Re^2: Unexpected OO accessor benchmarks

I read your node - interesting, thanks. But I can't get results like yours. Could it be because the number of iterations in your test are too low?

Putting get_value() and get_value2() declarations at the beginning of program after the "my $self" line, I get:

              Rate    normal optimized    direct
normal    174625/s        --      -27%      -60%
optimized 240602/s       38%        --      -45%
direct    435374/s      149%       81%        --
[download]

Moving the subs to just above sub a() declaration, I get

              Rate    normal optimized    direct
normal    172275/s        --      -28%      -61%
optimized 239252/s       39%        --      -46%
direct    444444/s      158%       86%        --
[download]

I've tried a few more re-arrangements of the code layout, and I can't get the reversal that you did. All re-arrangements I've tried have 'direct' at 145-155%. Any chance you can post the code so I can duplicate the results? Logically, I would think that asking for the value of an arrayref element is going to be faster than calling a sub to do so, so the benchmarks appear to be logical to me.

Comment on Re^2: Unexpected OO accessor benchmarks Select or Download Code

Replies are listed 'Best First'.
Re^3: Unexpected OO accessor benchmarks by chromatic (Archbishop) on Feb 10, 2007 at 02:17 UTC
I revised the code slightly: `use Benchmark 'cmpthese'; sub get_value { my $self = shift; return $self->[0]; } sub get_value2 { $_[0]->[0]; } my $self = bless [123], '::main'; cmpthese( 16000000, { normal => sub { $self->get_value() }, optimized => sub { $self->get_value2() }, direct => sub { $self->[0] }, }); exit(0); $ perl oo_benchmark.pl Rate normal optimized direct normal 1462523/s -- -21% -90% optimized 1858304/s 27% -- -88% direct 15238095/s 942% 720% --` [download] Yes, it took 16 million iterations to get an accurate reading on the direct attribute access. With a variant of your original benchmark modified to avoid the double-dispatch penalty, I get: `use Benchmark 'cmpthese'; my $self = bless [123], '::main'; cmpthese( 16000000, { normal => sub { $self->get_value() }, optimized => sub { $self->get_value2() }, direct => sub { $self->[0] }, }); exit(0); sub get_value { my $self = shift; return $self->[0]; } sub get_value2 { $_[0]->[0]; } $ perl oo_benchmark.pl Rate normal optimized direct normal 1725998/s -- -14% -94% optimized 2007528/s 16% -- -93% direct 29629630/s 1617% 1376% --` [download] That represents speedups of around 20%, 11%, and 100%, respectively. I can make the case for a 10% speedup perhaps, but the others are way outside anything I can explain. Thus I'm not sure this benchmark is useful.	[reply] [d/l] [select]

Replies are listed 'Best First'.

Re^3: Unexpected OO accessor benchmarks
by chromatic (Archbishop) on Feb 10, 2007 at 02:17 UTC

I revised the code slightly:

use Benchmark 'cmpthese';

sub get_value {
        my $self = shift;
        return $self->[0];
}

sub get_value2 {
        $_[0]->[0];
}

my $self = bless [123], '::main';

cmpthese( 16000000, {
        normal    => sub { $self->get_value() },
        optimized => sub { $self->get_value2() },
        direct    => sub { $self->[0] },
});

exit(0);

$ perl oo_benchmark.pl 
                Rate    normal optimized    direct
normal     1462523/s        --      -21%      -90%
optimized  1858304/s       27%        --      -88%
direct    15238095/s      942%      720%        --
[download]

Yes, it took 16 million iterations to get an accurate reading on the direct attribute access. With a variant of your original benchmark modified to avoid the double-dispatch penalty, I get:

use Benchmark 'cmpthese';

my $self = bless [123], '::main';

cmpthese( 16000000, {
        normal    => sub { $self->get_value() },
        optimized => sub { $self->get_value2() },
        direct    => sub { $self->[0] },
});

exit(0);

sub get_value {
        my $self = shift;
        return $self->[0];
}

sub get_value2 {
        $_[0]->[0];
}

$ perl oo_benchmark.pl 
                Rate    normal optimized    direct
normal     1725998/s        --      -14%      -94%
optimized  2007528/s       16%        --      -93%
direct    29629630/s     1617%     1376%        --
[download]

That represents speedups of around 20%, 11%, and 100%, respectively. I can make the case for a 10% speedup perhaps, but the others are way outside anything I can explain. Thus I'm not sure this benchmark is useful.

[reply]
[d/l]
[select]