Performance issues - Benchmark theme

Interesting. I can't re-produce those results. On my Dev Environment I -as you said- could barely see a difference in rendering times. Could you please give me a list of your enabled modules? Also, do you re-build your theme registry on every page call or was that turned off? Omega might be slower when rebuilding the theme registry over and over again because it has more to cache, but after caching there should not be any difference at all.

uh... that was a quarter of a second. No Really.

Hmm, I still can't reproduce this... Not even with the mentioned setup AND rebuilding the theme registry on page load. Let's keep this issue active and see if others encounter the same problem.

mkay...

Still stuck in London after Irene pushed our return flights out until late Wed night to get back to Boston, but trying to get back on top of some work items today.

So I may have used some "misleading" marketing text when last updating the project page regarding performance.

However:

When comparing the performance of themes, it is simply rather impossible to do. No two themes are alike. You can compare the performance of a couple of views that were created to extract the same data using different methods. But no two themes provide the exact same functionality.

Omega provides, hands down more features than ANY base theme out there between the grid setup and customization, the preprocessing features, and the advanced setup for responsive layouts. There is NO way to compare Omega to any other theme until another theme provides the exact same functionality as Omega.

With ALL the features included in Omega, I would consider running 200ms slower than Zen for example a huge accomplishment. Same goes for comparing it something out of core like Bartik or Garland. Those themes do pretty much nothing but maybe a couple of theme function overrides, and some CSS loading.

I consider the current performance level of Omega to be a huge leap forward in providing a wealth of functionality with only a tiny bit of performance hit.

I will, when I get back to the states, or if I run out of work playing catch up over the next few days still in London, take the time to update the entirety of the project page with a lot of the missing information on functionality currently included in Omega, and revise my "performance" statement to be more accurate.

I would have noticed such a performance hit during the development process. There is no way that the current code produces a 200ms performance penalty. If it does then there is a bug somewhere which oddly only occurs on your test environment. As I said: We need to wait for others to report similiar issues in order to narrow it down to something specific. I have about the same testing environment setup as you do, Jeff... When switching themes there is not even the tiniest slow-down, not even compared to Bartik or the like. Let's keep an eye on this anyways and look out for similiar issues in our queue. There MIGHT be a bug.

As an outsider who admires and uses themes from both of you, I would hope you would trade access to each other's test sites and see what's going on.

I don't see any aggressiveness, only one sarcastic remark about 250ms not being a lot (I think it is) and unbelieve amongst the dev's who show they take it seriously and will try to reproduce and find the cause of your findings.

@dhalbert I cannot believe there would be any competitiveness playing any role here (call me naive).

My previous comment implying 250ms was acceptable stems from my belief that Omega's use of CSS Media Queries, as part of a total responsive web design solution, would be a bit slower to render than themes that did not use this architecture and that is okay by me. It is my understanding that CSS media queries are parsed and interpreted before the browser starts to lay out the page. That takes time and is probably evaluated somewhat continuously. It is not that 250ms is not a long time to some, it is that is acceptable to me for using such a great theme.

If CSS Media Queries are not the culprit to the 250ms difference, then I am VERY, VERY disappointed in the Omega development team! Shame, shame on you. (Now that's sarcasm.)

;-D

Perhaps test with and without media queries.

To an outsider, you look a lot like a man with an agenda here Jeff. Why don't we wait and see if anyone else is experiencing the same issue...

Jeff, I didn't want to sound disrespectful in any way. Sorry if I did. I only wanted to say that I have never encountered any performance penalty like that. I am really interested in seeing where this goes and maybe have other ppl report the same problem. I REALLY cannot reproduce this, so we need more ppl with the same problems in order to narrow it down to something specific in the theme. Again, sorry if I sounded disrespectful or aggresive in any way. I greatly appreciate the work that you are doing (and have done) for Drupal and all the projects that you work on. Jake once said that my tone might sometimes sound a little bit disrespectful to someone who is a nativer speaker (as I am not). Please consider that!

Thanks Jeff. So you are reporting about the overall rendering time (not only server-side loading but also client-side rendering of CSS), is that right? In that case we talked at cross purposes since I was only referring to the server-side performance. Interesting point, I read a great article about CSS performance and how overrides can introduce a performance penalty for the browser a few weeks ago. Omega makes heavy use of such overrides and therefore the problem might even lie there then. My PC is extremely fast and that might be a reason why I cannot reproduce this. I am going to dig deeper (using my notebook) in a few days when I am back from my holidays.

@sammyd56: Please read the Drupal Code of Conduct, your reply to Jeff was very disrespectful and I'm sorry you've gone a full year thinking it was acceptable. Jeff is trying to help the Omega team by mentioning he saw performance problems - rather than attacking him maybe you could instead help to identify the problems and then help to solve them?

Old one, but performance is still important. Bumping this to 7.x-4.x. Perhaps we should have a docker image that we could use to compare core, Omega, Zen and other on. Working on D8 and Twig some nice tools was build to ease the profiling bit - https://www.drupal.org/contributor-tasks/profiling