How do you use BenchmarkDotNet to accurately measure C# performance, and what are the golden measurement rules?

Short answer

Create a separate Console project and mark methods with [Benchmark]; the library automatically warms up, isolates the processor, and measures time and memory allocations with [MemoryDiagnoser].

Explanation

Manual Stopwatch measurements can be misleading because of JIT compilation, GC pauses, and processor-frequency changes. BenchmarkDotNet isolates the runtime environment, runs repeated Warmup cycles until code reaches full optimization (Tier 1 JIT), performs hundreds of statistical iterations, and generates precise tables of nanoseconds and Heap allocation bytes.

Code example

[MemoryDiagnoser]
public class StringBenchmarks {
    [Benchmark]
    public string UsingStringBuilder() {
        var sb = new StringBuilder();
        for (int i = 0; i < 100; i++) sb.Append(i);
        return sb.ToString();
    }
}

Common mistakes

• Running benchmarks in Debug instead of Release, producing distorted results without compiler optimizations.

Follow-up questions

• What is Method Inlining, and when should you use [MethodImpl(MethodImplOptions.AggressiveInlining)]?

Sources

• Microsoft Learn — Performance testing and benchmarking in .NET