LS Cache slower than WP Simple Cache

#1
I've tested SC alone (page cache enabled, memcache for object cache, gzip turned on), LSC alone and even LSC + SC together (with advanced-object unchecked)....and using WP Simple Cache plugin alone is still the fastest in terms of perceived user responsiveness. LSC might in some cases get faster pingdom load times but not always.

It's very unfortunate but I will probably abandon LScache and maybe even try Varnish if I want something more aggressive. I don't understand how a server-level cache wouldn't be able to beat a plugin but it certainly was the case for me. Keep in mind I'm only testing on my staging site and not on my live site where I have at least 30-60 users browsing at all times.

From what i see…it seems that SC is faster because it’s an object cache (saving to user device) with many of the same objects being shared between different pages so it’s easier to get a cache hit. LSC, on the other hand requires that you (or any other user) have visit the exact same page before in order for it to cache. This is why it’s easier to get a cache miss with LSC. I suppose this problem can be easily rectified by using the crawler or releasing it on my live site where it's more likely that users will activate the cache for each other.

So basically:
- SC with client-cached objects and minimal php processing
is faster (and/or smaller downloaded size) than...
- LSC with server-cached full page with no php processing

I am curious to know if there's anything else I might have overlooked that could have prevented LScache from operating at it's full potential.
 

mistwang

LiteSpeed Staff
#2
If you are using "ab " to do the benchmark test, you should use "ab -H 'Accept-Encoding: gzip' ..." to fetch the compress cache data.
Internally, litespeed cache every page in compressed format, without that, server need to decompress the compressed cache for each request.
 
#3
If you are using "ab " to do the benchmark test, you should use "ab -H 'Accept-Encoding: gzip' ..." to fetch the compress cache data.
Internally, litespeed cache every page in compressed format, without that, server need to decompress the compressed cache for each request.
Thank you for the fast reply but I'm not sure I understand what I need to do. Also not sure what you're referring to when you say "ab".
 
#5
I initially clicked around judged with my own eye, and then used pingdom and also asked my programmer to see if he felt the same. The difference is noticeable enough...you can feel one is definitely snappier and more responsive than the other. Uncached initial load, LS is definitely slower...once cached, they are equal at best (LSC usually slower). But if I go away for 20 minutes and come back, LSC is annoyingly slower again on initial load.
 
Last edited:

NiteWave

Administrator
#6
this wiki is for you:
https://www.litespeedtech.com/suppo...ache:lscwp:customizations:multi-cache-plugins
lscache plugin can work with other wordpress cache plugin well, for example w3 total cache.
it should be able to work with simple cache as well, and keep its object cache advantage etc. as the wiki stated, it only replace "full page cache" part in other cache plugin.
if you found lscache plugin not work with simple cache well, or lscache + SC performance is less than SC or any other, please let us know. our lscache plugin is actively developed and has addressed lots of issues user reported here or on wordpress.org.
 
#7
:/
I'm really sorry, NiteWave. I tried that again just now and even incrementally adjusted settings on SC and such. Combining both LSC + SC is definitely slower than SC alone. The best performance is SC alone (even better than LSC alone). I even tried LSC with the crawler on (everything finished crawling) and it's STILL noticeably slower (despite showing equally fast or even faster pingdom times).
 

NiteWave

Administrator
#8
before our lscache developer kick in(their timezone is EDT). can you save your test results in http://www.webpagetest.org/ ?
LSC alone
LSC + SC
so we can compare the results later on.
I think we may not have tested SC before(too many wordpress cache plugin), so we may test it on our lab first, to see if any tricky in it.
I searched out 2 pages for your reference:
https://www.litespeedtech.com/benchmarks/wordpress
https://blog.litespeedtech.com/2015/08/31/getting-the-best-wordpress-performance/
w3 total cache, wp super cache, wp-rocket cache are mentioned, I think it's similar with your case.
I'm also curious about SC: is it really outperforms other competitions ?
 
#9
Ok, I'll do it tomorrow and get you the screenshots as it's late for me now. Believe me, I've seen all the pages on LiteSpeed Cache and comparison benchmarks...that's what made me curious to try it.

Simple Cache is better performance to me but there's also an issue of compatibility. Many of the other cache plugins did not work with all my plugins. W3TC is awfully annoying to set-up and we never found the balance between good performance + good compatibility. Either it's slow, or it's somewhat fast (a little sticky/hesitant) and doesn't work with my logged-in users. WP Super Cache was the similar problem. WP Rocket Cache, we didn't try to be honest. There were some others we tried too like Hyper Cache, etc...and can't remember others.

Ultimately, we decided to try Simple Cache because my programmer said their developer was a really good one from 10up and voila--it worked easily! It doesn't have much settings to play with so there aren't many things that can go wrong. I like that it only caches instead of trying to help you minify and compress images, etc. Please see for yourself.
 

lclarke

Administrator
Staff member
#10
You may already know this, but it's worth mentioning: currently, LSCache for WP doesn't cache pages for logged-in users. If you've been running your tests while logged in, that would explain why it doesn't seem faster to you.
If you haven't already, you should check your response headers and look for "X-LiteSpeed-Cache: hit."
 
#11
Yes....absolutely. I'll log into the site on Chrome, then test on incognito window as well as another browser, as well as web page tests. And yes, I took note of the "HITS" and "MISSES". Also checked with my programmer to make sure he experienced the same.
 
Last edited:
#12
I just did the tests again and they show the same thing. Would any of you guys want access to my staging server to play/test it yourself?
 
#16
Was there any resolution to this? I'm also a big fan of SimpleCache, I use it when a client isn't on LiteSpeed.
I would expect Johnny's original assessment—that it has to do with object caching—might be correct. The litespeed WP plugin does not do anything for object caching at this time, correct?
 

hai

Well-Known Member
#19
Yes you can use LSMCD on server level. On script level, you can search `object cache` plugins. We are supposed to support this in couple weeks soon.
 
Top