is everything same ?
Server, Theme, add on and others ?
if something different I suggest you list so it can be analyzed

At this stage I have changed the theme slightly (improvements) to increase speed on the new sites. But i can make another exact copy and upgrade it for testing.

who do I raise/list this issue to?

when you upgrade to 5.5.1 you have to update a couple of plugins to get them working with 5.5.1 from 5.4.2.2 so its hard to do an exact like for like test.

Both sites below are exactly the same the only difference is the version.
http://www.551.white-design.com
http://www.5422.white-design.com
In firefox safari and chrome the 511 (new version) loads consistently slower when browsing cash has been emptied.

new 5.5.1www.www.551.white-design.com/projects... Size 6.14 KB Load Time 8.19 seconds Average Speed per KB 1.33 seconds

older 5.4.2.2www.www.5422.white-design.com/projects... Size 6.14 KB Load Time 3.6 seconds Average Speed per KB 0.59 seconds

so im not mad now I'm doing an upgrade on the 5.4.2.2 -> 5.5.1 and then I can give results between the 3 sites.

Look pretty similar to me (404 errors not withstanding)

http://www.webpagetest.org/result/120131_5P_32BGV/...
http://www.webpagetest.org/result/120131_DA_32BGY/...

Ok interesting results
I have a clean 5.4.2.2, 5.5.1 and an upgrade 5.4.22->5.5.1

all three have exactly the same settings and content just different versions.

5.4.2.2 consistently loads at least 1sec faster than when it was upgraded to 5.5.1

5.5.1 is consistently slower than 5.4.2.2 by .3 to 1sec

5.4.22->5.5.1 is consistently slower than a clean install of 5.4.2.2
here are the latest details fromhttp://www.webpagetest.org
..........................................................
5.4.2.2 results
First View
Load Time 3.039s | First Byte 0.631s | Start Render 2.381s
Repeat View
Load Time 2.792s | First Byte 0.539s | Start Render 0.966s
..........................................................
5.5.1 results
First View
Load Time 3.497s | First Byte 0.657s | Start Render 2.448s
Repeat View
Load Time 2.521s | First Byte 0.575s | Start Render 1.143s
..........................................................
5.4.2.2 upgraded to 5.5.1 results
First View
Load Time 4.154s | First Byte 0.946s | Start Render 2.766s
Repeat View
Load Time 3.719s | First Byte 2.362s | Start Render 2.751s
..........................................................

Can anyone comment or help with these results, or am I missing something?

Interesting, but I would like to know more about the test conditions.

How many pages are averaged for each result?
How big are the databases that each version is using?
Is the mySQL configuration identical?
How many old pages/versions?
Is this with or without cache turned on?
How big is the site?
Are the test pages heavy in text, images or javascript?
Is there anything like Google analytics or other third party code involved?

Not much to go on (better to paste the links to webpagetest.org)

But here's what I'm guessing.

First two - not much in it. Within the error margin. Similar time to first bytes and start renders so the difference it probably just one was a little slower to download images on that test run or maybe the slower one has 404 errors (as I found during my test). Run it 10 times and the 1 sec will probably disappear in averaging if it is the former.

Noteable point. Second view is a lot less than the first view for both, so headers are set on the server (for offsite caching) and probably compression too. They have the same or similar server settings.

The third one has about 1/2 second longer time to first byte which is an amalgam of the server and concrete response times. 5.5.1 has been proven to be a couple of hundred ms faster in response so it tends to suggest the server is the main cause for the difference. Second view is about the same as the first so headers aren't set (and probably not compression either). This one has a different server set-up to the other two and is probably not on the same server.

Just guessing!


Later....
Just noticed the time to first byte on the third site "second view" is about 2 seconds more. Scratch the different server. It's your host that is causing the delay. Cannot determine if it is a different setup or not.

thanks for your reply and investigating
here are 3 instances of concrete5

http://www.5422.white-design.com...
http://www.551.white-design.com...
http://www.white-design.org/test/upgrade...

the last is the first site upgraded to 5.5.1 and all sites reside on the same server and all have been freshly installed a couple of plugins added and a small small amount of content added. all other settings are as a default install basic cache no pacha unless forced on pages. the databases are all clean installed for each site 5422 =158kb 551=203kb the 5422 upgraded to 551 is 239kb

each site has only 3 pages and again iv not done anything to one that I haven't to another. i will do a range of 10tests on each url and feedback

Sweet.
Post the link to a "typical" result for each (as I did) and don't forget to choose a sever that is located in your own country.

first off while i run the ten tests x all three sites here is a visual representation attached fromhttp://www.webpagetest.org of the visual loading times of each of the sites

http://www.webpagetest.org/video/compare.php?tests=120131_WY_32C94%...

see attached image and you can clearly see the older version finishes loading the site first the a clean install of the new version followed by the upgraded version.

I promises I'm not mad.

Maybe. But it doesn't tell me jack.
It's not a case of believing. It's finding out what the difference is, and for that I need to know each and every file, when it started and its other times. Post the links to the results. There's far more information in them than can be shown in a pic or a video.

test results for version 5.5.1
http://www.webpagetest.org/result/120131_QR_32CKH/...
avg first page load 8.02s

test results for version 5.4.2.2
http://www.webpagetest.org/result/120131_3W_32CMC/...
avg first load 4.0836s

test results for version 5.4.2.2 -> upgraded to 5.1.1
http://www.webpagetest.org/result/120131_5D_32CMS/...
avg first load 4.5522s

on the 5.5.1 there was a large 29.825s load so if i negate that result and replace it with the new avg 5.5 the new result is 5.5875s

OK. So they are not that dissimilar (4.0-5.5 secs).
The main difference is that both the 5.5.1 sites have 404 errors. Fix those and you will probably see more parity.

Second.
The upgraded site shows a redirect on the first fetch from no trailing slash to a trailing slash. This is probably the reason for the longer time to first byte on the upgraded site in comparison to the others. That will earn you a few hundred ms.

Fix the 404 errors and use the following url (added backslash) to access the upgraded one to eliminate the redirect from the results for comparison purposes.

http://www.white-design.org/test/upgrade/...

I'm seeing the delay in "First Byte". Do you have Full Page Caching enabled?

on
http://testsite.white-design.com/...

My gut definitely points to either the twitter and/or rss feeds at the bottom. Try disabling them and see what speed you get.

thanks that link is an old site i have rebuilt the site from scratch with only 3pages running the old version and the new version.

in this new set of tests im looking at exactly the same content but on different version and why they load significantly different - currently i have found the new concrete5 to load slower than the old and an upgraded version to load even slower.

by my 3tests that are currently running will hopefully shed some light

I suspect that the bulk of the time loading the home page (BTW, some really nice photography!) is due to the Picassa slideshow interface. So, I'm not sure how accurate a comparison this is between the two versions overall.

I'm just confused because I've done a whole bunch of testing on a whole bunch of hosts with full page caching and never had anything over 2 seconds ever. Which leads me to believe that a block involved is not set to cache in it's controller code.

Thanks for your comments - Can someone please explain what levels of CACHE should be turned on for best results.

mkly how do I check if a block is caching or not? and what should i do if its not?

on the page i have tested there are only the default autonav, hrml and content plucks loaded - and yes some java slideshow for the images.

my thoughts are if I load the slideshow at the end of the page at least you have a fast loading page to look at as the slideshow loads.

Also any suggestions of what i should or should not have in my .htaccess would be good to know.

http://www.concrete5.org/community/forums/jobs1/seo-and-speed-enhancements-on-existing-site/#266281

Yes I already have that in my .htaccess i read your posts about that before.

Can you suggest the best cache options - as i get confused about that is best.

Any thoughts on the best cache settings or should i turn it off completely?