KDE 4 Uses 40% Less Memory Than 3 Despite Eye-Candy

An anonymous reader writes "Pro-Linux reports that KDE 4, scheduled to be released in January 2008, consumes almost 40% less memory than KDE 3.5, despite the fact that version 4 of the Free and Open Source desktop system includes a composited window manager and a revamped menu and applet interface. KDE developer Will Stephenson showcased KDE 4's 3D eye-candy on a 256Mb laptop with 1Ghz CPU and run-of-the-mill integrated graphics, pointing out that mini-optimizations haven't even yet been started." Update: 12/14 22:40 GMT by Z : Or, not so much. An anonymous reader writes "The author of the original KDE 3.5 vs KDE 4.0 memory comparison has come out with a more accurate benchmark. In reality, KDE 4.0 uses 110 MB more memory than KDE 3.5.8.

Wow.

[log1385 (1199377)]

Someone call Bill Gates and tell him to read this.

Actually...

[DaedalusHKX (660194)]

And this is coming from a defender of the free market and devout believer in its virtues, but since Microsoft has largely benefited from partnering up with other large manufacturers of hardware and assemblers of said parts into systems to be sold, it would not be that hard to believe that they designed to a certain market level.

I.E.... "here you go gentlemen, the standard system you are able to use is X Ghz, and X Gigabytes of DDR 1600, anything less than that will be obsolete by the first service pack anyways, so get crackin'!!"

Linux people and most of the OSS folks (Unix as well) have been server dedicated systems for a long time, and built on a robust or rather "efficient" (perhaps a better term is "effective"?) platform. As a result, they've been building to extract as many cycles and memory space as possible for use by client applications, not the Host Operating System.

As a result, Microsoft has it in its best interests to PUSH the upgrade cycle. If they can be depended to push the upgrade cycle to keep selling new boxes, the retail computer builders will continue to give Microsoft the plugs and keep shipping their OS as the "default" or "preferred" or "Supported" Operating System for their Big Bad Ass Kicking Rigs (tm).

Re:Actually...

[Kadin2048 (468275)]

I think this is probably true.

As a matter of corporate policy on a high level, Microsoft obviously benefits from and feeds into the upgrade treadmill. I don't think it's hard to believe that there's a quid pro quo with the hardware manufacturers on this; at the very least it's an obvious symbiosis. Microsoft craps out a new OS every few years with vastly increased system requirements (at least in order to run well), and in return the hardware manufacturers continue to bundle Windows. (There's more to the relationship, obviously, such as Microsoft's pricing structure for OEM licenses, but I think the hardware/software upgrade path is a part.)

However, I don't think most of Microsoft's programmers necessarily go into work every day saying to themselves "today, I'm going to build the shittiest, most resource-hogging chunk of code I can, so help me God." I suspect they probably just code for whatever their higher-ups tell them the target platform is going to be. If you're an overworked programmer, and if management makes it obvious that they care more about shoveling in the features than in optimizing code for performance and footprint, you're not going to optimize.

I think that's Windows in a nutshell. Somewhere along the line, some suit decides what the target platform is going to be; at the beginning of the development cycle it's probably pretty top-of-the-line kit. Everything is targeted towards this, and the end result is massive increases in bloat. Optimization is hard and unless you emphasize it and reward it, it's not just going to happen all by itself.

On the OSS side, you see a lot of optimization happen because many developers are working with limited resources and aren't in a position (or have the desire) to go out and buy a faster computer to make some chunk of code run faster. If you write an OSS application that requires your users to go out and buy a new system in order to use it, you've just alienated a lot of potential users -- or, hopefully, created a demand for someone to optimize the code and get it running on existing, slower hardware.

In short, I don't think Windows' footprint and mediocre (or negative) performance gains is due to bad coding as much as it's a direct result of institutional culture. It's a good example of what can happen to any product or project if performance isn't a key consideration, and particularly if it takes a back seat to featuritis.

Well

[markov_chain (202465)]

The laptop was recent, but he limited the memory use and throttled down the CPU to 1GHz. So it still had fancy instructions and a much bigger cache, bus, etc.

Re:Well

[JonLatane (750195)]

1. There's no need for Compiz on KDE4; KWin supports composited window management built-in, and that's what he was using.
2. The computer has Intel integrated graphics, you don't get much lower than that.

Shows what is possible....

[gweihir (88907)]

... with careful work. And a primary focus on excellence, instead of making money. And people that do care about their product.

Re:Shows what is possible....

[CajunArson (465943)]

Uh... if you saw the number of bugs currently open in this "release candidate" (and I use the term loosely) you might be a little more realistic and less idealistic. I use KDE exclusively, but I'm holding off a big permanent jump until this gets A LOT more polish. One problem with OSS is that there's plenty of work that needs to get done that isn't "fun" and people don't like to do the stuff that isn't "fun" for free. Unfortunately, that doesn't mean it's not important.

ressourcenhungriger

[by Anonymous Coward]

"The fact that a new version of an application does not always ressourcenhungriger must prove the KDE project with the next generation of the environment."

I think I just found my new word-of-the-week

Bad measurements

[Percy_Blakeney (542178)]

I'm sorry, but just adding up the memory usage columns from something like 'top' is a horrible way to measure actual memory usage. Why? Well, shared libraries is one big reason. Most of those applications are likely to use a similar set of shared libraries, which the operating system only loads once in memory and then uses for all of the applications. However, things like 'top' include the memory usage of those libraries in every application that uses them. Thus, if 'libkdeprint' is 1 MB and is used by 10 KDE programs, the ACTUAL memory usage of that library would be 1 MB, but top would report 10 MB of memory used (1 MB for each app).

This effect is very noticeable with desktop environments like KDE and GNOME, where there are a ton of programs that all use the same set of shared libraries. If you reduced the size of a few very basic libraries (e.g. 'libkdecore') by a sizable amount, then you could show a fake "huge savings" across the ~30 KDE/GNOME apps that were running.

It isn't that I doubt that KDE 4 uses less memory -- it undoubtedly does -- it's just that using overly simplistic methods to measure the difference in usage is misleading and somewhat pointless.

See a longer discussion of the issue at: http://virtualthreads.blogspot.com/2006/02/understanding-memory-usage-on-linux.html [blogspot.com]

Results are completely false

[by Anonymous Coward]

Has no one pointed out that the numbers are actually completely, utterly wrong? See Lubos and Thiagos (two high-ranking KDE and Qt devs) comments here:

http://www.kdedevelopers.org/node/3138

See the original authors retraction, here:

http://www.jarzebski.pl/read/kde-3-5-vs-4-0-round-two.so

In similar conditions KDE 3 consumed 97 MB on memory, whereas KDE 4 about 170 MB.

So really, it should be "KDE4 uses 75% more memory", which is actually incredibly lame, but doesn't make for as good a title. I'm absolutely amazed that usually cynical slashdot readers have accepted this so uncritically.

Re:New Headline:

[aminorex (141494)]

That would be wrong, though. A basic KDE 3.5 desktop environment uses mid-way between what a Gnome 2.14 and an XFCE 4.2.2 will use. This suggests that a 4.0 desktop may consume less than XFCE does now.

Re:less memory!

[by Anonymous Coward]

The "unused" RAM won't be nice and empty. It'll be used as the system cache to store file data etc. that then can be accessed very quickly. Modern operating systems do not waste RAM by leaving it unused.

4...3....2......1.......

[rustalot42684 (1055008)]

FLAMEWAR!!!
just to speed things up a bit:

• The GNOME devs are interface nazis
• KDE has intolerable configuration menus and is ugly
• XFCE has no functionality
• Other window managers are for freaks and deviants
• Except TWM. Bow before the TWM gods!!!!!!!

CLI FTW!!!

[by Anonymous Coward]

What!? Whatever happened to the "GUIs are for infants and grandmas. if you can't do it on the command line you shouldn't be allowed to use a computer in the first place" flame?

It's a sad day in Linuxland. What became of the holier than thou, I program in assembly, certifiable *nix prick?

Oh, and don't forget, "Desktop environment x is so bloated."

bash? pffft...

[Junta (36770)]

WAY too much bloat for features most never use. Real men use dash (if you *must* have a program that's a shell and only a shell) or if you don't mind something a bit more versatile to save disk space at potentially the risk of slightly higher memory consumption when all you have is a shell, you use a symlink to busybox for your shell. But not with that glibc cruft mind you, uClibc is the only path to efficiency.

Also, you don't use init, you have the kernel run the aforementioned shell directly instead. Who needs all the cruft of startup services and a well set up tty, after all.

What Balderdash!

[spun (1352)]

You young whippersnappers and your fancy shell this and tty that. Real men feed their programs into a time share systems as big as a barn using punch cards, you young hooligan! Why, when I was a lad, all we had were toggles and lights, and we were grateful! Now get off my lawn before I shake my cane at you a second time!

Re:bash? pffft...

[timeOday (582209)]

Point taken, but for KDE to expand on the old functionality while reducing memory footprint by 40% is not a tradeoff - it's just a flat-out improvement.

Re:To compare with GNOME...

[Elladan (17598)]

Keep in mind that even basic modern graphics wastes more memory than that. That background image you have on your 1600x1200 desktop? 5.4 megs. Need a few composite buffers? 5.4 megs each.

Don't have a background? Just the frame buffer to activate that graphics mode itself is 5.4 megs, regardless of what you put on it.

Just to keep things in perspective here. That Commodore 64 you had ran nicely in 64k of ram, but it also only had 320x200 graphics (160x200 in 4-color mode). :-)

Re:To compare with GNOME...

[IGnatius T Foobar (4328)]

* It is not a project of the self-proclaimed "free" or open-source software movement, and does not suffer from popular fads among it, such as Xft/fontconfig and autoconf.

Ok, maybe Ion can run on smaller hardware, but it isn't exactly a feature worth trumpeting that the fonts are going to look like crap. Xft/fontconfig was a brilliant piece of work that finally put to rest all of the moronic "X11 is obsolete and must be completely replaced" ranting. While the dorks were chanting for X11 to be replaced, the Xft/fontconfig people were fixing the exact problems that were supposedly insurmountable. And they did so in a way that preserves X11's legendary network transparency.

Omit this functionality if you wish but don't advertise it as a "feature."

Re:To compare with GNOME...

[bendodge (998616)]

I'd have to back you up there: when I first installed Ubuntu I went with KDE because it seemed less foreign than GNOME. (And I'm quite happy.)

Re:To compare with GNOME...

[h4rm0ny (722443)]

Again, the same. I have used both DEs a lot and I find KDE more capable technically, but Gnome is easier on the eye. I like the top bar and separate window bar at the bottom. I also like the distinct lack of clutter. However, the new KDE looks better in terms of performance and with the new QT libraries, it's apparently easier to create apps that work on Linux and Windows.

If anyone created a KDE "theme" that made it look and work like Gnome, I'd be extremely happy with it. Probably the KDE camp would find that distasteful though.

Re:Just tried

[value_added (719364)]

Still, covering 1/4 of the screen sure didn't take much memory!

Speaking of wasted space and distractions, and not to be trollish, but I've always wondered why it is that KDE and Gnome insist on using large-to-oversized-to-supersized icons for everything, KDE being notable in that it traditionally distinguishes itself with icons of brighter colors, in wilder designs, and offers greater customisability?

Seems to me that the term eye-candy, while often used in a disparaging fashion, should refer to a certain kewl aesthetic, rather than literal candy of the M&M variety. It's almost the inverse of a Queer Eye for the Straight Guy episode -- instead of getting a great design from three flaming queers, you get a flaming queer design from a bunch of straight guys. Well, maybe not that bad, but still.

I mean, really, do people really need toolbars that takes up a 1/3 of the space of an application window? Is the boredom threshold so low that everything has to be decorated with bright colours, or is it that people find it hard to to hit things with their mouse? Sure, both KDE and Gnome are better than Windows, but by the time you've customised things to be less ... well, goofy, you might as well have installed something like Fluxbox or go back to using nothing but xterms, learning to do without the more subtle but useful effects available or being developed elsewhere.

Re:Please post a new story.

[Pulse_Instance (698417)]

I wouldn't admit to having actually read it 10 times.

Re:misleading article

[SoapDish (971052)]

Well, it works fine on an eeePC 900MHz celeron M (as has been noted earlier): http://www.youtube.com/watch?v=Wedw701Gy8s [youtube.com]

Also, since there's so much integration within KDE, the RAM usage doesn't jump that much when using an application. I'm running KDE 3.5 with opera, kmail, ktorrent, amarok, and yakuake, plus all the services, and I'm at about 300MB of RAM used - not much higher than when none of the apps are running.

Re:setup

[pherthyl (445706)]

I did in fact use the setup I described... and you can check that imacs were sold with 32 megs on wikipedia. Please check your facts before calling me a lier.

Sorry but you are completely full of shit. OS X does not run for any reasonable definition of "run" on 32mb of RAM.

Have a look at the minimum requirements for OS X 10.1 which you say was the most efficient OS X.

http://en.wikipedia.org/wiki/Mac_OS_X_v10.1#System_Requirements [wikipedia.org]

Notably: RAM required 128 megabytes

And you're saying you did OpenGL development on a quarter of the minimum requirements. Riiight.

Troll. Nice one though. The moderators believed you at least.