bench2 results
Marc Lehmann
schmorp at schmorp.de
Thu Dec 2 06:41:12 CET 2010
On Thu, Dec 02, 2010 at 06:24:35AM +0100, Marc Lehmann <schmorp at schmorp.de> wrote:
> > > use POE::Kernel { loop => 'POE::XS::Loop::EPoll' };
Anyways, I did it, and here are the results with everything from current
CPAN (the POE::Loop::EV test still hangs, so not included):
name sockets create request
EV 16 14.75 4.56
Perl 16 17.61 10.82
POE 16 173.87 155.32 XS::Loop::Poll
POE 16 177.62 153.15 XS::Loop::EPoll
POE 16 200.87 155.84 Loop::Event
POE 16 175.74 153.69 Loop::Select
name sockets create request
EV 20000 71.00 8.27
Perl 20000 75.64 32.58
POE 20000 278.48 5387.89 XS::Loop::Poll
POE 20000 275.60 5344.13 XS::Loop::EPoll
POE 20000 300.43 5999.13 Loop::Event
POE 20000 277.34 5366.55 Loop::Select
I had POE POE::XS::Loop::EPoll POE::XS::Loop::Poll POE::XS::Queue::Array
POE::Loop::Event POE::Loop::EV installed, and I assume
POE::XS::Queue::Array is picked up automatically.
I verified that EPoll is indeed being used, so the POE backend seems to be
of little relevance.
Just for fun, here are results after deleting POE::XS::Queue::Array, so
seems it indeed has a speedup effect on watcher creation:
name sockets create request
EV 16 14.63 4.56
Perl 16 17.49 10.78
POE 16 191.75 170.74
POE 16 191.00 170.99
POE 16 223.99 174.42
POE 16 187.87 172.68
name sockets create request
EV 20000 70.62 8.34
Perl 20000 75.78 32.66
POE 20000 289.91 5399.23
POE 20000 286.87 5426.24
POE 20000 311.88 5659.91
POE 20000 288.66 5467.37
--
The choice of a Deliantra, the free code+content MORPG
-----==- _GNU_ http://www.deliantra.net
----==-- _ generation
---==---(_)__ __ ____ __ Marc Lehmann
--==---/ / _ \/ // /\ \/ / schmorp at schmorp.de
-=====/_/_//_/\_,_/ /_/\_\
More information about the anyevent
mailing list