Public Forums 71,145 posts

[investx Friend investxJoin date: April 2015Posts: 58Downloads:0Uploads:4 Thanks: 3 Thanked: 1 times in 1 posts]

#153630

Hi everyone,

I have made some speed tests with the new 1.1.4 Version and I receive very strange and different results that I will post below.

I ask everyone who uses this new template to test their sites and post 1 to 3 different test in this threat to compare and find solutions for optimization.
Please indicate where do you host your site and wich server type do you have.

I will start by posting 4 test of my own site.

1. http://www.webpagetest.org/result/100818_35BC/
2. http://www.webpagetest.org/result/100818_35A7/
3. http://www.webpagetest.org/result/100818_35A9/
4. http://www.webpagetest.org/result/100818_35AF/

From my test I noticed 2 issues:

1. It seems that the first 3 requests generated by the template are taking too much time from the Initial Connectin until they generate the template and send the first byte to the browser.

2. The optimized js and css file are requested for the Repeat View too although it should not with the cache option enabled. This make me think that this 2 files aren’t cached.:((:((

Please post your test to confirm or not this issues and find solutions to them.

[Khanh Le Moderator Khanh LeJoin date: November 2013Posts: 1884Downloads:41Uploads:31 Thanks: 44 Thanked: 203 times in 131 posts]

#353113

Before making the test, you should turn the “Development mode” off. You also should make the test at least 2 times, from the second times, the result should be better.

[Khanh Le Moderator Khanh LeJoin date: November 2013Posts: 1884Downloads:41Uploads:31 Thanks: 44 Thanked: 203 times in 131 posts]

#353210

The performance is a big thing we want to improve. We’ve implemented many technique in T3v2 to make it faster. However, performance is not a story of only template.

@investx: can you make a performance test to your site with another templates, then share the result with us.

Thank you,
Khanh Le

[investx Friend investxJoin date: April 2015Posts: 58Downloads:0Uploads:4 Thanks: 3 Thanked: 1 times in 1 posts]

#353281

Hi,

Khanh I have made 2 tests with the JA Purity II and JA T3 Blank Template V 1.1.4.

1. JA Purity II http://www.webpagetest.org/result/100820_3AC3/
2. Framework2 V 1.1.4 http://www.webpagetest.org/result/100820_3AD8/

You must take into consideration that in first test the css and js files aren’t combined and compressed Yet with all that it have a better result.
The second test have a worst result and what is very strange is that the time from the Repeat View is higher then in the First View what is not normal.

Maybe other users can post their results to help us to make more comparisions and take some conclusions.

[investx Friend investxJoin date: April 2015Posts: 58Downloads:0Uploads:4 Thanks: 3 Thanked: 1 times in 1 posts]

#353337

I made a new series of test with different options and I take some new conclusions.

1. Test with no JS and CSS Optimization http://www.webpagetest.org/result/100821_3B86/
2. With Join & Minify http://www.webpagetest.org/result/100821_3B8E/
3. With Join & Minify & Compress http://www.webpagetest.org/result/100821_3B8K/

Now analyze these results

……………………….. Load Time ………. First Byte…………..Start Render

1……First View ……….6.694s………………1.006s ………………….4.950s
…….Repeat View……..1.353s ……………..0.945s……………….1.954s

2….First View ………….. 9.370s…………….5.441s……………. 7.593s
…..Repeat View …………5.630s ……………5.179s …………….5.728s

3…..First View……………10.589s……………5.645s………………8.867s
…….Repeat View……….. 5.621s…………….2.373s…………….. 6.214s

2. As you can see from the above table the Join & Minify option increases the Render Time and not decreases as it should.
The generation of the page increase but the times for the css and js files it seems to be ok http://www.webpagetest.org/result/100821_3B8E/3/details/

3.The Join & Minify & Compress option increases the Render Time even more and beside that in the Repeat View it adds 2 new requests ( that are actually the commpressed js and css files that should not be there)
The generation of the page increase but the times for the css and js files are not ok anymore.
http://www.webpagetest.org/result/100821_3B8K/3/details/ As you can see in this link the Time To First Byte for css and js file increases very much. :((
Also as you can see from this link http://www.webpagetest.org/result/100821_3B8K/1/details/cached/ in the Repeat View the optimized css and js files are requested again although it should not.

Now please see this test result made with the JA T3 V 1.1.1 http://www.webpagetest.org/result/100727_1XFA/
In this version I think the optimization worked fairly well.
A difference that I noticed is that in this version the optimized css and js files are stored in cache/t3 folder and not in the t3-assets folder.
By the way , what is the reason to save this files in the t3-assets folder which is never emptied and not in the cache/t3 folder that are emptied from 15 to 15 min ???

So in the end I think this is the place in wich you must pay more attention, in the code that make the css and js files optimizations and the place where they are stored and the mode in wich they are handled.

Well this is all that I can help you until now and in the hope and trust that you may solve this important issue I wish you success.

[Khanh Le Moderator Khanh LeJoin date: November 2013Posts: 1884Downloads:41Uploads:31 Thanks: 44 Thanked: 203 times in 131 posts]

#353338

Thank you very much investx. Your help is much appreciated. We are trying to find out the solution to improve the performance of T3v2 and hope that the next release of T3v2 will resolve those issues we face at this moment.

[veeco Friend veecoJoin date: February 2009Posts: 94Downloads:0Uploads:1 Thanked: 1 times in 1 posts]

#353671

Hello this is interesting matter since page speed now has become one of google ranking factor… i will have my first experience with this frameworks soon… so i will monitoring this thread and hope some user can help JA team to solve this problem.

@khan lee, why you plan to focus in this issue on V3 (i think V2 is only at early stage), does the issue is too complicated to track and solve ? please share your view ? i think it could be serious issue since its a template framework issue, not a content issue.

[Khanh Le Moderator Khanh LeJoin date: November 2013Posts: 1884Downloads:41Uploads:31 Thanks: 44 Thanked: 203 times in 131 posts]

#353813

<em>@veeco 191605 wrote:</em><blockquote>
@khan lee, why you plan to focus in this issue on V3 (i think V2 is only at early stage)</blockquote>
Sorry, it’s typing mistake :). We have no plan for V3 at this time.

[investx Friend investxJoin date: April 2015Posts: 58Downloads:0Uploads:4 Thanks: 3 Thanked: 1 times in 1 posts]

#354029

First of all I want to congratulate you. You are on the right road.

I made a new series of test with different options and with the new 1.1.5 version and I take some new conclusions.

1. Test with no JS and CSS Optimization http://www.webpagetest.org/result/100829_3RXZ/
2. Test with Join for JS files http://www.webpagetest.org/result/100829_3RY3/
3. With Join & Minify http://www.webpagetest.org/result/100829_3RY8/
4. With Join & Minify & Compress http://www.webpagetest.org/result/100829_3RYK/

Now let’s analyze these results step by step

……………………….. Load Time ………. First Byte…………..Start Render

1……First View ……….6.434s………………1.011s ……………….4.766s
…….Repeat View……..1.365s ……………..0.944s……………….1.910s

2……First View ……….6.211s………………1.229s ……………….4.537s
…….Repeat View……..1.292s ……………..0.876s……………….1.163s

3….First View ………….. 7.352s…………….3.724s……………. 5.856s
…..Repeat View …………2.862s ……………2.380s …………….2.896s

4…..First View……………5.632s……………1.425s… …………..4.077s
…….Repeat View………..1.538s……………1.049s…………….. 1.374s

The results are categorically better than those obtained with the previous version, however there is much room for improvement because the difference between the first (with no optimizations) and the last test (with all the optimizations) is not great at all.

2.As you can see from this link http://www.webpagetest.org/result/100829_3RY3/2/details/ the Join option it seems it do his job. It increase the js file dimension and download time (wich is normaly) but will all that it decreases a little the Start Render time and keep the Time to First Byte small enough for the js file and the template too ( see 1. and 3 file from the above link). So I belive that the Join option it’s OK.

3. The first problem I think is here. It seems that the Minify option increase the Time to First Byte.

4. Also as you can see from this link http://www.webpagetest.org/result/100829_3RYK/3/details/ the Compress option reduce the download time (as it should) but increase the Time to First Byte for the first 3 requests. So the gained speed is very small.

So in the end I think that the code for the Minify and Compress functions must be improved.

1 user says Thank You to investx for this useful post

1. Khanh Le

[Author]

Posts