Welcome to Perl::Formance.
This is about benchmarking Perl.
It combines several projects: the actual benchmark suite, tools to bootstrap a complete Perl+CPAN from git, a results database and chart rendering.
Where are the results?
Click here for Perl 5 results:
For your local use try the whole mix in one go with:
curl -Ls https://raw.githubusercontent.com/renormalist/app-bootstrap-perl/master/bin/bootstrap-perl | perl - --perlformance
This will by default build a "blead" Perl in
~/.bootstrapperl/, install CPAN with dependencies, run benchmarks and print out the results on screen at the end.
--perlformance you can provide other versions
any rev that "git" understands.
The Benchmark Storage
benchmarkanything.org has more info about the used benchmark result abstraction.
The Benchmark Suite
Benchmark::Perl::Formance is the framework for running benchmarks. It requires just a few dependencies and runs a basic set of benchmarks.
Task::PerlFormance provides the full set of benchmarks via dependencies from CPAN.
- You have a code snippet worth to measure over many Perl versions? Create a ticket with details, or even better a Pull Request.
- You have a server from where you want to regularly report results to my central results DB? Contact me.
App::Bootstrap::Perl is yet another Perl builder. It focuses on
- building from git
- set up CPAN.pm with distroprefs for unattended installation
- lightweight work-arounds for known build problems (e.g., for old 5.8/5.9 Perl versions)
- inject meta information into the built Perl to augment the benchmark results
- optionally start programs from the built Perl (in particular the benchmark suite)
- Overall vision at YAPC::EU 2011
- Early numbers at YAPC::EU 2012
- Rekapitulacija 2015 at YAPC::EU 2015
- Rekapitulacija 2019 at PerlCon 2019