Performance regression tests

[AveryClapp]

Feature Description

Track cache simulator performance over time and alert on regressions.

Metrics

• Analysis throughput (events/second)
• Memory usage (peak RSS)
• Binary size (cache-sim executable)
• Build time

Benchmark Workloads

• Small: 10K events (target: <100ms)
• Medium: 1M events (target: <5s)
• Large: 100M events (target: <5min)

CI Integration

- name: Performance benchmark
  run: |
    ./bench/run-perf-tests.sh
    # Compare with baseline (main branch)
    # Fail if >10% slower

Storage

• Store historical results in git (bench/results/)
• Plot trend over time (Actions artifact)

Alerts

• Slack/Discord notification on regression
• Block PR if >20% slower

Implementation

• Benchmark harness (hyperfine or criterion)
• Historical tracking (JSON time series)
• Visualization (plot.py)
• CI integration

Files

• bench/run-perf-tests.sh (new)
• bench/baseline.json (historical results)
• .github/workflows/perf-tests.yml

References

• hyperfine - Benchmarking tool