Skip to content

Fix benchmark p95 calculation - #165

Closed
codeforester wants to merge 1 commit into
mainfrom
bug/141-20260807-benchmark-p95
Closed

Fix benchmark p95 calculation#165
codeforester wants to merge 1 commit into
mainfrom
bug/141-20260807-benchmark-p95

Conversation

@codeforester

Copy link
Copy Markdown
Contributor

Summary

  • compute an interpolated p95 from benchmark samples instead of aliasing the maximum
  • add regression coverage proving p95 and maximum remain distinct

Closes #141.

Validation

  • pytest -q tests/test_benchmark_runtime.py
  • ruff check scripts/benchmark_runtime.py tests/test_benchmark_runtime.py
  • git diff --check

@codeforester
codeforester deleted the bug/141-20260807-benchmark-p95 branch August 7, 2026 20:18
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Benchmark's reported "p95" metric is actually just the max, not a real percentile — mislabels the CI perf gate

1 participant