call for benchmark – joomlaart templates are to slow

Hello everyone.

doing some server side benchmarking to speed up shops I hat to realize that joomla art templates are fare away from nice load times.

with typical magento templates we can serve 12-17 page requests per second.

joomlaart templates results in only 0.71 page deliveries.

a simple test withe the apache benchmark tool
ab -kc 5 -t 60 http://tube. removed .com/test-product.html

Server Software: Apache/2.2.9
Server Hostname: tube. removed .com
Server Port: 80

Document Path: /test-product.html
Document Length: 23736 bytes

Concurrency Level: 5
Time taken for tests: 63.758 seconds
Complete requests: 45
Failed requests: 0
Write errors: 0
Keep-Alive requests: 0
Total transferred: 1094670 bytes
HTML transferred: 1068120 bytes
Requests per second: 0.71 [#/sec] (mean)
Time per request: 7084.182 [ms] (mean)
Time per request: 1416.836 [ms] (mean, across all concurrent requests)
Transfer rate: 16.77 [Kbytes/sec] receive

Using Zend Server with the included APC Cache on the old test machine…this result is shocking!

I guess JoomlaArt can do better work

17 answers

cgc0202 0.00 $tone August 30, 2009
Public

Hmmm…

Just wondering. Are polls initiated by Members allowed again? I remember during the contentious period of the revisions of JA Teline II, I had a poll in a thread, it was removed. First time that happened. I used to place Polls before and other members did too. So, I thought it was OK.

Since there was no note why the Poll was removed. I placed it back and requested to know who removed it and why.

Hung responded that he removed it. Hung responded that Polls should be initiated by Joomlart Staff and Moderators only. I countered that it has not been the policy before. However, since then I complied with Hung’s directive. After all, it is their site.

I have not been monitoring many template threads though. By accident I saw this thread with a Poll.

Has the policy been changed? Does Joomlart allow members to initiate polls again?

I am in no position to vote in this poll itself because I do not use Magento. I am more interested in the consistency in the implementation of a policy.

Cornelio

#1
Profile photo of Matthew Wilkinson 0.00 $tone August 30, 2009
Public

I expect better quality templates than JA are coming out with for Magento too.

The templates are very good visually, but technically.. they aren’t so hot. very bulky, even for magento.

I have stripped out a number of JA’s addons from my live stores using any JA Magento templates. I find the templates I have modified to be fast loading once they get going, but getting a return request from the server is too slow. It is definitely not my server as it is near empty bar 3 stores, 2 active. I’m not sure why the delay exists with JM templates, as it does not for the free Magento templates.

JA’s support system is so incredibly slow I do not intend to continue with any return business to these guys until they get themselves sorted. Clearly they have too much work to handle, which is disappointing at best..

Come on JA, we all know you are capable of much more, live up to your name

All the best in sorting your systems out.

Matt

#2
wiery 0.00 $tone August 31, 2009
Public

Joomlart templates are definitely bulky… but they do look nice. Hopefully with the one template release per month now it will give them more time to get good designs plus speed up load times.

#3
Profile photo of Kathrin Wolf 0.00 $tone August 31, 2009
Public

unit is in page delivery per second, displaying a sample product page

jm_purity 5.64
jm_larix: 5,68
jm_mesolite: 5,68
jm_sulfur: 0,75
jm_topaz: 0,70
jm_tube: 0,78
jm_lead: 0,77

every benchmark test has be done three times to validate the results and process the APC cache.

#4
nefar 0.00 $tone August 31, 2009
Public

Hmmm…

Just wondering. Are polls initiated by Members allowed again? I remember during the contentious period of the revisions of JA Teline II, I had a poll in a thread, it was removed. First time that happened. I used to place Polls before and other members did too. So, I thought it was OK.

Since there was no note why the Poll was removed. I placed it back and requested to know who removed it and why.

Hung responded that he removed it. Hung responded that Polls should be initiated by Joomlart Staff and Moderators only. I countered that it has not been the policy before. However, since then I complied with Hung’s directive. After all, it is their site.

I have not been monitoring many template threads though. By accident I saw this thread with a Poll.

Has the policy been changed? Does Joomlart allow members to initiate polls again?

I am in no position to vote in this poll itself because I do not use Magento. I am more interested in the consistency in the implementation of a policy.

Cornelio

I started a poll last week regarding peoples experience if they upgraded to VM 1.13. I would assume they turned the function back on at some point.

#5
Profile photo of Arvind Chauhan Staff 30.00 $tone September 2, 2009
Public

would somebody of the joomlart support staff be so kind to post some reply?

Hi zamb,

  1. First of all >> Polls are allowed to be created by Mods and above only. We have been considerate for this Poll of yours, as JA Magento Club is new and Poll guidelines have not be defined for Magento Club. Regardless, the poll topic is valid and speed issues seems to be issue with magento itself, aggravated by use of feature rich templates.
  2. To spill the beans, I am no Magento user and it would be difficult to comment on the speed issues caused by JA theme on Magento Platform.
  3. I would suggest that you run different Magento themes on http://websiteoptimization.com/services/analyze/ and analyze the data and compare with JA theme.
  4. Once you have that in your hand, you would be in a position to point out the files / scripts adding to the extra time taken by Joomlart templates.
  5. May be you can put up a detailed analysis and how and where joomlart can help in speeding up the theme.
  6. However, still I shall forward a special request to the Developers (via Mr. Hung) to have a look at your thread.

Few points to ponder :

  • JA Products are feature rich and fully loaded, its upto the user to customize them to their taste.
  • Its a simple calculation >> More Feature >> more JS & more css files >> more css images >> More loading time, increased HTTP requests >> Load on server. This is the case with any template for any platform.
  • Faster running templates would need to sacrifice the features or desings (decrease of css images) or as a last resort use of css sprites >> which shall decrease the scope for customization on user end.
  • Probable reasons for joining JA Magento can be attributed to the appealing designs or features, these additional things are bound to take up some resources of any server.

We would be grateful if you could elaborate on this and specify specific features or areas of JA template where you think improvements can be considered. Please PM me your site url, so that i can analyze the same and put up some suggestions myself.

With warm regards & Thanks,

Arvind

—————-

@ cgc0202 >> Please do not use the Poll related issue in here or quote me in template forums for the same. For now (till a final decision is announced), polls are only Mods and above, this poll has been kept running after consultation with Mr. Hung. It would be great if you could raise your opinion about the poll issue in JA general forums, you may move the content of your post to it. I would be deleting all offtopic posts from in here (including this comment), so that this thread stays strictly on topic, as it might be basis of some concrete tips or changes in future.

@ nefar >> same as above.

#7
Profile photo of Kathrin Wolf 0.00 $tone September 2, 2009
Public

Arvind, thx for your reply.

But you missed that I’m not talking about load times within the browser.
Apache Benchmark does just query a url to get the html code back from th server:
So no css js files or any other gets loaded..the web server has just to process the php code and deliver it to the benchmark client.

As you see in my second post:
on the test server, apache could only deliver 0.7 magento product sites per second. ZERO point SEVEN.. other templates results to 5 sites/second and above.

I’m talking about the really worse php code within JM Templates that sucks up cpu resources on the webserver side.

..not the extension code, not the load times on the server I’m talking just about the php code within your templates.

Other templates like the one from template monster or silver template provide clean code.
So it is not an magento issue..is a code design issue within your templates.

We do benchmark testing for a lot of clients and I have never seen that a template had such negative impact on the webserver…since I made the apache benchmark test by myself with the above mentioned templates.

Anyway..I do not criticize your design or your support -- just looking forward that the speed issue gets solved.

you will find more information about the benchmark suite on http://httpd.apache.org/docs/2.0/programs/ab.html

I hope I could make myself clear.

#8
Profile photo of Arvind Chauhan Staff 30.00 $tone September 2, 2009
Public

Dear zamb,

Thanks a lot for correcting me. Yes, its quite clear that i misunderstood the point raised by you. I am much more of Joomla Guy. I do assure you some comments coming in from Developers through Mr. Hung at an earliest.

I am going through the Apache Benchmark link given by you.

Arvind

#9
Profile photo of Mv Tools 0.00 $tone September 20, 2009
Public

Do not use in any case the quickstart.
Why I do not know.
But Magento install by means of downloader and afterwards manual install the template, ensures 50% speed profits.
I have now tested this at all templates of JA.

#10
nefar 0.00 $tone October 1, 2009
Public

I am just curious if you have been able to do any tests comparing JA templates to other companies Magento templates? It would be interesting to see if there is a marked difference in speed/loading times.

#12
Profile photo of tomc 0.00 $tone October 1, 2009
Public

I am just curious if you have been able to do any tests comparing JA templates to other companies Magento templates? It would be interesting to see if there is a marked difference in speed/loading times.

VERY good point and suggestion !!!

#13
Profile photo of Matthew Wilkinson 0.00 $tone October 3, 2009
Public

http://demo.hellothemes.com/hellomini/

Browser load speeds..
Homepage -- 4.07 seconds.
(this is slower than it actually would be as they are querying an image that does not exist)

Category page 3.66 seconds (this speed is more like it!)

Product page -- 4.17 seconds

Not major improvement on JA’s, if much at all. will see how fast i can get this template running on my own server.

#14
nefar 0.00 $tone October 3, 2009
Public

I do like the feature set Magento, the more professional look, and that they actually update the system on a regular basis. However, I keep finding horror stories of slow load times… Especially on servers that are not specifically optimized for Magento.

#15
Profile photo of Kathrin Wolf 0.00 $tone October 5, 2009
Public

I am just curious if you have been able to do any tests comparing JA templates to other companies Magento templates? It would be interesting to see if there is a marked difference in speed/loading times.

we have done client side load time testing too, but I think we should give joomlaart some time to fix there issue before we post the results -> since 1-2seconds server time processing is a huge part of the client load time

-> but as a clue..with some zend modifications most magento templates have load times under 2 seconds.

@mattw just a browser test on a demo site is not a real benchmark, no cache, no server side preloaded cache and so on…this number is not usefull.

@netar "…Especially on servers that are not specifically optimized for Magento…."
Variens intenion is not a small light cart like os:c ..so it will never be really fast on a shared host.
If you are looking for a fast cart, give opencart a try.

#16
nefar 0.00 $tone October 6, 2009
Public

@netar "…Especially on servers that are not specifically optimized for Magento…."
Variens intenion is not a small light cart like os:c ..so it will never be really fast on a shared host.
If you are looking for a fast cart, give opencart a try.

Well if a cart actually requires a dedicated server In order to function I have my doubts if it’s programmed with server resources in mind. Which tells me about the quality of the programming, you get lazy on taking resource usage into consideration it’ will come to bite you in the arse down the road.

#17

This question is now closed

Written By

Comments