BenchmarkDotNet

Uses simple attributes like `[Benchmark]` and `[Params]` to write benchmarks similarly to unit tests, making it accessible without deep performance expertise, as shown in the example code.

Incorporates perfolizer and pragmastat engines to handle pilot experiments, warm-up iterations, and steady-state detection automatically, ensuring high measurement precision and warning about pitfalls.

Supports benchmarking across various runtimes (.NET Framework, .NET Core, Mono, NativeAOT), operating systems, and architectures with simple job configurations, demonstrated in the multi-runtime example.

Can measure managed and native memory traffic and provide disassembly listings for deep performance analysis, going beyond simple timing metrics.