Re: Unexpected OO accessor benchmarks

Am I missing something, or is that just plain weird?

Yes, you are. Try adding this to your script

use Data::Dumper;
print Dumper {
    normal    => $self->a(),
    optimized => $self->b(),
    direct    => $self->c(),
};
[download]

and you will see this:

$VAR1 = {
          'direct' => 123,
          'normal' => 123,
          'optimized' => 123
        };
[download]

Essentially you are benchmarking integers :)

-stvn

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

Replies are listed 'Best First'.
Re^2: Unexpected OO accessor benchmarks by Fletch (Bishop) on Feb 09, 2007 at 20:10 UTC
Yup, what he should have done is pass an anonymous coderef instead: `normal => sub { $self->a() },` [download] Which gives reasonable results: `Rate normal optimized direct normal 662252/s -- -15% -56% optimized 775194/s 17% -- -49% direct 1515152/s 129% 95% --` [download]	[reply] [d/l] [select]
Re^3: Unexpected OO accessor benchmarks by cLive ;-) (Prior) on Feb 09, 2007 at 20:23 UTC
Thank you - I feel much better now. Oh the fun of pre-processing... `Rate normal optimized direct normal 172740/s -- -28% -60% optimized 239252/s 39% -- -45% direct 433898/s 151% 81% --` [download] That looks a little more like what I expected.	[reply] [d/l]
Re^4: Unexpected OO accessor benchmarks by Fletch (Bishop) on Feb 09, 2007 at 20:42 UTC
It's not preprocessing. You pass a hashref to `cmpthese`, the values of which are computed when that line of code is executed. Those values in this case are method invocations which are evaluated where they occur. It should be no more surprising than `some_sub( $obj->method() )` calling that sub with the value returned from the method call.	[reply] [d/l] [select]
Re^5: Unexpected OO accessor benchmarks by cLive ;-) (Prior) on Feb 10, 2007 at 00:55 UTC
Re^6: Unexpected OO accessor benchmarks by eric256 (Parson) on Feb 10, 2007 at 02:29 UTC