Performance Benchmarks with ASV

A test to demonstrates performance tracking over time using ASV

Setup

pip install asv

How to Add a New Performance Point

Step 1: Modify benchmarks to simulate performance change

For performance improvement, edit benchmarks/benchmarks.py:

• Reduce loop iterations
• Reduce data sizes
• Simplify calculations

For performance regression:

• Increase loop iterations
• Add more complex calculations
• Increase data processing

Step 2: Create a new version

# Make your changes to benchmarks/benchmarks.py
git add -A
git commit -m "v2.6.0: Your performance change description"  # Use your version number
git tag v2.6.0  # Use your own version number (v2.6.0, v2.7.0, etc.)

Step 3: Run benchmarks on the new version

asv run v2.6.0^!  # Replace with your version number

Step 4: Update graphs

asv publish
asv preview

Step 5: View results

Open http://localhost:8080 to see the new data point on the graphs.

Current Benchmark History

• v2.1.0: Initial test benchmarks
• v2.2.0: Performance optimizations
• v2.3.0: Feature expansion (regression)
• v2.4.0: Partial recovery
• v2.5.0: Major optimization breakthrough

Each shows different performance patterns in the ASV graphs