Jump to content

Performance Minify and Compress decrease page load speed


Recommended Posts

Hi,

 

I've been trying to increase performance of an 1.5.4.1. shop and noticed that turning -on- 'Minify HTML' slows down the page loading time by about 2 seconds, from under 2 seconds to over 4 seconds.

 

Turning on 'Compress inline JavaScript' adds another .4 seconds to the page loading time.

 

My other settings are currently:

'Smarty Recompile' never.

'Smarty Cache' on.

'Smart cache for CSS' on.

'Smart cache for Javascript' on.

'Apache optimization' on.

 

I tried clearing the smarty cache and recompiling smarty after I've checked 'Minify' to on, but it does not change anything and the same performance drop occurs.

 

 

Any ideas anyone?

Link to comment
Share on other sites

  • 3 months later...

I started out with a WAMP server based on Coral (UniServer), currently I have another installation on a XAMPP server and one on a Red Hat LAMP stacked cloud server.

 

On all these installations I'm measuring about 20%-30% page load drop when 'Minify HTML' is turned on.

Turning on 'Compress Inline Javascript' drops the page load speed by another 2 or 3 percent.

 

On the XAMPP server page loads drop from 3.05 seconds to 4.33 seconds avg.

On the LAMP server page loads drop from 1.09 seconds to 1.36 seconds avg.

 

I'm kind of assuming that this 'feature' (?) only should be enabled if proper caching is also enabled, but it's a wild guess.

Link to comment
Share on other sites

I have also noticed that the minify html adds to the processing time. Personally I think this is a trade off. When you add more processing to the files, it adds more time to processing. 

 

At the same time I wouldn't trust any tests on the localhost unless you are running in a container environment. 

Link to comment
Share on other sites

@Dh24

Could you please elaborate? I'm confused now...

 

I have also noticed that the minify html adds to the processing time. Personally I think this is a trade off. When you add more processing to the files, it adds more time to processing. 

 

At the same time I wouldn't trust any tests on the localhost unless you are running in a container environment. 

 

If turning these options on is a trade off, then what is the benefit? If a drop in page load speed is the result, it seems kind of out of place to me to put these options on the 'Performance' page.

 

Why not test on localhost (assuming you mean a dev machine)? If properly set up, any machine can be used for testing. I've tested on one dev machine and two production servers. On all of them the performance drop occurs.

Link to comment
Share on other sites

The trade off is smaller file size to be downloaded from the server, as compared with the server responding faster. Depending on what you are after depends on how you use it.

 

Testing on a dev machine is fine, but not speed testing. You cannot accurately judge the loading time of a website running off of a desktop computer. You have things like monitors attached, video cards, and other things that will slow the system down. Plus more than likely you are using a home class processor, storage, and memory. Also, you are using Xampp which is loading the services over top of windows services. Speed testing is always best done on a webserver in my opinion. 

Link to comment
Share on other sites

The trade off is smaller file size to be downloaded from the server, as compared with the server responding faster. Depending on what you are after depends on how you use it.

 

Testing on a dev machine is fine, but not speed testing. You cannot accurately judge the loading time of a website running off of a desktop computer. You have things like monitors attached, video cards, and other things that will slow the system down. Plus more than likely you are using a home class processor, storage, and memory. Also, you are using Xampp which is loading the services over top of windows services. Speed testing is always best done on a webserver in my opinion. 

 

>>>The trade off is smaller file size to be downloaded from the server, as compared with the server responding faster. Depending on what you are after depends on how you use it.

Okay, I'm still missing the point, how is a drop in page load speed in any way beneficial? How should I use it?

 

>>>Testing on a dev machine is fine, but not speed testing.

I agree with you there, but in this case I'm not doing a 'speed test'. I'm comparing two different Prestashop configurations on of which is showing major perfomance drops on several completely different environments.

Link to comment
Share on other sites

Ok, take this case point. You are running a fast web server in a country with slow internet connections. The few milliseconds that it takes to minify the html might not be a problem because it reduces the amount of data sent. The reduced data sent makes the page load .75 seconds faster. That would be a case point to use it.

 

On your end, it looks more like you have a slower web server, but faster local internet connections. i would not use it in that case point. 

 

On the same token, if you want the code minified without the processing overhead, you can hand minify it. 

  • Like 1
Link to comment
Share on other sites

  • 2 weeks later...
  • 1 year later...
×
×
  • Create New...