Why did initial load time improve?

General questions pertaining to how certain issues can be resolved. This forum is closed with effect from February 2019. Please post future requests on https://github.com/sayontan/suffusion.
Forum rules
This forum is closed with effect from February 2019. Please post future requests on https://github.com/sayontan/suffusion.

1. No offensive language and no mocking
2. Please do a thorough search before you post something. Trust us, there is a high probability that the question you are asking has been asked previously.
3. No soliciting. You cannot post here soliciting bids for people offer you quotes, or even offer money to people for some work. You will be moderated if you do so. If you are looking for help, please post your request on http://jobs.wordpress.net or http://codepoet.com
4. Please be reasonable. You are getting software and support. For free. Complicated requests from a general purpose theme are not welcome and some volunteers might lose patience with you.
5. Please do your due diligence. If you posted a query and we answered with a link, take the trouble to go through the link contents.
6. Please post with complete information. Requests for help MUST be accompanied with your URL, particularly if you are asking something like "Why am I seeing a blank space?"
Locked
digitalham
Posts: 6
Joined: 18 Mar 2015, 02:30

Why did initial load time improve?

Post by digitalham » 14 Mar 2016, 03:47

Sorry this is a bit long. The actual crux of the matter is simple but the background info is necessary.

I have several WP sites on a the same server. I keep track of performance and problems by running a monitoring routine on a raspberry pi. It just uses CURL to do initial loads of the entry page on each site and records the results.

The site in question was converted from a hand coded html/php/mysql site to WP. Most WP post entries were generated by a script and just pull in php to generate the content with pretty much the same php as the original site. There are over 3000 pages and I put the poor load times compared to my other WP sites down to that but the only reported issues on performance and large page counts on google are for the back end so I looked at plugins. It's a campsite directory and owners can log in to update their own listing. When I converted to WP I used a plugin to allow those logins. I thought maybe it was that plugin so I've split the site into two - one for display and one for maintenance. Dropping the login plugin didn't improve things and all the other plugins are common to all sites.

Today I thought I'd compare Suffusion options between the sites using a file compare of the exports and there was only cosmetic differences. I thought I'd spotted a couple of items that were maybe not optimal and updated skinning/header only to find it was correct and had to change it back. I also changed Backend/Custom Includes to enable default feed as I have a plugin that removes feeds links and thought not having one may cause a problem. I then found that it was disabled in all the other sites with no problem. I have left it enabled but am pretty much certain from looking at the load times now that it makes no difference.

So other than altering a couple of minor things and changing one back again I've done nothing.

The very strange thing is that all the response times since doing that have been around half of the fastest time before. Any ideas? Transients?

Some time back I posted here about issues with a different WP site which I blamed on menus despite finding no other issues with them as it was the one thing different about that site. I went over to a hand coded menu and performance improved. I'm beginning to wonder if that was more down to re-saving some of the options pages but can'y for the life of me see how.

Here is the table of the 30 fastest load times. I was making the changes around 9:10

2016-03-14 10:45:06 Datasize: 5919 Load time: 0.652
2016-03-14 10:10:04 Datasize: 5919 Load time: 0.667
2016-03-14 10:40:05 Datasize: 5919 Load time: 0.693
2016-03-14 10:15:05 Datasize: 5918 Load time: 0.696
2016-03-14 09:05:04 Datasize: 5919 Load time: 0.698
2016-03-14 10:05:05 Datasize: 5919 Load time: 0.712
2016-03-14 08:55:04 Datasize: 5919 Load time: 0.717
2016-03-14 09:45:05 Datasize: 5919 Load time: 0.733
2016-03-14 09:25:04 Datasize: 5919 Load time: 0.735
2016-03-14 10:20:05 Datasize: 5919 Load time: 0.741
2016-03-14 09:30:06 Datasize: 5919 Load time: 0.750
2016-03-14 10:30:04 Datasize: 5919 Load time: 0.750
2016-03-14 09:35:04 Datasize: 5919 Load time: 0.753
2016-03-14 09:55:04 Datasize: 5919 Load time: 0.771
2016-03-14 09:10:07 Datasize: 5919 Load time: 0.790
2016-03-14 09:40:04 Datasize: 5919 Load time: 0.866
2016-03-14 09:15:09 Datasize: 5919 Load time: 0.935
2016-03-14 09:20:05 Datasize: 5919 Load time: 1.036
2016-03-14 09:50:04 Datasize: 5919 Load time: 1.185
2016-03-13 02:15:04 Datasize: 5912 Load time: 1.205
2016-03-12 01:55:04 Datasize: 5912 Load time: 1.211
2016-03-13 04:45:05 Datasize: 5912 Load time: 1.211
2016-03-12 11:05:05 Datasize: 5912 Load time: 1.218
2016-03-12 23:05:04 Datasize: 5912 Load time: 1.218
2016-03-12 02:40:04 Datasize: 5912 Load time: 1.221
2016-03-13 17:25:04 Datasize: 5912 Load time: 1.224
2016-03-14 01:15:05 Datasize: 5912 Load time: 1.226
2016-03-12 18:55:04 Datasize: 5912 Load time: 1.227
2016-03-12 13:10:03 Datasize: 5912 Load time: 1.229
2016-03-12 08:45:04 Datasize: 5912 Load time: 1.230

I will obviously keep checking to make sure that this improvement does continue. Unfortunately I have no idea what the extra 7 bytes are in the pages since the change.
===update===
The 7 bytes is a red herring. The page is dynamic and was 5919 before I played about with it today.

Locked