Most of the code I write is library code aimed to help other developers be more productive. As a consequence, a lot of this code is part of the critical path of an application and is execute a lot. This means that performance is critical. To measure the impact of a code change, I like to use BenchmarkDotNet, a powerful library for benchmarking .NET code.
When of the benchmark tests I got was to compare the impact between logging enabled and disabled.
I removed most the implementation in the code above but what is important is that the implementation of both benchmark methods was completely the same with only the addition of an EnableLogging() call as a difference.
An alternative(read better) way to implement this, is by using parameters. I added an extra property IsLoggingEnabled and specified the values using the [Params] attribute:
The benchmark is run for every combination of parameters you’ve provided and nicely visualized in the results:
// * Summary *
BenchmarkDotNet v0.13.10, Windows 10 (10.0.19044.4046/21H2/November2021Update) (Hyper-V)
Intel Xeon Gold 6154 CPU 3.00GHz, 1 CPU, 4 logical and 4 physical cores
.NET SDK 8.0.101
[Host] : .NET 8.0.1 (8.0.123.58001), X64 RyuJIT AVX2
DefaultJob : .NET 8.0.1 (8.0.123.58001), X64 RyuJIT AVX2
| Method | IsLoggingEnabled | Mean | Error | StdDev | Gen0 | Gen1 | Allocated |
|------------ |----------------- |----------:|----------:|-----------:|-----------:|---------:|----------:|
| GetProducts | False | 91.78 ms | 1.830 ms | 3.481 ms | 750.0000 | 250.0000 | 9.65 MB |
| GetProducts | True | 345.69 ms | 58.352 ms | 172.051 ms | 29333.3333 | - | 362.61 MB |
If you need more values or want to generate them dynamically, you can also use [ParamsSource]. You need to specify the name of a public method or property that provides the values (implementing IEnumerable). The source must be within the benchmarked type: