Performance Benchmarks
Last Updated
These benchmarks were automatically generated on 2026-01-31 from the latest CI run.
Environment: Ubuntu Latest • .NET SDK 10.0.102
🚀 Runtime Benchmarks
Click on any benchmark to view detailed results:
- AsyncTests - Detailed performance analysis
- DataDrivenTests - Detailed performance analysis
- MassiveParallelTests - Detailed performance analysis
- MatrixTests - Detailed performance analysis
- ScaleTests - Detailed performance analysis
- SetupTeardownTests - Detailed performance analysis
🔨 Build Benchmarks
- Build Performance - Compilation time comparison
📊 Methodology
These benchmarks compare TUnit against the most popular .NET testing frameworks:
| Framework | Version Tested |
|---|---|
| TUnit | 1.12.93 |
| xUnit v3 | 3.2.2 |
| NUnit | 4.4.0 |
| MSTest | 4.0.2 |
Test Scenarios
The benchmarks measure real-world testing patterns:
- DataDrivenTests: Parameterized tests with multiple data sources
- AsyncTests: Realistic async/await patterns with I/O simulation
- ScaleTests: Large test suites (150+ tests) measuring scalability
- MatrixTests: Combinatorial test generation and execution
- MassiveParallelTests: Parallel execution stress tests
- SetupTeardownTests: Expensive test fixtures with setup/teardown overhead
Environment
- OS: Ubuntu Latest (GitHub Actions)
- Runtime: .NET 10.0.2 (10.0.2, 10.0.225.61305), X64 RyuJIT x86-64-v3
- SDK: .NET SDK 10.0.102
- Hardware: GitHub Actions Standard Runner (Ubuntu)
- Tool: BenchmarkDotNet v0.15.8, Linux Ubuntu 24.04.3 LTS (Noble Numbat)
Why These Numbers Matter
- No Mocking: All tests use realistic patterns, not artificial micro-benchmarks
- Equivalent Logic: Each framework implements identical test scenarios
- Warm-Up Excluded: Measurements exclude JIT warm-up overhead
- Statistical Rigor: Multiple iterations with outlier detection
Source Code
All benchmark source code is available in the tools/speed-comparison directory.
Continuous Benchmarking
These benchmarks run automatically daily via GitHub Actions.
Each benchmark runs multiple iterations with statistical analysis to ensure accuracy. Results may vary based on hardware and test characteristics.
Last generated: 2026-01-31T00:33:21.366Z