NAME

Bencher::Scenario::TimeHiRes::time - Benchmark getting current time

VERSION

This document describes version 0.003 of Bencher::Scenario::TimeHiRes::time (from Perl distribution Bencher-Scenarios-TimeHiRes), released on 2018-12-21.

SYNOPSIS

To run benchmark with default option:

% bencher -m TimeHiRes::time

To run module startup overhead benchmark:

% bencher --module-startup -m TimeHiRes::time

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.

Time::HiRes 1.9741

BENCHMARK PARTICIPANTS

  • Time::HiRes::time (perl_code)

    Function call template:

    Time::HiRes::time
  • CORE::time (perl_code)

    Code template:

    time()
  • Time::HiRes::tv_interval (perl_code)

    Function call template:

    Time::HiRes::tv_interval

SAMPLE BENCHMARK RESULTS

Run on: perl: v5.26.0, CPU: Intel(R) Core(TM) i5-2400 CPU @ 3.10GHz (4 cores), OS: GNU/Linux LinuxMint version 18.2, OS kernel: Linux version 4.8.0-53-generic.

Benchmark with default options (bencher -m TimeHiRes::time):

#table1#
+--------------------------+-----------+-----------+------------+---------+---------+
| participant              | rate (/s) | time (ns) | vs_slowest |  errors | samples |
+--------------------------+-----------+-----------+------------+---------+---------+
| Time::HiRes::tv_interval |   1200000 |     850   |        1   | 1.7e-09 |      20 |
| Time::HiRes::time        |  12300000 |      81.6 |       10.5 | 6.9e-11 |      20 |
| CORE::time               | 200000000 |       6   |      100   | 2.1e-10 |      20 |
+--------------------------+-----------+-----------+------------+---------+---------+

Benchmark module startup overhead (bencher -m TimeHiRes::time --module-startup):

#table2#
+---------------------+-----------+------------------------+------------+---------+---------+
| participant         | time (ms) | mod_overhead_time (ms) | vs_slowest |  errors | samples |
+---------------------+-----------+------------------------+------------+---------+---------+
| Time::HiRes         |        12 |                      7 |        1   | 2.8e-05 |      20 |
| perl -e1 (baseline) |         5 |                      0 |        2.4 | 3.8e-05 |      20 |
+---------------------+-----------+------------------------+------------+---------+---------+

To display as an interactive HTML table on a browser, you can add option --format html+datatables.

BENCHMARK NOTES

Use floating-point time(), because tv_interval() is about 1 order of magnitude slower!

Ref: David Golden's TPC 2017 talk "Real World Optimization" https://www.youtube.com/watch?v=_PJIVVGAZqA

HOMEPAGE

Please visit the project's homepage at https://metacpan.org/release/Bencher-Scenarios-TimeHiRes.

SOURCE

Source repository is at https://github.com/perlancar/perl-Bencher-Scenarios-TimeHiRes.

BUGS

Please report any bugs or feature requests on the bugtracker website https://rt.cpan.org/Public/Dist/Display.html?Name=Bencher-Scenarios-TimeHiRes

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 <perlancar@cpan.org>

COPYRIGHT AND LICENSE

This software is copyright (c) 2018, 2017, 2016 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.