XFCE memory usage: a simpler test

Alpay Erturkmen kalpaye at ii.metu.edu.tr
Wed May 26 08:17:00 CEST 2004

as already been pointed out, this comparison is a mess. with 1GB of RAM 
Linux is guaranteed that it will be using every bit of it...

but with guys like me, who has a p2 with 128Mb of RAM, you have to 
choose every application you use very carefully.

Personally i even hate to hear about KDE or GNOME running on my machine 
since it takes longer for them to start than Linux to boot.

This is why i prefer xfce over them. it is fast and furious. however i 
do not like xffm in this sense, so i combine rox with xfce. I tried to 
use nautilus but it is slooooowww...

i personally invite you to perform this tests on machines with at most 
128mb of ram :) i would do them myself but i do not have full GNOME or 
KDE installed on my machine due harddisk constraints.

Have a very nice day :)
Alpay K. Erturkmen
Research Assistant
METU Informatics Institute M.Sc 2003
METU Industrial Engineering 2001

Personal web page on http://www.ii.metu.edu.tr/~kalpaye
Journal on http://my.opera.com/kalpaye

Windows is not the answer, it is the question. The answer is "no!".
Running on Fedora Core 1 Linux and XFCE4.

Ylosar Goer wrote:
> purslow at sympatico.ca wrote:
>> thanx to OF BT ET AL for their intelligent comments on the Gentoo thread:
>> i forwarded it for interest & -- as BT concludes -- it is practically 
>> useful.
>> <snip>
> I think that knowing how much RAM an Environment eat at startup may be 
> of interrest if the amount of RAM at your disposal is severely 
> restricted. But startup and initial usage are not everything.
> When you are limited in RAM, i mean really limited like when you take 
> care of how many xterms you already opened (the next one will make the 
> OS swap...), the way your Environment process the display of a simple 
> menu could be critical. Simply consider the display of a menu. If that 
> menu contains icons (i don't even talk about transparency nor droped 
> shadows...), displaying it involves more hardware resources (which then 
> cannot be used by something else at that moment) than if it did not 
> contains icons. Icons have to be read from disk, so they reduce the 
> efficiency of the disk buffers (by discarding more usefull data) and 
> consequently will reduce the global reactivity of your system to your 
> next action (launching an xterm!) and take other resources to be 
> rendered on screen.
> This 'icon' point is only an example but i think it's a good 
> illustration of what can lead an Environment to the 'bloated' category 
> when RAM (or CPU cycles) become precious.
> But, this is obviously not the only factor that make (if it does) an 
> Environment bloated or lightweight. The difference between the two may 
> be in *how* that Environment uses its resources (initialy acquired or 
> not). One Environment that takes 20MB at one moment may easyly be much 
> more efficient than an other one that takes only 10MB of RAM to process 
> the same fonctionalities, depending on how it has been designed _and_ 
> coded. Imagine forgotten loops, debug statements, verbose logs or 
> uncatched exceptions that may slow down every single things you do... 
> who knows.
> So, in my opinion, numbers, and especially those ones are not very 
> usefull to classify Environments, even in the original post perspective.
> Oh, and i'm sure i'm forgotting many things like having a gtk 
> Environment while using qt apps (or the opposite) that may 'bloat' a 
> 'lightweight' Environment.
> -- 
> yP
> _______________________________________________
> Xfce mailing list
> Xfce at xfce.org
> http://lunar-linux.org/mailman/listinfo/xfce
> http://www.xfce.org

More information about the Xfce mailing list