ozneilau wrote: glynster wrote:
These results might not be entirely scientific as some of the demonstration websites are in frames or similar so that other templates can be selected other than the one currently being displayed. Some demonstration websites also have more content than others. Hopefully the test is still sufficient for a ball park comparison.
W3C valid (or close to it) templates loading in under three seconds:
YouTheme Warp http://www.yootheme.com/themes/warp-framework
was just over 3s
See the attachment for the full details.
Edit: Motif http://manoscrafted.com
loaded in 1.9s
and might also be worth a look.
With all due respect your test isn't a test of template performance at all. Your test is a test of server response, performance and site content. Template performance has very little, and quite indistinguishable in this instance, effect on these results.
The amount of content on the demo site and the location and performance of the server are what's showing in your results.
In the case of content, for example, Construct's demo is loading no images whereas Joomlafuture loads 1 huge banner image and 6 smaller ones. The six small images are pulled from Gravatar (which means page rendering is blocked while they download). These images aren't a part of the template, they're demo content.
In respect of server performance location has a huge impact. GMetrix's free service doesn't allow you to specify a test server location (you can do that with pingdom tools though, a similar service which I think is better than GMetrix). The GMetrix test tool you linked to is in Vancouver. Yootheme's demo server is in Germany and mine is in London. Your tests are showing trans-continental latency more than anything else I think.
There's only one way to accurately test template performance, install all of them on a single server ensure they are all showing the same demo content (or preferably none at all) and test them all from a single known location, either local or from a specific node as you can with Pingdom tools. I'm confident that, tested correctly, you will see a vastly different set of results.
Before I'm accused of bashing your results because my template looks so slow please be aware that I (like the authors of Construct and Motif) have zero commercial interest in doing so as we don't sell templates or extensions.
Personally, this means that a) I haven't put very much effort into tailoring my demo site, (it took me about 3 hours to put together) b) it's on a low-grade development server, and c) I've made zero effort in terms of server set-up (I haven't enabled gzipping or caching) on my server because I don't actually care if you use it or not. I make this template primarily for my own benefit and to act as an example of best practice.
That said, I do think it's important to have the correct facts in place on a thread like this.
Top-tip: marking your posts as "Urgent!", when they clearly aren't, will most likely get you ignored.
Please don't PM me with anything other than *Personal Messages*! If you have a Joomla question, post it in the forum.