Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:38444
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Received-SPF: error (pb1.pair.com: domain chiaraquartet.net from 38.99.98.18 cause and error)
Message-ID: <485B5874.9000204@chiaraquartet.net>
Date: Fri, 20 Jun 2008 02:12:52 -0500
User-Agent: Thunderbird 2.0.0.14 (X11/20080505)
MIME-Version: 1.0
To: internals Mailing List <internals@lists.php.net>
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Subject: interesting update on phar performance
From: greg@chiaraquartet.net (Gregory Beaver)

Hi all,

I decided to run my standard phpMyAdmin test without APC enabled and got
startling results from siege:

     Date & Time,  Trans,  Elap Time,  Data Trans,  Resp Time,  Trans
Rate,  Throughput,  Concurrent,    OKAY,   Failed
2008-06-20 02:02:35,    915,      60.01,           1,       0.98,      
15.25,        0.02,       14.88,     915,       0 <-- phpMyAdmin on disk
2008-06-20 02:05:04,    911,      60.04,           1,       0.98,      
15.17,        0.02,       14.86,     911,       0 <-- phpMyAdmin in phar
with phar.cache_list

That's right - they are identical in performance.  With APC, there is a
performance difference (I'm not sure why, to be honest, this one is
really hard to profile):

     Date & Time,  Trans,  Elap Time,  Data Trans,  Resp Time,  Trans
Rate,  Throughput,  Concurrent,    OKAY,   Failed
2008-06-20 01:34:00,   2735,      59.72,           5,       0.33,      
45.80,        0.08,       14.95,    2735,       0 <-- phpMyAdmin on disk
2008-06-20 01:36:11,   2409,      60.14,           5,       0.37,      
40.06,        0.08,       14.95,    2409,       0 <-- phpMyAdmin in phar
with phar.cache_list

However, the difference is negligible.  Thanks to Gopal for the
mini-tutorial on using copy-on-write to implement phar.cache_list, and
kcachegrind for finding the obvious bottlenecks.

Greg