performance-benchmark

name	performance-benchmark
description	Benchmark all sorting algorithm generators: measures step counts and execution time across input sizes and patterns (random, sorted, reverse, duplicates). Use when adding or modifying an algorithm to verify Big-O behaviour.
user-invocable	true

Performance Benchmark

Runs all algorithm generators across multiple input sizes and patterns, outputting step counts and timing.

When to Use

• After implementing a new algorithm (verify step count matches expected Big-O)
• After modifying an existing algorithm (regression check)
• When comparing two implementations

Procedure

1. Run the benchmark:

   npx tsx .github/skills/performance-benchmark/scripts/benchmark.ts
   

2. Review the output table showing step counts per algorithm × input pattern × size

3. Cross-reference against complexity-expectations.md to verify that step counts are within expected ranges

Interpreting results

• Compare counts against theoretical values in the expectations file
• A compare count significantly above the expected range indicates a bug
• Timing is indicative only — generator overhead dominates at small n