Nicholas Clark wrote:
> [patch 28974] and this with -O2
>
> Rate str tid
> str 500/s -- -50%
> tid 1009/s 102% --
Great job on this, Nick. Thanks.
> So the overloading code is still slowing things down for
> non-overloaded stringification, but not as badly.
It's not the use of overload.pm that is still the issue.
It's just that generating the Class=TYPE(0xADDR) string
takes so much time. Your patch puts stringification of
overloaded and non-overloaded objects nearly on a par.
However, both still pale in comparison to getting a TID:
use strict;
use warnings;
no warnings 'void';
use Benchmark qw(cmpthese);
use threads;
my $thr = threads->create(sub {});
my $obj = bless(\(my $x=1), 'foo_bar');
cmpthese( -1, {
'obj' => sub { ''.$obj for (1..1000) },
'thr' => sub { ''.$thr for (1..1000) },
'tid' => sub { ''.$thr->tid for (1..1000) }
});
$thr->join();
Results:
Rate thr obj tid
thr 131/s -- -2% -63%
obj 133/s 2% -- -62%
tid 355/s 171% 166% --
Changing the above to "use threads 'stringify';", and
the results are:
Rate obj thr tid
obj 133/s -- -59% -62%
thr 328/s 146% -- -7%
tid 352/s 164% 7% --
which shows the 7% overload overhead, but also shows
that ""-overloading is so much better than C=T(0xA)
stringification.
__________________________________________________
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com
