ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/AnyEvent/lib/AnyEvent.pm
(Generate patch)

Comparing AnyEvent/lib/AnyEvent.pm (file contents):
Revision 1.61 by root, Fri Apr 25 01:55:25 2008 UTC vs.
Revision 1.70 by root, Fri Apr 25 07:25:44 2008 UTC

… …
369 369
370There is no support for WxWidgets, as WxWidgets has no support for 370There is no support for WxWidgets, as WxWidgets has no support for
371watching file handles. However, you can use WxWidgets through the 371watching file handles. However, you can use WxWidgets through the
372POE Adaptor, as POE has a Wx backend that simply polls 20 times per 372POE Adaptor, as POE has a Wx backend that simply polls 20 times per
373second, which was considered to be too horrible to even consider for 373second, which was considered to be too horrible to even consider for
374AnyEvent. Likewise, other POE backends can be used by Anyevent by using 374AnyEvent. Likewise, other POE backends can be used by AnyEvent by using
375it's adaptor. 375it's adaptor.
376
377AnyEvent knows about L<Prima> and L<Wx> and will try to use L<POE> when
378autodetecting them.
376 379
377=item AnyEvent::detect 380=item AnyEvent::detect
378 381
379Returns C<$AnyEvent::MODEL>, forcing autodetection of the event model 382Returns C<$AnyEvent::MODEL>, forcing autodetection of the event model
380if necessary. You should only call this function right before you would 383if necessary. You should only call this function right before you would
… …
431no warnings; 434no warnings;
432use strict; 435use strict;
433 436
434use Carp; 437use Carp;
435 438
436our $VERSION = '3.2'; 439our $VERSION = '3.3';
437our $MODEL; 440our $MODEL;
438 441
439our $AUTOLOAD; 442our $AUTOLOAD;
440our @ISA; 443our @ISA;
441 444
… …
448 [Coro::Event:: => AnyEvent::Impl::CoroEvent::], 451 [Coro::Event:: => AnyEvent::Impl::CoroEvent::],
449 [EV:: => AnyEvent::Impl::EV::], 452 [EV:: => AnyEvent::Impl::EV::],
450 [Event:: => AnyEvent::Impl::Event::], 453 [Event:: => AnyEvent::Impl::Event::],
451 [Glib:: => AnyEvent::Impl::Glib::], 454 [Glib:: => AnyEvent::Impl::Glib::],
452 [Tk:: => AnyEvent::Impl::Tk::], 455 [Tk:: => AnyEvent::Impl::Tk::],
456 [Wx:: => AnyEvent::Impl::POE::],
457 [Prima:: => AnyEvent::Impl::POE::],
453 [AnyEvent::Impl::Perl:: => AnyEvent::Impl::Perl::], 458 [AnyEvent::Impl::Perl:: => AnyEvent::Impl::Perl::],
454 # everything below here will not be autoprobed as the pureperl backend should work everywhere 459 # everything below here will not be autoprobed as the pureperl backend should work everywhere
455 [Event::Lib:: => AnyEvent::Impl::EventLib::], # too buggy 460 [Event::Lib:: => AnyEvent::Impl::EventLib::], # too buggy
456 [Qt:: => AnyEvent::Impl::Qt::], # requires special main program 461 [Qt:: => AnyEvent::Impl::Qt::], # requires special main program
457 [POE::Kernel:: => AnyEvent::Impl::POE::], # lasciate ogni speranza 462 [POE::Kernel:: => AnyEvent::Impl::POE::], # lasciate ogni speranza
… …
855 $quit->broadcast; 860 $quit->broadcast;
856 }); 861 });
857 862
858 $quit->wait; 863 $quit->wait;
859 864
865
866=head1 BENCHMARK
867
868To give you an idea of the performance and overheads that AnyEvent adds
869over the event loops directly, here is a benchmark of various supported
870event models natively and with anyevent. The benchmark creates a lot of
871timers (with a zero timeout) and io watchers (watching STDOUT, a pty, to
872become writable, which it is), lets them fire exactly once and destroys
873them again.
874
875=head2 Explanation of the columns
876
877I<watcher> is the number of event watchers created/destroyed. Since
878different event models feature vastly different performances, each event
879loop was given a number of watchers so that overall runtime is acceptable
880and similar between tested event loop (and keep them from crashing): Glib
881would probably take thousands of years if asked to process the same number
882of watchers as EV in this benchmark.
883
884I<bytes> is the number of bytes (as measured by the resident set size,
885RSS) consumed by each watcher. This method of measuring captures both C
886and Perl-based overheads.
887
888I<create> is the time, in microseconds (millionths of seconds), that it
889takes to create a single watcher. The callback is a closure shared between
890all watchers, to avoid adding memory overhead. That means closure creation
891and memory usage is not included in the figures.
892
893I<invoke> is the time, in microseconds, used to invoke a simple
894callback. The callback simply counts down a Perl variable and after it was
895invoked "watcher" times, it would C<< ->broadcast >> a condvar once to
896signal the end of this phase.
897
898I<destroy> is the time, in microseconds, that it takes destroy a single
899watcher.
900
901=head2 Results
902
903 name watcher bytes create invoke destroy comment
904 EV/EV 400000 244 0.56 0.46 0.31 EV native interface
905 EV/Any 100000 610 3.52 0.91 0.75 EV + AnyEvent watchers
906 CoroEV/Any 100000 610 3.49 0.92 0.75 coroutines + Coro::Signal
907 Perl/Any 16000 654 4.64 1.22 0.77 pure perl implementation
908 Event/Event 16000 523 28.05 21.38 0.86 Event native interface
909 Event/Any 16000 943 34.43 20.48 1.39 Event + AnyEvent watchers
910 Glib/Any 16000 1357 96.99 12.55 55.51 quadratic behaviour
911 Tk/Any 2000 1855 27.01 66.61 14.03 SEGV with >> 2000 watchers
912 POE/Event 2000 6644 108.15 768.19 14.33 via POE::Loop::Event
913 POE/Select 2000 6343 94.69 807.65 562.69 via POE::Loop::Select
914
915=head2 Discussion
916
917The benchmark does I<not> measure scalability of the event loop very
918well. For example, a select-based event loop (such as the pure perl one)
919can never compete with an event loop that uses epoll when the number of
920file descriptors grows high. In this benchmark, only a single filehandle
921is used (although some of the AnyEvent adaptors dup() its file descriptor
922to worka round bugs).
923
924C<EV> is the sole leader regarding speed and memory use, which are both
925maximal/minimal, respectively. Even when going through AnyEvent, there is
926only one event loop that uses less memory (the C<Event> module natively), and
927no faster event model, not event C<Event> natively.
928
929The pure perl implementation is hit in a few sweet spots (both the
930zero timeout and the use of a single fd hit optimisations in the perl
931interpreter and the backend itself). Nevertheless tis shows that it
932adds very little overhead in itself. Like any select-based backend its
933performance becomes really bad with lots of file descriptors, of course,
934but this was not subjetc of this benchmark.
935
936The C<Event> module has a relatively high setup and callback invocation cost,
937but overall scores on the third place.
938
939C<Glib>'s memory usage is quite a bit bit higher, features a faster
940callback invocation and overall lands in the same class as C<Event>.
941
942The C<Tk> adaptor works relatively well, the fact that it crashes with
943more than 2000 watchers is a big setback, however, as correctness takes
944precedence over speed. Nevertheless, its performance is surprising, as the
945file descriptor is dup()ed for each watcher. This shows that the dup()
946employed by some adaptors is not a big performance issue (it does incur a
947hidden memory cost inside the kernel, though).
948
949C<POE>, regardless of backend (wether using its pure perl select-based
950backend or the Event backend) shows abysmal performance and memory
951usage: Watchers use almost 30 times as much memory as EV watchers, and 10
952times as much memory as both Event or EV via AnyEvent. Watcher invocation
953is almost 700 times slower as with AnyEvent's pure perl implementation.
954
955Summary: using EV through AnyEvent is faster than any other event
956loop. The overhead AnyEvent adds can be very small, and you should avoid
957POE like the plague if you want performance or reasonable memory usage.
958
959
860=head1 FORK 960=head1 FORK
861 961
862Most event libraries are not fork-safe. The ones who are usually are 962Most event libraries are not fork-safe. The ones who are usually are
863because they are so inefficient. Only L<EV> is fully fork-aware. 963because they are so inefficient. Only L<EV> is fully fork-aware.
864 964
865If you have to fork, you must either do so I<before> creating your first 965If you have to fork, you must either do so I<before> creating your first
866watcher OR you must not use AnyEvent at all in the child. 966watcher OR you must not use AnyEvent at all in the child.
967
867 968
868=head1 SECURITY CONSIDERATIONS 969=head1 SECURITY CONSIDERATIONS
869 970
870AnyEvent can be forced to load any event model via 971AnyEvent can be forced to load any event model via
871$ENV{PERL_ANYEVENT_MODEL}. While this cannot (to my knowledge) be used to 972$ENV{PERL_ANYEVENT_MODEL}. While this cannot (to my knowledge) be used to
… …
879 980
880 BEGIN { delete $ENV{PERL_ANYEVENT_MODEL} } 981 BEGIN { delete $ENV{PERL_ANYEVENT_MODEL} }
881 982
882 use AnyEvent; 983 use AnyEvent;
883 984
985
884=head1 SEE ALSO 986=head1 SEE ALSO
885 987
886Event modules: L<Coro::EV>, L<EV>, L<EV::Glib>, L<Glib::EV>, 988Event modules: L<Coro::EV>, L<EV>, L<EV::Glib>, L<Glib::EV>,
887L<Coro::Event>, L<Event>, L<Glib::Event>, L<Glib>, L<Coro>, L<Tk>, 989L<Coro::Event>, L<Event>, L<Glib::Event>, L<Glib>, L<Coro>, L<Tk>,
888L<Event::Lib>, L<Qt>, L<POE>. 990L<Event::Lib>, L<Qt>, L<POE>.
… …
892L<AnyEvent::Impl::Tk>, L<AnyEvent::Impl::Perl>, L<AnyEvent::Impl::EventLib>, 994L<AnyEvent::Impl::Tk>, L<AnyEvent::Impl::Perl>, L<AnyEvent::Impl::EventLib>,
893L<AnyEvent::Impl::Qt>, L<AnyEvent::Impl::POE>. 995L<AnyEvent::Impl::Qt>, L<AnyEvent::Impl::POE>.
894 996
895Nontrivial usage examples: L<Net::FCP>, L<Net::XMPP2>. 997Nontrivial usage examples: L<Net::FCP>, L<Net::XMPP2>.
896 998
999
897=head1 AUTHOR 1000=head1 AUTHOR
898 1001
899 Marc Lehmann <schmorp@schmorp.de> 1002 Marc Lehmann <schmorp@schmorp.de>
900 http://home.schmorp.de/ 1003 http://home.schmorp.de/
901 1004

Diff Legend

– Removed lines
+ Added lines
< Changed lines
> Changed lines