I'm new to IPB.  I'm not seeing the option anywhere to enable APC, memcache, xcache or etc.  Can someone point me towards it?

In ACP, just search for memcache. Its always the easiest way to find something.

ACP -> System -> Advanced Configuration -> Data Storage 

By default IPS have file cached enabled but you can chose Mysql and any cache method you want 

The data storage method is used by various systems to save arbitrary data. If caching is enabled, using the MySQL Database provides the best performance for most environments, otherwise using the File System is usually best.

Edited May 25, 201510 yr by Ali Majrashi

Thanks, I found it.  Anyone have comments on using Redis vs Memcache vs APC?  I am a fan of Redis in my own development projects, but not sure about IPB performance.

 

E3-1275 v2, 32GB, 2x1TB Samsung 850 Pro R1, Debian 8, kernel 4.0.4, MariaDB 10.0, IPB 4.0.  Dedicated server.  Zero activity on this box except for my test.  Brand new IPB 4.0 installation, there is one post on the forum, default skin.  All tests used the same source server config with webpagetest site.  All caches are local (127.0.0.1).  On each new config, the services were restarted so that the 'first view' would be an accurate representation of an uncached thread being viewed.

 

 

No cache:

 

						Document Complete			Fully Loaded
	Load Time	First Byte	Start Render	Speed Index	DOM Elements	Time	Requests	Bytes In	Time	Requests	Bytes In	Cost
First View	2.482s	0.321s	2.395s	2401	382	2.482s	21	1,020 KB	2.537s	21	1,022 KB	$$$--
Repeat View	0.722s	0.287s	0.724s	701	382	0.722s	1	12 KB	0.722s	1	12 KB

 

 

Redis cache:

 

 

						Document Complete			Fully Loaded
	Load Time	First Byte	Start Render	Speed Index	DOM Elements	Time	Requests	Bytes In	Time	Requests	Bytes In	Cost
First View	2.890s	0.693s	2.891s	2902	382	2.890s	21	1,020 KB	2.945s	21	1,022 KB	$$$--
Repeat View	1.295s	0.586s	1.279s	1302	382	1.295s	1	12 KB	1.295s	1	12 KB

 

APC cache:

 

 

						Document Complete			Fully Loaded
	Load Time	First Byte	Start Render	Speed Index	DOM Elements	Time	Requests	Bytes In	Time	Requests	Bytes In	Cost
First View	2.920s	0.611s	2.795s	2801	382	2.920s	21	1,103 KB	2.983s	21	1,104 KB	$$$--
Repeat View	0.960s	0.559s	0.962s	1001	382	0.960s	1	12 KB	0.960s	1	12 KB

 

Memcached:

 

						Document Complete			Fully Loaded
	Load Time	First Byte	Start Render	Speed Index	DOM Elements	Time	Requests	Bytes In	Time	Requests	Bytes In	Cost
First View	2.929s	0.662s	2.894s	2900	382	2.929s	21	1,020 KB	2.989s	21	1,022 KB	$$$--
Repeat View	0.653s	0.236s	0.655s	700	382	0.653s	1	12 KB	0.653s	1	12 KB

 

These results are interesting.  Aside from the disappointing performance, I also noticed the APC test version generated an extra 83kb of traffic for reasons unknown.  I re-ran it a few more times and couldn't duplicate that behavior.

Does anyone have results that contradict mine?  As it stands, it seems "no cache" produces the fastest consistent load times.  Which is beyond reason... ??

In ACP, just search for memcache. Its always the easiest way to find something.

​I like the search feature.  But even though I have minimum 2 char length, the search box requires 4 characters before it starts producing results.

show variables like 'ft_min_word_len';
+-----------------+-------+
| Variable_name   | Value |
+-----------------+-------+
| ft_min_word_len | 2     |
+-----------------+-------+

 

Thanks, I found it.  Anyone have comments on using Redis vs Memcache vs APC?  I am a fan of Redis in my own development projects, but not sure about IPB performance.

 

​Memcache is the only one where you can enable cache pages for guests. So in the end, is the best one for IPS.

​Memcache is the only one where you can enable cache pages for guests. So in the end, is the best one for IPS.

​The option for guest caching is present on Redis and APC as well, according to my admin cp.

Hmm i think i remember in the betas to be only with memcache. Maybe im wrong or they change it.

 

I wanted to mention the test above was on 4.0.6.1.

One quick question, where can I find the version number of IPB in the admin cp?  I must be blind, I don't see it anywhere!

I wanted to mention the test above was on 4.0.6.1.

One quick question, where can I find the version number of IPB in the admin cp?  I must be blind, I don't see it anywhere!

​http://community.invisionpower.com/files/file/7515-suite-version-on-acp-dashboard/

Interesting test results. I was just setting up a redis container and enabled caching. Less scientific, but it felt slower, maybe I'm getting old. It's only me using it between a couple of incognito sessions...

I've also got a memcache container waiting. Once I've sorted another issue, I'll probably get to throw some load testing on and see. I suspect, the non-cache response time & load will rocket as we head up the RPS chart.

So no cache has the best results?

Does anyone have results that contradict mine?  As it stands, it seems "no cache" produces the fastest consistent load times.  Which is beyond reason... ??

@ctrlbrk What did you use to test/produce report?  I can run it against my test setup as well and see how it compares to yours.

Edited June 20, 201510 yr by tjk

As it stands right now, my settings are:

Data Storage Method: File system
Caching Method: - No caching -

Hi, with my 4.0.8.1 and a 6gb database of a forum since 2005 the memcache performance is awful while apcu is pretty fast.

Hi, with my 4.0.8.1 and a 6gb database of a forum since 2005 the memcache performance is awful while apcu is pretty fast.

Interesting...what are you using for opcache?  Are you running php-fpm too?  In my testing, memcached was the fastest option, it blew away Redis too.  I'm using Zend Opcace, MySQL for "data storage method', and memcached for Caching Method.

Interesting...what are you using for opcache?  Are you running php-fpm too?  In my testing, memcached was the fastest option, it blew away Redis too.  I'm using Zend Opcace, MySQL for "data storage method', and memcached for Caching Method.

php-fpm with 5.6.9's built in opcache on debian 8 with nginx.

 

 

What version of memcached are you using? And what memcache extention for php?

What version of memcached are you using? And what memcache extention for php?

Running on Ubuntu 14.04.2 with the following stock packages:

 

apt-cache showpkg memcached
Package: memcached
Versions:
1.4.14-0ubuntu9 (/var/lib/apt/lists/us.archive.ubuntu.com_ubuntu_dists_trusty_main_binary-amd64_Packages) (/var/lib/dpkg/status)
 
apt-cache showpkg php5-memcached
Package: php5-memcached
Versions:
2.1.0-6build1 (/var/lib/apt/lists/us.archive.ubuntu.com_ubuntu_dists_trusty_universe_binary-amd64_Packages) (/var/lib/dpkg/status)

 

php-fpm with 5.6.9's built in opcache on debian 8 with nginx.

 

 

Would you mind sharing how you are doing your testing?  I can run it against my test install and test the various caching methods, and we can compare apples to apples then!

Thanks!

Running on Ubuntu 14.04.2 with the following stock packages:

 

apt-cache showpkg memcached
Package: memcached
Versions:
1.4.14-0ubuntu9 (/var/lib/apt/lists/us.archive.ubuntu.com_ubuntu_dists_trusty_main_binary-amd64_Packages) (/var/lib/dpkg/status)
 
apt-cache showpkg php5-memcached
Package: php5-memcached
Versions:
2.1.0-6build1 (/var/lib/apt/lists/us.archive.ubuntu.com_ubuntu_dists_trusty_universe_binary-amd64_Packages) (/var/lib/dpkg/status)

 

You are using a Memcached build that is 3 years old. Maybe its the cause of being slow.

Edited June 22, 20159 yr by RevengeFNF

You are using a Memcached build that is 3 years old. Maybe its the cause of being slow.

2 things...

1. I have watched this thread, and have not received emails on any of the updates, sounds like a bug if any IPS admins are watching.

2. I'm not complaining of slowness, quite the opposite, I'm able to get over 700 RPS using apache bench.  As for memcached I've looked for someone packaging newer releases for 14.04 but haven't found them, any pointers here would be appreciated.