MODx / Etomite speed testing

Howdy y’all-

A while ago I did a speed performance test between Etomite and MODx. I thought I’d share the results. I think most of the powers that be at both Eto and MODx have seen it but perhaps the general usership might also be interested:

Jared’s Etomite vs. MODx Shootout

It should be noted that many improvements to BOTH systems have been made recently. I hope to update the shootout soon-ish with new installations for more up to the minute comparisons. Prolly when Etomite 0.6.1 final and MODx 0.9.1 is released.

Enjoy.

Awesome, it will be great if there is another CMS being tested and compares . You can actually make a performance test website for CMS.

Good job!

Hah, as the ultimate speed freak, I of course can’t resist to comment on this one ;-)

And so, here are a few comments:

Looking at the php processing time etomite/modx shows you isn’t completely fair. There is more involved in displaying a page. I would rather go for apache bench or similar, and see what the _real_ pagerequests/second is, rather then looking at those values.

3 runs per page is statistically not a good measurement. 20/30 measures is allready much better. Count in that php accelerators cache the result, I used 100 requests for stressing a page, which gave reasonable consistent results.

Modx install doesn’t use caching by default IIRC. Did you check that the generated pages in both modx and etomite had:
- the exact same template
- with the exact same snippets/chuncks/etc to parse
- and finally, used cached snippet output or not? ([[Menu]] vs [!Menu!])

I profiled modx a while ago, and it really took me a long time to understand the results, or better, to correctly interpret them.

Greetz,

Remon

P.S.
Just to show you how hard it is to test something, this is the result of the cached home page of the default modx install on my amdk6 400 server:
MySQL: 0.0000 s, 0 request(s), PHP: 0.0149 s, total: 0.0149 s,

FYI, there are many parsing improvements in the 0.9.1 core which should speed up things (I think.\...)

The shoot-out was about TP3.2 (as Jared mentioned), we left 3.2 far behind.

TP3.2 -> TP3.3 -> x9 (0.9.0) -> 0.9.1 (very soon)

Just to clarify a bit

regards,

zi

Yep- I used whatever was available for download (by the general public) on the respective sites. I used what I had and did what I could in reasonable time.

I’d love to do another with newer versions of both, but it looks like Remon might be better suited to the job! Take it R.S.!!

Haha, sure I’ll do some stress testing, but right now I don’t have apache installed due a new install. Hopefully I’ll get it solved next week.

But if Eto didn’t change the parser significantly, then there shouldn’t be much difference in parsing speed, so I was a bit surprised that your measurements showed modx performing orders worse then eto.
Also, I didn’t get such bad results on my PIII box either...

Btw, your site with the test results are really nice, clear to read etc.

Have a nice day,

Remon