PHP APC Performance Tests with Apache ab

Submitted by Mike Milano on September 5, 2008 - 11:00pm.

Today I ran some quick tests with PHP APC using Apache's ab benchmarking tool. The goal was to see how much of a performance gain I could see between PHP APC disabled, then enabled.

So what is op-code caching?

From wikipedia.org: A PHP accelerator is an extension designed to boost the performance of software applications written using the PHP programming language. Most PHP accelerators work by caching the compiled bytecode of PHP scripts to avoid the overhead of parsing and compiling source code on each request (some or all of which may never even be executed). For best performance, caching is to shared memory with direct execution from the shared memory and the minimum of memory copying at runtime. A PHP accelerator typically reduces server load and increases the speed of PHP code anywhere from 2-10 times, depending on factors such as the inherent execution time of the PHP application and the percentage of source code actually executed on a given request. While a code optimizer may even slow down overall performance when used in isolation, it can provide an additional performance boost when coupled with a code cache as the optimization effort is performed just once.

The Testing System
The system is just dev box I have here in the office. These are the stats:

Processor: P4 3Ghz
Memory: 1 Gb
OS: Ubuntu Feisty Fawn
Apache: 2.2.3
PHP: 5.2.1
Target Web App: Drual 5, fresh home page

Without PHP APC: 7.86 Requests per second
Here is the output of ab command with PHP APC disabled.

# ab -n 50 http://localhost/drupal5/
This is ApacheBench, Version 2.0.40-dev <$Revision: 1.146 $> apache-2.0
Copyright 1996 Adam Twiss, Zeus Technology Ltd, http://www.zeustech.net/
Copyright 2006 The Apache Software Foundation, http://www.apache.org/

Benchmarking localhost (be patient).....done


Server Software:        Apache/2.2.3
Server Hostname:        localhost
Server Port:            80

Document Path:          /drupal5/
Document Length:        5144 bytes

Concurrency Level:      1
Time taken for tests:   6.362471 seconds
Complete requests:      50
Failed requests:        0
Write errors:           0
Total transferred:      284150 bytes
HTML transferred:       257200 bytes
Requests per second:    7.86 [#/sec] (mean)
Time per request:       127.249 [ms] (mean)
Time per request:       127.249 [ms] (mean, across all concurrent requests)
Transfer rate:          43.54 [Kbytes/sec] received

Connection Times (ms)
              min  mean[+/-sd] median   max
Connect:        0    0   0.0      0       0
Processing:   125  126   3.5    125     143
Waiting:      125  126   3.5    125     143
Total:        125  126   3.5    125     143

Percentage of the requests served within a certain time (ms)
  50%    125
  66%    126
  75%    126
  80%    126
  90%    131
  95%    134
  98%    143
  99%    143
 100%    143 (longest request)

With PHP APC: 52.92 Requests per second
Here is the output of the Apache ab command with PHP APC enabled.

# ab -n 50 http://localhost/drupal5/
This is ApacheBench, Version 2.0.40-dev <$Revision: 1.146 $> apache-2.0
Copyright 1996 Adam Twiss, Zeus Technology Ltd, http://www.zeustech.net/
Copyright 2006 The Apache Software Foundation, http://www.apache.org/

Benchmarking localhost (be patient).....done


Server Software:        Apache/2.2.3
Server Hostname:        localhost
Server Port:            80


Document Path:          /drupal5/
Document Length:        5144 bytes

Concurrency Level:      1
Time taken for tests:   0.944904 seconds
Complete requests:      50
Failed requests:        0
Write errors:           0
Total transferred:      284150 bytes
HTML transferred:       257200 bytes
Requests per second:    52.92 [#/sec] (mean)
Time per request:       18.898 [ms] (mean)
Time per request:       18.898 [ms] (mean, across all concurrent requests)
Transfer rate:          293.15 [Kbytes/sec] received

Connection Times (ms)
              min  mean[+/-sd] median   max
Connect:        0    0   0.0      0       0
Processing:    18   18   0.3     18      19
Waiting:       18   18   0.0     18      18
Total:         18   18   0.3     18      19

Percentage of the requests served within a certain time (ms)
  50%     18
  66%     18
  75%     18
  80%     18
  90%     19
  95%     19
  98%     19
  99%     19
 100%     19 (longest request)

In Conclusion
As you can see the performance gain in terms of requests per second went up over 600% due to op-code caching that PHP-APC provides.

The downside to PHP APC, and other op-code apps, is that they will cause segmentation faults, although not often. I came across this article from 2bits.com. They have posted a script which will monitor the Apache logs and restart the service when a segmentation fault happens.

In my opinion, that's a small price to pay for such high performance gains.

Coder1

PHP APC Performance Tests with Apache ab

Poll

Random Tip

	Home PHP APC Performance Tests with Apache ab Submitted by Mike Milano on September 5, 2008 - 11:00pm. op-code Performance PHP PHP APC Today I ran some quick tests with PHP APC using Apache's ab benchmarking tool. The goal was to see how much of a performance gain I could see between PHP APC disabled, then enabled. So what is op-code caching? From wikipedia.org: A PHP accelerator is an extension designed to boost the performance of software applications written using the PHP programming language. Most PHP accelerators work by caching the compiled bytecode of PHP scripts to avoid the overhead of parsing and compiling source code on each request (some or all of which may never even be executed). For best performance, caching is to shared memory with direct execution from the shared memory and the minimum of memory copying at runtime. A PHP accelerator typically reduces server load and increases the speed of PHP code anywhere from 2-10 times, depending on factors such as the inherent execution time of the PHP application and the percentage of source code actually executed on a given request. While a code optimizer may even slow down overall performance when used in isolation, it can provide an additional performance boost when coupled with a code cache as the optimization effort is performed just once. The Testing System The system is just dev box I have here in the office. These are the stats: Processor P4 3Ghz Memory 1 Gb OS Ubuntu Feisty Fawn Apache 2.2.3 PHP 5.2.1 Target Web App Drual 5, fresh home page Without PHP APC: 7.86 Requests per second Here is the output of ab command with PHP APC disabled. # ab -n 50 http://localhost/drupal5/ This is ApacheBench, Version 2.0.40-dev <$Revision: 1.146 $> apache-2.0 Copyright 1996 Adam Twiss, Zeus Technology Ltd, http://www.zeustech.net/ Copyright 2006 The Apache Software Foundation, http://www.apache.org/ Benchmarking localhost (be patient).....done Server Software: Apache/2.2.3 Server Hostname: localhost Server Port: 80 Document Path: /drupal5/ Document Length: 5144 bytes Concurrency Level: 1 Time taken for tests: 6.362471 seconds Complete requests: 50 Failed requests: 0 Write errors: 0 Total transferred: 284150 bytes HTML transferred: 257200 bytes Requests per second: 7.86 [#/sec] (mean) Time per request: 127.249 [ms] (mean) Time per request: 127.249 [ms] (mean, across all concurrent requests) Transfer rate: 43.54 [Kbytes/sec] received Connection Times (ms) min mean[+/-sd] median max Connect: 0 0 0.0 0 0 Processing: 125 126 3.5 125 143 Waiting: 125 126 3.5 125 143 Total: 125 126 3.5 125 143 Percentage of the requests served within a certain time (ms) 50% 125 66% 126 75% 126 80% 126 90% 131 95% 134 98% 143 99% 143 100% 143 (longest request) With PHP APC: 52.92 Requests per second Here is the output of the Apache ab command with PHP APC enabled. # ab -n 50 http://localhost/drupal5/ This is ApacheBench, Version 2.0.40-dev <$Revision: 1.146 $> apache-2.0 Copyright 1996 Adam Twiss, Zeus Technology Ltd, http://www.zeustech.net/ Copyright 2006 The Apache Software Foundation, http://www.apache.org/ Benchmarking localhost (be patient).....done Server Software: Apache/2.2.3 Server Hostname: localhost Server Port: 80 Document Path: /drupal5/ Document Length: 5144 bytes Concurrency Level: 1 Time taken for tests: 0.944904 seconds Complete requests: 50 Failed requests: 0 Write errors: 0 Total transferred: 284150 bytes HTML transferred: 257200 bytes Requests per second: 52.92 [#/sec] (mean) Time per request: 18.898 [ms] (mean) Time per request: 18.898 [ms] (mean, across all concurrent requests) Transfer rate: 293.15 [Kbytes/sec] received Connection Times (ms) min mean[+/-sd] median max Connect: 0 0 0.0 0 0 Processing: 18 18 0.3 18 19 Waiting: 18 18 0.0 18 18 Total: 18 18 0.3 18 19 Percentage of the requests served within a certain time (ms) 50% 18 66% 18 75% 18 80% 18 90% 19 95% 19 98% 19 99% 19 100% 19 (longest request) In Conclusion As you can see the performance gain in terms of requests per second went up over 600% due to op-code caching that PHP-APC provides. The downside to PHP APC, and other op-code apps, is that they will cause segmentation faults, although not often. I came across this article from 2bits.com. They have posted a script which will monitor the Apache logs and restart the service when a segmentation fault happens. In my opinion, that's a small price to pay for such high performance gains. Login or register to post comments	Subscribe to coder1.com Click to view this site as XML. Poll What cacherouter plugin do you prefer? APC 50% Memcache 0% DB 50% XCache 0% Eaccellerator 0% Total votes: 2 Login or register to post comments Older polls Random Tip VI Editor: Revert to last save Use the following command to revert to the last saved version of the current file in VI: (make sure you are not in insert mode) :e!
	Home \| Articles \| Screencasts \| Polls \| Contact Copyright © 2008, Mike Milano