CMS Made Simple Forums

yes, but me and somebody else here like to speek to you, it's better find speeder algorithm to generating menu, against thinking about little bites of memmory ...when theses informations are not right(probably, and not your fault).

about little bites of memmory

Oh no little  -  3 - 4 MB !!!! I think you don't know of what we here are speaking.

using xdebug and kcachegrind I got the following results (see attached picture)
It seems obvious where the bottleneck lies.
(this callgraph is 'sorted' bij time)



[attachment deleted by admin]

That's really interesting petert !

How exactly did you manage to get that graph ? Is there such a tool available for Windows ? My linux box does not have any X manager, so could this work without KDE ?

melix wrote: How exactly did you manage to get that graph ?

'using xdebug and kcachegrind'

melix wrote: Is there such a tool available for Windows ?

http://www.google.nl/search?q=xdebug+kc ... :official

melix wrote: My linux box does not have any X manager, so could this work without KDE ?

Read up on xdebug please.
XDebug is a debugger and profiler, the output of the profiler is fed into a program like kcachegrind. XDebug runs at the server and needs no X at all.

I don't think kcachegrind will work without X, but thanks for your help

melix wrote: I don't think kcachegrind will work without X, but thanks for your help

You are right about that, but where did I say it would?
I thought you had to use windows cr.p for a gui program, that's the google link.

melix wrote: I don't think kcachegrind will work without X, but thanks for your help

Please look at http://sourceforge.net/projects/wincachegrind/

cyberman wrote:

melix wrote: I don't think kcachegrind will work without X, but thanks for your help

Please look at http://sourceforge.net/projects/wincachegrind/

Thanks, seems to be what I was looking for

maxi2 wrote: Here are two of the best profiler classes:

Problem with profiler classes is that you have to change your scripts. Those classes alter the behaviour of your scripts and slow them down and have nasty side-effects.
The XDebug route is transparent and in combination with kcachegrind it will allow you to emulate a 'run' of the program.

It uses the processor emulation of Valgrind to run the executable, and catches all memory accesses for the trace. The user program does not need to be recompiled; it can use shared libraries and plugins, and the profile measuring doesn't influence the trace results. The trace includes the number of instruction/data memory accesses and 1st/2nd level cache misses, and relates it to source lines and functions of the run program.

melix wrote:

cyberman wrote:

melix wrote: I don't think kcachegrind will work without X, but thanks for your help

Please look at http://sourceforge.net/projects/wincachegrind/

Thanks, seems to be what I was looking for

it's mentioned on the very first hit of the google link

Don't be upset : I've been working both very hard and late to hunt those PHP4 vs PHP5 reference management differences, and it's been quite stressing

melix wrote: Don't be upset : I've been working both very hard and late to hunt those PHP4 vs PHP5 reference management differences, and it's been quite stressing

I am not upset, just disapointed
(get some rest, will do you good)

So don't be disapointed : you managed to get me right to xdebug Well, I don't think it's a good idea to have software named xsomething if it doesn't require X, but that's how the X goes

In my daily computer hour in my hollyday i play something with phpedit and his debugger.

The questions - what happens if i load the default page with a fresh installed 0.11

All Smarty plugins are registered (plugins there are never used too)
All Smarty Informations are stored 9 times !!
All informations of the complete tables content and content_props are stored in Object(CmsObject) contentcache
All datas of the modules are stored (plugins there are never used too)
Many datas of the logger are stored , but no user need it.

You loose a lot of memory and time.