well he must have a slowish server tho..

the average execution time on my site is around 0.17

the average time for my forum ranges from 0.045 - 0.145 (IPBM 1.4d)

I told myself I wouldn't get into this but here I go again...

As stated they did a comparitive test under equal conditions. I won't nit pick the server choices and I stopped reading at the statement basic install: 1 news listing. Who would use any CMS with one item?

Perhaps the stress test is indicitive of what a real site with more content would perform like but maybe not. Scalability and sustained load under real world conditions often returns much different results. Some can handle the conditions better than others and could be different than the base test. I don't know how well XOOPS would be but it works for me. The Nukes and others are also fine in their world too I'm sure.

Here is my analogy: My car is pretty basic and there is a much more expensive and powerful version of the same car available. In a similar one on one test the other car will prove much more powerful than mine. But with my careful set-up (very limited budget) and the type of tight rough courses I run on I frequently beat that other car. Real world conditions and a good driver can give unexpected results.


The 0-60 times are only one (outdated) part of the story.

I found XOOPS to be a bit of a dog of a system to start off with too. Then I realised the wonders of the caching system, inparticular the block caching system. If you run the MySQL debug option it'll tell you how many queries are being run. Get this down using the caching and/or by removing non essential content and XOOPS will perform like it's got a dirty big supercharger bolted to it . On my reasonably modest 1.2Ghz Athlon server it takes around 1.5~2 seconds to process requests if you aren't hitting cache, and about .9~1.2 secinds if you do. And that is with the server running a lot of other very resource hungary services such as Exchange, Active directory and several SNMP stats gethering programs.

If you know what your doing you can speed up any cms even by 50%. I've managed to speed up my md-pro from 2.3s to about 1.2 by eliminating a lot of not used code. But it is impossible to have a cms with lot of features which would be super fast.

Quote:
Right - as stated, more and more features and the OO design does take a little beating on performance.

Smarty may give a little overhead - at least the first time the page is accessed due to the compilation of the template - but the benefits this template system gives are worth it, I think. The caching is good and can help a lot.

However, there are also quite some modules out there, which are rather inefficient in their SQL queries (not pinning particular examples here, just stating it) and I have optimised my earlier versions of my modules a LOT by making the calls more efficient.
E.g. the first Server Booking module had a query for every day of the month - i.e. 28-31 calls - just to get the bookings. I have now reduced that to 1.

So lots of places to begin - and I don't know where in the core is the best place to start optimising, but do keep your eyes open for any inconveniences.

In conclusion, I am of course disappointed in seeing XOOPS so low on the list, but that's not the fault of the tester - and it *may* look different when scaled up in size, but we will have to accept that +features = -performance.