/

Comparing results

A good thing when you benchmark your code is comparing results from different methods against a baseline and check what changed.

To do so, check this snippet:

private static void ActionToTest()
{
    // Your code to profile here
}

private static void ActionToTest1()
{
    // Your code to profile here
}

private static void ActionToTest2()
{
    // Your code to profile here
}

var actions = new Action[]
{
    ActionToTest1,
    ActionToTest2
};

BenchmarkComparison[] comparisons = Benchmarker.StartAndCompare(ActionToTest, parameters, actions);