A simple Julia package to plot performance and data profiles.

Citation	Documentation	Linux/macOS/Windows/FreeBSD	Coverage

This package contains Julia translations of original scripts by Elizabeth Dolan, Jorge Moré and Stefan Wild. See http://www.mcs.anl.gov/~wild/dfo/benchmarking.

The original code was not accompanied by an open-source license. Jorge Moré and Stefan Wild have kindly provided their consent in writing to allow distribution of this Julia translation. See the consent folder for a full transcription.

It appears that performance profiles date back to at least 1996!

Watch out for the pitfalls of profiles!

julia> Pkg.add("BenchmarkProfiles")

No plotting backend is loaded by default so the user can choose among several available plotting backends. Currently, Plots.jl and UnicodePlots.jl are supported. Backends become available when the corresponding package is imported.

julia> using BenchmarkProfiles
julia> T = 10 * rand(25,3);  # 25 problems, 3 solvers
julia> performance_profile(PlotsBackend(), T, ["Solver 1", "Solver 2", "Solver 3"], title="Celebrity Deathmatch")
ERROR: ArgumentError: The backend PlotsBackend() is not loaded. Please load the corresponding AD package.
julia> using Plots
julia> performance_profile(PlotsBackend(), T, ["Solver 1", "Solver 2", "Solver 3"], title="Celebrity Deathmatch")  # Success!

• A. L. Tits and Y. Yang, Globally convergent algorithms for robust pole assignment by state feedback, IEEE Transactions on Automatic Control, 41(10), pages 1432–1452, 1996. DOI 10.1109/9.539425.
• E. Dolan and J. Moré, Benchmarking Optimization Software with Performance Profiles, Mathematical Programming 91, pages 201–213, 2002. DOI 10.1007/s101070100263.
• J. J. Moré and S. M. Wild, Benchmarking Derivative-Free Optimization Algorithms, SIAM Journal on Optimization, 20(1), pages 172–191, 2009. DOI 10.1137/080724083.

If you use BenchmarkProfiles.jl in your work, please cite using the format given in CITATION.bib.