v1.5.0

BenchmarkTools v1.5.0

Diff since v1.4.0

Main changes

• Warm up is now systematic in every individual benchmark and suite (#330)

Merged pull requests:

• warm up at the beginning of every measurement (#330) (@willow-ahrens)
• Update BenchmarkTools version constant (#349) (@Zentrik)
• Only run GC 3 times instead of 4 to scrub (#350) (@Zentrik)
• Allow loading saved benchmark results from version <= 1.3.2 (#351) (@Zentrik)
• Bump actions/cache from 3 to 4 (#352) (@dependabot[bot])
• Deserialize Inf correctly in some cases (#354) (@Zentrik)
• Bump to 1.5.0 (#357) (@gdalle)
• Add 1.10 to CI (#358) (@gdalle)
• Allow nightly to fail (#359) (@gdalle)

Closed issues:

• Warmups are skipped when running a benchmark suite (#291)
• Should hardcoded version number be updated? (#348)
• cannot read benchmark JSON if run & written with seconds=Inf (#353)

v1.4.0

BenchmarkTools v1.4.0

Diff since v1.3.2

Merged pull requests:

• Make @benchmarkset work with for loop (#284) (@skleinbo)
• Add Example Code to Common Workflow (#290) (@FelixBenning)
• Use nameof instead of Type.name.name (#296) (@jishnub)
• Create dependabot.yml (#302) (@ranocha)
• Bump actions/checkout from 2 to 3 (#303) (@dependabot[bot])
• Bump codecov/codecov-action from 1 to 3 (#304) (@dependabot[bot])
• Bump actions/cache from 1 to 3 (#305) (@dependabot[bot])
• Automatically create nested BenchmarkGroup on access (#309) (@MilesCranmer)
• Fix a few typos (#310) (@pitmonticone)
• Clarify multiple setup values (#312) (@gdalle)
• Add CI on Julia 1.9 and fix Ubuntu builds for unsupported versions (#313) (@gdalle)
• Semicolons instead of tuple in multiple setup docs (#317) (@gdalle)
• Don't tune manually set evals parameter (#318) (@gdalle)
• Add logo (#320) (@gdalle)
• Fix logo (#321) (@gdalle)
• Fix logo path (#322) (@gdalle)
• Logo img was in docs/build (#323) (@gdalle)
• Add Aqua to tests (#326) (@gdalle)
• Apply JuliaFormatter with BlueStyle and add it to tests (#327) (@gdalle)
• Bump actions/checkout from 3 to 4 (#331) (@dependabot[bot])
• don't export leaves (#333) (@lucaferranti)
• Add compat lower bound Statistics >= 1 (#334) (@gdalle)
• Reexport std and var from Statistics (#336) (@milesfrain)
• Update attribution in readme (#341) (@LilithHafner)
• Revert "Update attribution in readme" (#342) (@vchuravy)
• Bump version to 1.4.0 (#346) (@gdalle)

Closed issues:

• tune! overrides evals=1, causing errors in destructive benchmarks (#24)
• judge working with memory (#70)
• Replace markdown manual/reference documentation with actual docstrings + Documenter.jl (#133)
• Change median and mean export, or update documentation (#146)
• Request: @BProfile macro for profiling (#169)
• Change return or print units (#170)
• @Btime not printing as expected in a loop in Atom (#184)
• No stable documentation (only dev) (#219)
• @benchmarkset seems to be missing cases (#221)
• Incorrect(?) bar heights and plot labels from @benchmark (#246)
• Spurious performance penalty for single-element Union (#262)
• @btime errors because tune! does not execute setup (#264)
• tune! ignores evals=1 of @benchmarkable (#266)
• Number of significant digits in @Btime seems to be an overkill (#288)
• How do you save benchmark results? (#289)
• Silence warnings when using @benchmark (#295)
• Tuning ignores explicit evals parameter (#297)
• Compiler optimization example outdated? (#298)
• Unexpected Behaviour of @btime - cubic splines evaluation (#299)
• @btime has constant runtime whereas @time is dependent on input (#301)
• Automatically create keys in BenchmarkGroup (#308)
• Re-edit docs for multiple setup (#315)
• Add logo? (#316)
• tune! on benchmarkable with evals set. (#328)
• Removing leaves export (#332)
• Remember keyword parameters for tune! (#337)
• [Feature Request] Comparing two functions (#338)
• Feature request: asynchronously build a benchmark group (#345)

v1.3.2

BenchmarkTools v1.3.2

Diff since v1.3.1

Closed issues:

• Display fails with only one sample (#261)
• std and var broken (#293)

Merged pull requests:

• Fix name in comment (#278) (@LilithHafner)
• Fix a few typos in the documentation (#282) (@pitmonticone)
• Fix few broken links (#286) (@Whojo)
• Fix std and var on trials and groups. (#294) (@maleadt)

v1.3.1

BenchmarkTools v1.3.1

Diff since v1.3.0

Merged pull requests:

• Make @btime correctly return the value of the expression in v1.8 (#276) (@giordano)

v1.3.0

BenchmarkTools v1.3.0

Diff since v1.2.2

Closed issues:

• Local variable defined in testset not visible in benchmark macro (#269)
• BenchmarkTools test fails on Windows (#273)

Merged pull requests:

• Update ExecutionTests.jl to work with Windows path (#274) (@strickek)
• Make use of Base.donotdelete if available (#275) (@Keno)

v1.2.2

BenchmarkTools v1.2.2

Diff since v1.2.1

Closed issues:

• Interpolating variables does not work anymore in v1.2.1 (#270)

Merged pull requests:

• Use Expr(:tuple) instead of Expr(:vect) to avoid vect semantics (#271) (@vchuravy)

v1.2.1

BenchmarkTools v1.2.1

Diff since v1.2.0

Closed issues:

• Weird precompilation errors (#252)

Merged pull requests:

• Support manual range for hist display (#255) (@timholy)
• Simplify control logic for :logbins (#257) (@timholy)
• bprofile: disable gcscrub by default (#259) (@timholy)
• Ensure interpolated variables are garbage-collectable (#265) (@ettersi)
• Exclude internals from reference (#267) (@tkf)
• fix dead links in README (#268) (@Moelf)

v1.2.0

BenchmarkTools v1.2.0

Diff since v1.1.4

Merged pull requests:

• Add @bprofile (#254) (@timholy)

v1.1.4

BenchmarkTools v1.1.4

Diff since v1.1.3

Closed issues:

• Confusing numbers in histogram (#249)

Merged pull requests:

• Make sure times are sorted in the at-benchmark histogram (#250) (@giordano)

v1.1.3

BenchmarkTools v1.1.3

Diff since v1.1.2