view release on metacpan or  search on metacpan

lib/Bench.pm  view on Meta::CPAN

            map { Dumbbench::Instance::PerlSub->new(code => $subs{$_}) }
                keys %subs
        );
        $bench->run;
        $bench->report;

    } else {
        require Benchmark;
        Benchmark::timethese(
            $opts->{n},
            \%subs,
        );
    }

    $bench_called++;
}

END {
    _set_interval;
    say _fmt_num($ti, "s") unless $bench_called || $ENV{HARNESS_ACTIVE};
}

1;
# ABSTRACT: Benchmark running times of Perl code

__END__

=pod

=encoding UTF-8

=head1 NAME

Bench - Benchmark running times of Perl code

=head1 VERSION

This document describes version 0.10 of Bench (from Perl distribution Bench), released on 2014-05-14.

=head1 SYNOPSIS

 # time the whole program
 % perl -MBench -e'...'
 0.0123s

 # basic usage of bench()
 % perl -MBench -e'bench sub { ... }'
 100 calls (58548/s), 0.0017s (0.0171ms/call)

 # get bench result in a variable
 % perl -MBench -E'my $res = bench sub { ... }'

 # specify bench options
 % perl -MBench -E'bench sub { ... }, 100'
 % perl -MBench -E'bench sub { ... }, {n=>-5}'
 304347 calls (60665/s), 5.017s (0.0165ms/call)

 # use Dumbbench as the backend
 % perl -MDumbbench -MBench -E'bench sub { ... }'
 % perl -MBench -E'bench sub { ... }, {dumbbench=>1, dumbbench_options=>{...}}'
 Ran 26 iterations (6 outliers).
 Rounded run time per iteration: 2.9029e-02 +/- 4.8e-05 (0.2%)

 # bench multiple codes
 % perl -MBench -E'bench {a=>sub{...}, b=>sub{...}}, {n=>-2}'
 % perl -MBench -E'bench [sub{...}, sub{...}]'; # automatically named a, b, ...
 b: 100 calls (5357/s), 0.0187s (0.1870ms/call)
 a: 100 calls (12120/s), 0.0083s (0.0825ms/call)
 Fastest is a (2.267x b)

=head1 DESCRIPTION

This module is an alternative interface for L<Benchmark>. It provides some nice
defaults and a simpler interface. There is only one function, B<bench()>, and it
is exported by default. If bench() is never called, the whole program will be
timed.

This module can utilize L<Dumbbench> as the backend instead of L<Benchmark>.

=head1 FUNCTIONS

=head2 bench SUB(S)[, OPTS]

Run Perl code(s) and time it (them). Exported by default. SUB can be a coderef
for specifying a single sub, or hashref/arrayref for specifying multiple subs.

Options are specified in hashref OPTS. Available options:

=over 4

=item * n => INT (default: 100)

Run the code C<n> times, or if negative, until at least C<n> CPU seconds.

=item * dumbbench => BOOL

If 0, do not use L<Dumbbench> even if it is available. If 1, require and use
L<Dumbbench>. If left undef, will use L<Dumbbench> if it is already loaded.

=item * dumbbench_options => HASHREF

Options that will be passed to Dumbbench constructor, e.g.
{target_rel_precision=>0.005, initial_runs=>20}.

=back

=head1 SEE ALSO

L<bench>, command-line interface for this module

L<Benchmark>

L<Dumbbench>

=head1 HOMEPAGE

Please visit the project's homepage at L<https://metacpan.org/release/Bench>.

=head1 SOURCE

Source repository is at L<https://github.com/sharyanto/perl-Bench>.