A fast path to benchmarking
summary
A new open source tool called Fastpath runs benchmarks interactively or in CI, stores raw runs in local or remote databases, and drives tests inside Docker-wrapped system images to avoid rebuilding OS artifacts. It focuses on detecting and bisecting regressions using statistical summaries and confidence intervals, stopping bisections early when intervals don’t overlap to save time. Example: a trans huge pages optimisation produces a clear, large delta in averaged runs; metadata and per-run distributions are retained for deeper analysis.
AI-written summary · about
chapters (8)
- 0:00Introduction to benchmarking
- 1:46Benchmarking challenges explained
- 8:41Introducing FastPath tool
- 11:55Statistical analysis of results
- 20:00Visualizing benchmark data
- 24:20Storing metadata and results
- 25:31Automating pass-fail decisions
- 26:55Closing remarks and Q&A
related talks
44mBenchmarking your commits: The Postgres Performance Farm
46mPerformance Testing Framework
40mAutomated Performance Testing In Preproduction
39mCreating & Scaling Meaningful Performance Tests in CI
01h 02mThe performance loop—A practical guide to profiling and benchmarking
33mPerformance Testing in Agile and DevOps
23mStop Performance Regressions Before Release Valkey’s CI Benchmarking
10m{ Fast, Cost-effective } – Choose Two