Power & Memory Usage Of GNOME, KDE, LXDE & Xfce

[V!NCENT]

That's some unthoughtful advice. Some people have better things to spend their money on.

Most people, like me, are seeing a computer as a little bit more of a valuable nessecaty then others I guess. More than a car that is. If you' re doing it right you can upgrade your entire PC for no less than 300 dollars. If you can use that for 6 years than that's pretty much of a good investment. Unless you're talking food and a roof above your head I'd say there aren't a lot of things more important than a computer...

Which is a horrible price hike when ~2 years ago it cost 20$.

Realy? DDR3?

[curaga]

All I know is that one slot of 2GB of RAM costs 47 bucks:

Which is a horrible price hike when ~2 years ago it cost 20$.

[yotambien]

And text editing long lines in a Qt based editor (whether the lines are *dynamically* wrapped or not, but with syntax highlighting on) is unbearable (a bug in QTextEngine). Actually the same is true of gedit but not other GTK editors.

This is horrible. And it's been like this since I first tried the KDE4 versions of Kate and Kile, way, way before I actually installed the whole thing. Do you have a link for the actual bug?

I really like the latest KDE but seeing results like this and being primarily a laptop user, one sometimes gets the urge to switch back to gnome.

I observed that CPU spikes while doing what one would consider to be stupidly simple tasks are way less common now than a couple of releases ago (I don't want to imagine what KDE4 <= 4.2 was like, and I thank those poor bas...testers who got most of the bugs out of the way). My only complain about KDE4 now is the startup time really.

[molecule-eye]

Is consuming the same as utilising? Sometimes not.

All I know is that one slot of 2GB of RAM costs 47 bucks:

http://www.newegg.com/Product/Product.aspx?Item=N82E16820148221

And I know that KDE is much faster in running multiple apps at the same time than Gnome.

Computers these days can handle anything if ran at full speed, utilising an entire core (no timeslicing). Thus todays law is no Longer Moore's; it is load.

Memory footprint and speed (Enlightenment is set to run at 30fps, not 1000fps) is totaly out of the question today. Get over it and replace your 10yo PC with a 500 dollar one from Dell and never again worry about anything for the next 6 years.

That's some unthoughtful advice. Some people have better things to spend their money on.

[V!NCENT]

Is consuming the same as utilising? Sometimes not.

All I know is that one slot of 2GB of RAM costs 47 bucks:

http://www.newegg.com/Product/Product.aspx?Item=N82E16820148221

And I know that KDE is much faster in running multiple apps at the same time than Gnome.

Computers these days can handle anything if ran at full speed, utilising an entire core (no timeslicing). Thus todays law is no Longer Moore's; it is load.

Memory footprint and speed (Enlightenment is set to run at 30fps, not 1000fps) is totaly out of the question today. Get over it and replace your 10yo PC with a 500 dollar one from Dell and never again worry about anything for the next 6 years.

[molecule-eye]

Battery usage

I don't care much about the memory usage of KDE, what concerns me is the battery usage. I've noticed (monitoring cpu load when doing various tasks) that KDE seems to be, in general, more of a cpu pig than gnome, even with compositing disabled. E.g. hovering over the Task Manager widget in KDE (also Smooth Tasks and the like) is quite cpu intensive, whereas hovering over Window List in a gnome panel is far less intensive. I find that odd.

And text editing long lines in a Qt based editor (whether the lines are *dynamically* wrapped or not, but with syntax highlighting on) is unbearable (a bug in QTextEngine). Actually the same is true of gedit but not other GTK editors.

I really like the latest KDE but seeing results like this and being primarily a laptop user, one sometimes gets the urge to switch back to gnome.

[yotambien]

I think the problems with the numbers provided by Phoronix have been exposed several pages (and many days) ago. And I think it's also clear that KDE4 consumes more memory than the other DEs. I don't want to discuss that again. However, it is quite funny that you bring the updates you did today as some sort of argument against a benchmark performed two weeks ago.

[Shining Arcanine]

It seems today's updates have further reduced the memory footprint of my system to 139MB. My system is still running KDE 4.4.1, so the only change has been in background libraries. The ability for me to get this sort of memory footprint calls the numbers Phoronix provides into question. There is no way that the desktop environments, even the lightest weight one, LXDE, are using the majority of the RAM being reported in use.

[Shining Arcanine]

First of: ppl. with arch/gentoo asking for vanilla tests please do it yourself and post here the results you get. I think this approach is better and doing so you will contribute more knowledge and help ppl. be more informed about this issue.

Second, I think that (at least for me) the results are as expected.

LXDE is the lightest DE because it offers the least amount of gfx candy and least amount of functionality. But guess what, for some ppl. this is exactly what they are looking for.

Then we have Gnome, which is between LXDE and KDE. It's a full DE, plenty of apps., some gfx candy (if you want more you get it with compiz) and middle resource usage. For some ppl. this middle ground is the sweet spot.

Then you get KDE: the full DE platform. You get lots of gfx candy, lots of apps/services/addons/widgets. Ppl. that like customization will feel @ home here.

So they all fill a "niche" and there is place for all of them.

Here is a test. My laptop running Gentoo Linux uses 148MB of RAM in KDE 4.4.1 with only konsole open (and all of the various stuff that is set to auto-start running in the background, including Wicd), which is far below than any of the figures that Phoronix presented, even for LXDE.

All of the stuff Ubuntu runs in the background is skewing the results. The only reason KDE does so poorly in memory footprint on Ubuntu is probably more of an indication that someone did not properly replace Gnome with KDE (perhaps the Kubuntu people) prior Phoronix's tests than it is of any deficiency in KDE.

[mugginz]

I might add that even though XP took about a week and a half to boot, unlike Ubuntu it was far from as responsive and settled for quite a while after the start menu was available for use.