Re: Benchmark for echoing via comma, concat, inteprolation, andheredoc.

Robert Cummings wrote:
> That echo benchmark though... WTF!

Yup similar numbers here tho' not quite as staggering on my setup:

PHP 5.2.6 with Suhosin-Patch 0.9.6.2 (cli) (built: Aug 25 2008 10:09:21)
Copyright (c) 1997-2008 The PHP Group
Zend Engine v2.2.0, Copyright (c) 1998-2008 Zend Technologies

(Tho' Suhosin is disabled).

To summerise, the echo method took about twice as long as the others.


shell:~> time php -qC test.echo.php >/dev/null
8.52user 8.23system 0:16.97elapsed 98%CPU (0avgtext+0avgdata 0maxresident)k
608inputs+0outputs (2major+2721minor)pagefaults 0swaps

shell:~> time php -qC test.concat.php >/dev/null
7.16user 1.70system 0:08.94elapsed 99%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (0major+2724minor)pagefaults 0swaps

shell:~> time php -qC test.interpol.php >/dev/null
5.84user 1.76system 0:07.62elapsed 99%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (0major+2724minor)pagefaults 0swaps

shell:~> time php -qC test.heredoc.php >/dev/null
5.52user 1.73system 0:07.29elapsed 99%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (0major+2723minor)pagefaults 0swaps

shell:~> time php -qC test.echo.php >/dev/null
8.21user 8.08system 0:16.44elapsed 99%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (0major+2723minor)pagefaults 0swaps

shell:~> time php -qC test.concat.php >/dev/null
7.29user 1.66system 0:08.98elapsed 99%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (0major+2724minor)pagefaults 0swaps

shell:~> time php -qC test.interpol.php >/dev/null
6.34user 1.79system 0:08.15elapsed 99%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (0major+2723minor)pagefaults 0swaps

shell:~> time php -qC test.heredoc.php >/dev/null
6.10user 1.67system 0:07.79elapsed 99%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (0major+2724minor)pagefaults 0swaps



And just for sh1ts and giggles, I tried printf... which although I
expected to be slow, it still turned out faster than the echo!!!


shell:~> time php -qC test.printf.php >/dev/null
11.85user 1.67system 0:13.63elapsed 99%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (0major+2727minor)pagefaults 0swaps

shell:~> time php -qC test.printf.php >/dev/null
12.27user 1.69system 0:13.99elapsed 99%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (0major+2728minor)pagefaults 0swaps


--

Colin Guthrie
gmane(at)colin.guthr.ie
http://colin.guthr.ie/

Day Job:
Tribalogic Limited [http://www.tribalogic.net/]
Open Source:
Mandriva Linux Contributor [http://www.mandriva.com/]
PulseAudio Hacker [http://www.pulseaudio.org/]
Trac Hacker [http://trac.edgewall.org/]

On Wed, 2008-08-27 at 18:02 +0100, Colin Guthrie wrote:
> Robert Cummings wrote:
> > That echo benchmark though... WTF!
>
> Yup similar numbers here tho' not quite as staggering on my setup:

Hah, your values led me to think about it... then I realized I've been
playing around with Xen lately and was currently running the Xen kernel.
This was having a huge performance penalty on my system... which is odd
since I have a Quad Phenom Black Edition (must be at the kernel level
then since my CPU load is miniscule). Anyways, rebooting to the non Xen
kernel I get the following numbers which reflect the same distribution
but are MUCH faster lol.

shell:~> time php5 -qC test.echo.php > /dev/null

real 0m10.622s
user 0m5.380s
sys 0m4.680s

shell:~> time php5 -qC test.concat.php > /dev/null

real 0m4.875s
user 0m3.990s
sys 0m0.870s

shell:~> time php5 -qC test.interpol.php > /dev/null

real 0m4.799s
user 0m3.740s
sys 0m1.050s

shell:~> time php5 -qC test.heredoc.php > /dev/null

real 0m4.901s
user 0m3.750s
sys 0m1.070s


Cheers,
Rob.
--
http://www.interjinn.com
Application and Templating Framework for PHP