DevBench

Benchmark your developer PC's performance with real-world .NET build scenarios.

Quick Start

Windows (PowerShell)

git clone https://github.com/damianedwards/DevBench.git
cd DevBench
.\run-benchmarks.ps1

macOS / Linux (Bash)

git clone https://github.com/damianedwards/DevBench.git
cd DevBench
./run-benchmarks.sh

The interactive menu will guide you through selecting which benchmarks to run.

What It Measures

DevBench measures build performance across different scenarios:

• Cold builds - First build after clearing caches
• Warm builds - Subsequent builds with warm caches
• Incremental builds - Rebuilding after a small code change

Results capture detailed system information so you can compare performance across different hardware and configurations.

System Information Collected

• OS: Platform, version, architecture
• CPU: Model, cores, frequency
• Memory: Capacity, speed
• Storage: Type (SSD/HDD), file system (NTFS, ReFS, ext4, btrfs, APFS), model, free space
• Windows-specific: Dev Drive status, Windows Defender exclusions
• macOS-specific: Chip model (Apple Silicon/Intel), Rosetta 2 status
• Linux-specific: Distribution, kernel version

Submitting Results

After running benchmarks, your results are saved to results/results-<timestamp>-<hash>.json.

To share your results:

1. Fork this repository
2. Copy your results file to the results/ folder
3. Submit a pull request

Your results will appear on the DevBench Results page after merging.

Adding New Benchmarks

See docs/benchmark-schema.md for the benchmark manifest format.

Benchmarks can be:

• In-repo: Source code included in the benchmarks/ folder
• External-repo: Points to another GitHub repository

Project Structure

DevBench/
├── run-benchmarks.ps1      # Windows bootstrapper
├── run-benchmarks.sh       # Unix bootstrapper
├── src/
│   └── DevBench.cs         # Test harness (file-based .NET 10 app)
├── benchmarks/             # Benchmark definitions
│   └── dotnet-hello-world/
│       ├── benchmark.json
│       └── src/
├── results/                # Submitted benchmark results
├── site/                   # Results website (Eleventy)
└── docs/                   # Documentation

Requirements

• To run benchmarks: Just clone and run the script (downloads pre-built harness and benchmarks)
• To develop the harness or benchmarks: .NET 10 SDK, use --dev flag

Development Mode

When contributing to DevBench (harness or benchmarks), use the --dev flag:

# Windows
.\run-benchmarks.ps1 -Dev

# macOS / Linux
./run-benchmarks.sh --dev

This builds the harness locally and uses the local benchmarks/ folder instead of the release versions.

License

MIT License - see LICENSE