KDE's Leaner Experience On openSUSE Tumbleweed vs. Ubuntu 17.04

[starshipeleven]

But really, this is all *i*k measurement. You will really care about a hundred megabytes here or there if you are stuck on a machine with 2GB. With 4GB and beyond nobody really cares, except for a few freaks.

Meh, if you're stuck on a 2GB machine you've got bigger issues, like that the processor is very likely a turd that can't run smoothly a web browser.

[Shiba]

Nope. That is just Akonadi removed. But I hear that XFCE still has more functionality and customization than a "full GNOME environment" while consuming half the RAM....

If a minimal KDE is your idea of a "full desktop environment", I have no doubt you really believe Xfce is as well. Well, next time you need a PIM or a metadata indexed filesystem on your reduced environment, you could always... dunno, drag your panel around the screen... maybe it will serve some purpose.

Tumbleweed's KDE is not stripped down, it's Ubuntu's KDE that is configured like total crap.

What's missing then? The article says

The KDE user-space components by default on Tumbleweed were far shorter than Ubuntu.

which implies a lot of things are missing, but the forum thread is 13 pages...

[useless]

which implies a lot of things are missing, but the forum thread is 13 pages...

That's my question too. He should really elaborate it more. KDE PIM suite (Kontact) and its backend are not enabled by default in TW (if I recall correctly), don't know about Ubuntu. On the other hand, in that long thread I've posted the memory usage of Ubuntu with the Plasma Desktop session: 420 MiB according to ksysguard. Strange...

EDIT: Now that I think, there should really be some way to ignore users in these forums. Trolls are the reason for 13-ish pages threads, sometimes.

EDIT2: Just tested: In TW, KDE PIM takes plus ~ 260 MiB with no IMAP account, no calendar resources and no addressbook. There, you get ~ 700 MiB of memory usage, far from the GiB seen in Ubuntu. A KDE dev posted in that linked thread, stating that "removing" (apt-get remove) other DEs doesn't always remove their dependencies, their services, etc.

[juno]

The original article should be updated. It starts with a wrong headline and goes on like this.

[starshipeleven]

What's missing then? The article says

which implies a lot of things are missing, but the forum thread is 13 pages...

I think it's a typo (user space components are "far shorter" than Ubuntu? That's not proper english, seems like 2 sentences smashed into one).

I have Tumbleweed running on my laptop and it's not "slim" by any means, all stuff is there, wobbly windows and everything.

Can confirm that Kontact is not started by default (that's around 200 MB of RAM but takes mere seconds to start up on mechanical hard drives).
Other stuff like the (stupid) wallet system is on by default, as far as I can see that's a normal KDE installation.

Besides I don't see why a mail application should be on autostart by default, so I don't see the lack of Kontact startup as a "slimming down".
Also because most sane people use Thunderbird, not Kontact.

which implies a lot of things are missing, but the forum thread is 13 pages...

That thread is mostly me and other people flaming this forum's GNOME fanboy, I'll provide a best-effort tl;dr:

-People pointed out that if you install 2 DEs on the same system you'll end up with services from both running, and this is probably one of the reasons the first test is total crap, where the poor Budgie took ages to start up (this is valid in any distro, not just Ubuntu)

-there were a couple posts that linked to development blogs of KDE where the devs talked about how distros (Debian/Ubuntu mostly) routinely fuck up packaging of KDE by bundling useless dependencies or not bundling needed ones, or by not following upstream advice/memos causing preventable breakage, while praising OpenSUSE for being most responsive and providing best KDE support. There was also an interesting blog post about the bad side of the modern distros with maintainers and package management systems and whatever (well-known stuff for most linux users, but well-written nontheless).

-there was a bunch of people posting their experience and their feedback was that that KDE was using 400-600 MB on their systems and startup times were less than 10 seconds on a SSD.

-useless user posted screenshots of a Ubuntu VM with KDE and again it was using 450 MB of RAM or something like that.

[starshipeleven]

The original article should be updated. It starts with a wrong headline and goes on like this.

Clickbait at its finest. Attract clicks with that, and then attract them again with this one where he "rectifies" his "mistake".

Tinfoil, I need more of it.

[Azrael5]

Michael test Neon kde so to verify some kind of improvements. thanks

[useless]

-useless user posted screenshots of a Ubuntu VM with KDE and again it was using 450 MB of RAM or something like that.

I've set another VM with Ubuntu in it. Installed plasma-desktop, gnome-shell and ubuntu-gnome-desktop, just... because. Didn't remove any DE, always rebooted before switching session. Here is what I found:

KDE Plasma uses ~ 480 MiB
GNOME desktop uses ~ 700 MiB.

In Plasma session there are a couple of GNOME related process (Tracker, GOA, gnome keyring, etc), still, never over ~ 500 MiB of RAM.
In GNOME session, only kwallet seems to be running (6 MiB), still, never over 800 MiB.

So, again, what's up with that PTS result?

[smartalgorithm]

as a opensuse tumbleweed + kde plasma user i'm signing under every single test result here haha thanks for these tests.

[leipero]

Michael, please use "free -m" next time, I mean, you can even automatize the process with it, so it can output in "x time" and you can log it automatically (I did not look how you measured) and calculate avg automatically, plus it will give "more complete" picture on free memory, buffered etc.