NAME Bencher::Scenario::CSVParsingModules - Benchmark CSV parsing modules VERSION This document describes version 0.002 of Bencher::Scenario::CSVParsingModules (from Perl distribution Bencher-Scenario-CSVParsingModules), released on 2021-07-31. SYNOPSIS To run benchmark with default option: % bencher -m CSVParsingModules To run module startup overhead benchmark: % bencher --module-startup -m CSVParsingModules For more options (dump scenario, list/include/exclude/add participants, list/include/exclude/add datasets, etc), see bencher or run "bencher --help". DESCRIPTION Packaging a benchmark script as a Bencher scenario makes it convenient to include/exclude/add participants/datasets (either via CLI or Perl code), send the result to a central repository, among others . See Bencher and bencher (CLI) for more details. BENCHMARKED MODULES Version numbers shown below are the versions used when running the sample benchmark. Text::CSV_PP 2.01 Text::CSV_XS 1.46 BENCHMARK PARTICIPANTS * Text::CSV_PP (perl_code) Code template: my $csv = Text::CSV_PP->new({binary=>1}); open my $fh, "<", ; my $rows = []; while (my $row = $csv->getline($fh)) { push @$rows, $row } * Text::CSV_XS (perl_code) Code template: my $csv = Text::CSV_XS->new({binary=>1}); open my $fh, "<", ; my $rows = []; while (my $row = $csv->getline($fh)) { push @$rows, $row } * naive-split (perl_code) Code template: open my $fh, "<", ; my $rows = []; while (defined(my $row = <$fh>)) { chomp $row; push @$rows, [split /,/, $row] } BENCHMARK DATASETS * bench-100x100.csv * bench-10x10.csv * bench-1x1.csv * bench-5x5.csv BENCHMARK SAMPLE RESULTS Sample benchmark #1 Run on: perl: *v5.34.0*, CPU: *Intel(R) Core(TM) i7-4770 CPU @ 3.40GHz (4 cores)*, OS: *GNU/Linux LinuxMint version 19*, OS kernel: *Linux version 5.3.0-68-generic*. Benchmark command (default options): % bencher -m CSVParsingModules Result formatted as table (split, part 1 of 4): #table1# {dataset=>"bench-100x100.csv"} | participant | rate (/s) | time (ms) | pct_faster_vs_slowest | pct_slower_vs_fastest | errors | samples | |--------------+-----------+-----------+-----------------------+-----------------------+---------+---------| | Text::CSV_PP | 32.7 | 30.6 | 0.00% | 2073.95% | 2e-05 | 20 | | Text::CSV_XS | 640 | 1.6 | 1854.86% | 11.21% | 6.2e-06 | 20 | | naive-split | 710 | 1.4 | 2073.95% | 0.00% | 2.5e-06 | 20 | The above result formatted in Benchmark.pm style: Rate Text::CSV_PP Text::CSV_XS naive-split Text::CSV_PP 32.7/s -- -94% -95% Text::CSV_XS 640/s 1812% -- -12% naive-split 710/s 2085% 14% -- Legends: Text::CSV_PP: participant=Text::CSV_PP Text::CSV_XS: participant=Text::CSV_XS naive-split: participant=naive-split The above result presented as chart: Result formatted as table (split, part 2 of 4): #table2# {dataset=>"bench-10x10.csv"} | participant | rate (/s) | time (μs) | pct_faster_vs_slowest | pct_slower_vs_fastest | errors | samples | |--------------+-----------+-----------+-----------------------+-----------------------+---------+---------| | Text::CSV_PP | 1600 | 620 | 0.00% | 2364.58% | 1.1e-06 | 20 | | Text::CSV_XS | 19000 | 54 | 1061.89% | 112.12% | 1.1e-07 | 20 | | naive-split | 39000 | 25 | 2364.58% | 0.00% | 5.3e-08 | 20 | The above result formatted in Benchmark.pm style: Rate Text::CSV_PP Text::CSV_XS naive-split Text::CSV_PP 1600/s -- -91% -95% Text::CSV_XS 19000/s 1048% -- -53% naive-split 39000/s 2380% 116% -- Legends: Text::CSV_PP: participant=Text::CSV_PP Text::CSV_XS: participant=Text::CSV_XS naive-split: participant=naive-split The above result presented as chart: Result formatted as table (split, part 3 of 4): #table3# {dataset=>"bench-1x1.csv"} | participant | rate (/s) | time (μs) | pct_faster_vs_slowest | pct_slower_vs_fastest | errors | samples | |--------------+-----------+-----------+-----------------------+-----------------------+---------+---------| | Text::CSV_PP | 7500 | 130 | 0.00% | 2000.62% | 2.7e-07 | 20 | | Text::CSV_XS | 42700 | 23.4 | 466.40% | 270.87% | 2e-08 | 20 | | naive-split | 160000 | 6.3 | 2000.62% | 0.00% | 1.3e-08 | 20 | The above result formatted in Benchmark.pm style: Rate Text::CSV_PP Text::CSV_XS naive-split Text::CSV_PP 7500/s -- -82% -95% Text::CSV_XS 42700/s 455% -- -73% naive-split 160000/s 1963% 271% -- Legends: Text::CSV_PP: participant=Text::CSV_PP Text::CSV_XS: participant=Text::CSV_XS naive-split: participant=naive-split The above result presented as chart: Result formatted as table (split, part 4 of 4): #table4# {dataset=>"bench-5x5.csv"} | participant | rate (/s) | time (μs) | pct_faster_vs_slowest | pct_slower_vs_fastest | errors | samples | |--------------+-----------+-----------+-----------------------+-----------------------+---------+---------| | Text::CSV_PP | 3370 | 296 | 0.00% | 2429.52% | 2.5e-07 | 22 | | Text::CSV_XS | 33000 | 30 | 878.84% | 158.42% | 5.3e-08 | 20 | | naive-split | 85400 | 11.7 | 2429.52% | 0.00% | 3.2e-09 | 22 | The above result formatted in Benchmark.pm style: Rate Text::CSV_PP Text::CSV_XS naive-split Text::CSV_PP 3370/s -- -89% -96% Text::CSV_XS 33000/s 886% -- -61% naive-split 85400/s 2429% 156% -- Legends: Text::CSV_PP: participant=Text::CSV_PP Text::CSV_XS: participant=Text::CSV_XS naive-split: participant=naive-split The above result presented as chart: Sample benchmark #2 Benchmark command (benchmarking module startup overhead): % bencher -m CSVParsingModules --module-startup Result formatted as table: #table5# | participant | time (ms) | mod_overhead_time | pct_faster_vs_slowest | pct_slower_vs_fastest | errors | samples | |---------------------+-----------+-------------------+-----------------------+-----------------------+---------+---------| | Text::CSV_PP | 20 | 13.9 | 0.00% | 233.89% | 9.8e-05 | 20 | | Text::CSV_XS | 17 | 10.9 | 21.13% | 175.64% | 5.4e-05 | 20 | | perl -e1 (baseline) | 6.1 | 0 | 233.89% | 0.00% | 5.3e-05 | 20 | The above result formatted in Benchmark.pm style: Rate T:C_P T:C_X perl -e1 (baseline) T:C_P 50.0/s -- -15% -69% T:C_X 58.8/s 17% -- -64% perl -e1 (baseline) 163.9/s 227% 178% -- Legends: T:C_P: mod_overhead_time=13.9 participant=Text::CSV_PP T:C_X: mod_overhead_time=10.9 participant=Text::CSV_XS perl -e1 (baseline): mod_overhead_time=0 participant=perl -e1 (baseline) The above result presented as chart: To display as an interactive HTML table on a browser, you can add option "--format html+datatables". CONTRIBUTOR perlancar (on pc-office) HOMEPAGE Please visit the project's homepage at . SOURCE Source repository is at . BUGS Please report any bugs or feature requests on the bugtracker website When submitting a bug or request, please include a test-file or a patch to an existing test-file that illustrates the bug or desired feature. AUTHOR perlancar COPYRIGHT AND LICENSE This software is copyright (c) 2021, 2019 by perlancar@cpan.org. This is free software; you can redistribute it and/or modify it under the same terms as the Perl 5 programming language system itself.