0.6.1

What's Changed

• feat: Run all benchmarks in isolation by @hassila in #56

Full Changelog: 0.6.0...0.6.1

0.6.0

What's Changed

• feat: Integrate with hdr histogram by @dimlio in #54
• fix: correct histogram benchmark by @dimlio in #55

New Contributors

• @dimlio made their first contribution in #54

Full Changelog: 0.5.4...0.6.0

0.5.4

What's Changed

• chore: Cleanup jemalloc helper by @hassila in #52
• feat: exportable benchmarks by @ORyanHampton in #49
• chore: Fix SwiftLint warnings by @hassila in #53

New Contributors

• @ORyanHampton made their first contribution in #49

Full Changelog: 0.5.3...0.5.4

0.5.3

What's Changed

• fix: [sc-1033] Remove circular dependency on Benchmark by @hassila in #48

Full Changelog: 0.5.2...0.5.3

0.5.2

Just a bump release to test SwiftPackageIndex

0.5.1

What's Changed

• ci: 🔄 Synced file(s) with ordo-one/public-repository-templates by @ordo-ci in #42
• ci: sync workflow files by @freef4ll in #43
• feat: SwiftPackageIndex support by @hassila in #45

New Contributors

• @ordo-ci made their first contribution in #42
• @freef4ll made their first contribution in #43

Full Changelog: 0.5.0...0.5.1

0.5.0

What's Changed

• chore: Move to use DateTime package by @hassila in #41

This version moves up the dependency to Swift 5.7 and macOS v13 as we start using real Swift Duration and Clock/Instant types.

Full Changelog: 0.4.3...0.5.0

0.4.3

What's Changed

• chore: Update documentation by @hassila in #38
• chore: Move to dispatch conditions / NIOLock instead of raw pthread mutex/conditions by @hassila in #39

Full Changelog: 0.4.2...0.4.3

0.4.2

Significant reduction in benchmarking overhead for simpler benchmarks that only care about .wallClock and .throughput making very small microbenchmarks work better.

Overall measurements show that malloc metrics are the most expensive (factor 3-4 or gathering operating system stats).

What's Changed

• chore: Remove unnecessary Darwin/Glibc includes and clean up some numerics usage by @hassila in #33
• chore: Add basic benchmark for empty benchmark to allow measurement of overhead by @hassila in #35
• fix: Only query and check malloc and operating system stats when needed by @hassila in #36
• fix: Remove unnecssary jemalloc queries by @hassila in #37

Full Changelog: 0.4.1...0.4.2

0.4.1

Major change is a fix for correctly applying thresholds. Also helper functions for defining them, e.g.

    let customThreshold = BenchmarkResult.PercentileThresholds(absolute: [.p25 : .millseconds(3), .p50 : .milliseconds(23)])
    let customThreshold2 = BenchmarkResult.PercentileThresholds(absolute: [.p25 : .mega(2), .p50 : .mega(3)])

What's Changed

• chore: Minor polish of output by @hassila in #29
• fix: Fix comparisons of absolute thresholds, add helpers for defining them by @hassila in #32

Full Changelog: 0.4.0...0.4.1