Slashdot Mirror


Preliminary Benchmarks: Unity vs. Gnome-Shell

fatalGlory writes "Despite some initial reservations about Gnome-shell, it appears to be coming out very nicely. In some preliminary benchmarking tests I've been conducting, Ubuntu's Unity desktop on 11.04 Natty uses roughly double the memory that Gnome-shell uses."

27 of 258 comments (clear)

  1. I feel like trolling... by johnsnails · · Score: 2

    I feel the urge to troll because I loved using ubuntu until 11.04, but now have switched to kubuntu 11.04. Instead I will say I look forward to the continual improvements that will be made to both unity and gnome shell.

    1. Re:I feel like trolling... by vajorie · · Score: 2

      now have switched to kubuntu 11.04

      If you don't like Unity, don't use it; Gnome Shell 2 is still in Natty, and it works fine.

      -- There's no -1 for "I don't get it."

      What part of that did you not get? ;)

    2. Re:I feel like trolling... by thepike · · Score: 2

      Why does everyone bring this up? That support is leaving in 11.10 so the argument only works as a stopgap for another few months. Sure you can choose that now, but in the long run if you don't want to use unity (and who does?) you have to switch someday.

      The big question is where will people go; slackware? Fedora? Xubuntu? I know Ubunutu is trying to get more mainstream, but they'll lose some of their hardcore users, and I have to wonder how that'll affect their devs.

  2. Unity sucks by Anonymous Coward · · Score: 2, Informative

    What more do you need to know? I installed (and fully updated) Natty this weekend, and crashed it 3 times in 20 minutes with different Unity bugs. Then, I hit up the goog, and found out how to get my classic gnome interface back (it's in a dropdown at the login prompt). Waste of 22 minutes, if you ask me. I can't imagine how much time the Unity devs wasted on that crap.

  3. Why do people underthink memory usage? by Anonymous Coward · · Score: 4, Insightful

    It's really not that big a deal these days, you want to use memory, because memory is FAST, and in comparison to the old days, dirt cheap. Loading things into memory is not an automatic sign of bloat, sometimes it is a sign of doing what you should, putting memory into use.

    1. Re:Why do people underthink memory usage? by rubycodez · · Score: 2

      I want to use my memory myself, not have it used up by some bloated piece of shit window manager

    2. Re:Why do people underthink memory usage? by water-and-sewer · · Score: 2

      I'm hip with the "put all your memory to use" paradigm - that makes perfect sense. However, I want that memory to be in use for my applications. The Window Manager is just a utility, not an end in itself.

      At my place, my primary box is an x86_64 box with 2G of RAM and it flies. It would easily run Natty, if I cared to install it (I don't like Gnome). But I still boot up my old PIII 555 Mhz with 128MB RAM just to see how far we've gotten. My conclusion after 11 years (bought that old PIII in 2000) is that WE HAVEN'T GOTTEN FAR. That old KDE 3.1 desktop does everything I want a desktop to do, and in some ways WIndowmaker does as well. Fast forward to the land of beautifulu rendering, compiz effects, GPU-accelerated whiz-bang, and in terms of functionality I'm not much better off. I'd far prefer to use the old desktop and have that RAM available for better apps. My desktop and window manager don't deserve all that RAM for themselves. And if your solution is to buy more memory, you're mistaken, because the developers will quickly appropriate that new RAM for themselves in the next edition, for an even shinier-but-no-more-useful windowmanager.

      --
      If this were Usenet, I'd killfile the lot of you.
  4. Not a useful comparison in any regard by LordLimecat · · Score: 4, Insightful

    Absolutely worthless comparison, as it compares vastly different distributions. He isnt even comparing 2 debian based distros, or trying to control for different running services; why is there not even an attempt to isolate the memory usage of the DE / WM?

    Perhaps this could have been useful as a comparison of distro memory usage, but even in that it fails-- its comparing an installed Debian distro to live-CD based Fedora; why wasnt fedora installed and compared (perhaps using VMs?), or Ubuntu run from LiveCD?

  5. Hell of a benchmark by Osgeld · · Score: 2

    just saying

  6. Wow by Chuck+Chunder · · Score: 3, Insightful

    It's as if someone designed that "benchmark" to be criticisable in every possible way!

    --
    Boffoonery - downloadable Comedy Benefit for Bletchley Park
    1. Re:Wow by Chuck+Chunder · · Score: 2

      It would be if I was American.......

      --
      Boffoonery - downloadable Comedy Benefit for Bletchley Park
  7. hmm but linux doesn't crash by decora · · Score: 3, Funny

    everyone knows that! it's built on a solid, stable unix foundation, with a 'keep it simple' philosophy that guarantees performance and stability! with 12 overhead cams and a dual plated stainless steel cooking surface, your family will be sure to love the new Unity.

  8. Bad test, right result by Tester · · Score: 3, Informative

    He says Gnome-Shell uses software rendering.. It's not true, fedora ships free 3D drivers for Intel, AMD/ATI and NVidia now.

    Also, why use VLC when he could use Totem on all 3.And Why does he use apt-get in his test ?

    Even though I like the result, it seems like a pretty lame test.

  9. Re:Why not Gnome on Ubuntu? by rubycodez · · Score: 2

    Canonical's Gnome 3 PPA is a mess right now, spontaneously explodes, and when you try to go back to either classic or Unity there will be issues PPA Purge doesn't fix. Better spend your time looking for your next distro in 5 months when you won't have a classic mode. Debian, Mint, Arch, Puppy, Pardus, Mandriva, Fedora, FreeBSD, Gentoo, Sabayon, PCLinuxOS, PC-BSD, MEPIS all will be in better shape than 11.11, try 'em out and pick one.

  10. Comparing total memory usage is stupid by fudoniten · · Score: 4, Informative

    Others have correctly pointed out that comparing memory usage on two different distros is pointless. On top of that, comparing total memory usage is stupid.

    Look, you have memory in your system to be used. If you dug into it and found out that most of that memory consisted of massive, unused libraries, duplicate code, empty datastructures, or garbage that wasn't getting cleaned up, then sure, you could give it a hard time. But if it's full of cached images and icons so that the interface can be quicker and more responsive, well, isn't that why you have all that RAM?

    A perfect program/OS would very quickly gobble up all available memory by storing and caching useful stuff...and then free it up the instant it was needed elsewhere. That turns out to be harder than it sounds, since procs generally don't know or care about totally memory usage, but still, the ideal should not be the opposite extreme.

    1. Re:Comparing total memory usage is stupid by wrook · · Score: 2

      But if it's full of cached images and icons so that the interface can be quicker and more responsive, well, isn't that why you have all that RAM?

      A perfect program/OS would very quickly gobble up all available memory by storing and caching useful stuff...and then free it up the instant it was needed elsewhere.

      The problem with filling up all your memory is that you don't know what the priority for memory use is. Is all that RAM on my system there to cache icons, or is it there to cache my database? As an application developer, you don't know how important your program is to the system compared to another program. As an OS developer, you similarly don't know. Chewing up all the available RAM so that one application is optimally fast, may be exactly the wrong thing to do if, as a user, I don't care about the speed of that application all that much.

      RAM is a finite resource. Allocating and deallocating memory takes time. So even if we free up memory the instant we realize someone else wants it (pretty much impossible, but let's say we can do it), it will still take take. Swapping out to disk also takes time. Filling up all your memory pretty much guarantees that you will increase latency whenever you need to allocate memory (starting a new program, or changing between programs, etc, etc).

      As an application developer, you should be aware of the impact of memory that you using. Do you *really* need to cache that icon which is really never going to be shown again in all likely hood? Especially when talking about a desktop shell, I certainly don't want it eating up all my RAM. If my shell eats up all my RAM, then *any* application I run will result in swapping. Suddenly, the start up time for *all* my apps has been increased. Programs that like this should use as little memory as possible while still performing adequately quickly. This allows the applications *that do useful things* to use as much memory as possible and run as quickly as possible.

  11. Re:Test the thing that matters: Usability by Glonoinha · · Score: 2

    I just installed 11.04 this evening.
    The reason /, has it in for Unity is : Unity sucks. If you don't already know where to find something, you will never find it. In my 20+ years of using computers I've never had a UI hide the details of getting shit done nearly as well as Unity. Sure thing - if all you want to do is open Firefox or an office suite nobody on this board has ever used - it is pretty damn slick. But want to do anything 'normal' besides that (or God forbid : advanced!) and unless you know exactly what you are looking for before you start, you are fucked. There is no breadcrumb trail. There is no drilling down. You are pretty much shit out of luck.

    10.04 - my favorite now, and may be for a while it seems.

    --
    Glonoinha the MebiByte Slayer
  12. Re:"Every software engineer should be a creationis by SanityInAnarchy · · Score: 4, Insightful

    And hey, big surprise, comments disabled on that article. "You must log in to post a comment." And even if I did want yet another account just for commenting on a single blog, I don't see anywhere to register.

    Just for fun, I'll respond here. I might even try to email it to him, if that works.

    There are three fundamental entities that make up our universe, matter, energy and information.

    I'm not sure information is an "entity" in any relevant sense. It's a phenomenon. More in a moment...

    Now, creating and communicating information is demonstrably a mental process, requiring an intelligent mind to create and to receive the information.

    Actually, it's demonstrably a physical process, one which can be performed entirely by machines, unless you are willing to describe my laptop as an "intelligent mind." But it depends what you mean by "information", in this case, as you point out:

    In this way, information is distinct from data, which Shannon unfortunately referred to as "information" in his work on statistical-level information-theory, leading to the present ambiguity.

    If that is the way in which information is distinct from data, then I work with a hell of a lot more data than information. My computer creates, interprets, communicates, and manipulates all sorts of data that no "intelligent mind" will ever touch, unless, again, you're willing to allow that my cell phone is an "intelligent mind."

    This reality has been demonstrated amply in the book by Gitt and is expressed in a streamlined form in this lecture by Wilder-Smith.

    I'm not willing to buy and read a book, but maybe I'll listen to the lecture.

    Yet, so many software engineers remain evolutionists.

    ...what? Unless you're referring to the arguments you referenced via an amazon link and an mp3 file, I see nothing in your argument which requires intelligent design or negates evolution. Even if I accepted your premise that information must have an intelligent designer -- sorry, a god -- as its originator, as a "software engineer," I'd hope you understand that humans can and have written programs which simulate the genetic process at various levels -- why, then, could this god not design evolution as part of the "program" of the universe, fire it off and let it run, exactly as human beings do all the time?

    Despite interpreting and often designing language conventions every day, very few software engineers seem to have considered the implications of language and information-theory for genetics, biology and metaphysics.

    Again, out of the blue, you're introducing a new topic -- languages -- along with committing a stupidly trivial fallacy. Just guessing here, because you didn't actually deliver an argument, but if you did, I imagine it would look like this:

    1. Humans can create languages.
    2. Humans have intelligent minds.
    3. Given 1 and 2, intelligent minds can create languages.
    4. DNA is a language.
    5. Given 3 and 4, an intelligent mind can create DNA (the language).
    Therefore, only an intelligent mind could have created DNA.

    Both 5 and the conclusion are absurd on their face, and I hope you can see that. 5 is fallacious because 3 asserts only that intelligent minds can create languages, not that they can create all languages. Even if 5 were sound, the conclusion is fallacious because 5 asserts only that an intelligent mind can create DNA, and not that only an intelligent mind can create (or could have created) DNA.

    And again, what about this falsifies evolution? If it worked, it would falsify abiogenesis. Evolution can happen without intervention once we have DNA, just as a program can run without human intervention once we start it running.

    Why is this? Well that's a discussion for another post.

    If you're going to m

    --
    Don't thank God, thank a doctor!
  13. Re:Worthless shit by hawguy · · Score: 2, Interesting

    Worthless shit written by a christfag.

    Well to be fair, he had to rush it out before he's Raptured this weekend. Otherwise he probably would have spent a few more days on it.

    (for those that read this comment years from now: http://en.wikipedia.org/wiki/2011_end_times_prediction)

  14. smartphone interface. by MadMaverick9 · · Score: 2

    both gnome 3.0 and unity look like they were designed for smartphones and tablets.

    The reason why the devs did this is because users are familiar with the smartphone interface, so putting the same interface on the linux desktop would make transition easier.

    But - a 1900x1200 pixel pc desktop is not the same as a tiny 100x100 pixel smartphone screen. pc desktop users with a normal size keyboard and big sized monitor have very different requirements.

  15. Re:Test the thing that matters: Usability by Sepodati · · Score: 2

    Right click on the Applications icon in the launcher and your sorted categories are right there just like you use to have. Or use the drop down option when typing into the search area (to the far right) to limit to specific categories. Or just click Applications and expand the Installed section and everything you have is right there.

  16. While I'm at it... by SanityInAnarchy · · Score: 2

    So I just killed about an hour listening to the mp3 and thinking about how I'd respond... not much new there.

    Starts with a few intelligence-insulting analogies, like "Bricks don't build houses." Yeah, bricks also don't reproduce.

    "We ought to be able to put the contents of life into a test tube and see that it would build life." In other words, if we can't put something in a test tube and get something out of it, we require an intelligent agency for that process? Seriously? If we assume this guy isn't as dumb as he seems to be, and that by "test tube" he meant "any artificial apparatus", then we have experiments which show that things like amino acids can be formed through entirely natural, mechanical processes.

    He then goes on to describe intelligence as the "third column of the universe" -- that is, matter, energy, and information -- and also that "nobody knows how to define it." Actually, we can define information, and we can do so quite precisely. If you can't, then maybe you shouldn't be claiming it's a fundamental property of the universe. When we do use precise definitions, we find that either DNA is not information, or that information does not require a designer, but this argument is almost deliberately vague.

    He then goes on to talk about codes, pointing out that SOS is only a code by convention, and that if we didn't have that convention, SOS wouldn't communicate "distress" or, really, anything at all -- that sequences of syllables, or letters, or nucleotides, are meaningless unless some meaning is assigned to them -- that in this case, without the ribosomes, a particular sequence of DNA wouldn't correspond to a particular sequence of DNA into a protein. So he's already given up his premise, which he's repeated constantly, that a convention can only be established when two intelligences come together and agree on something, unless he's willing to call ribosomes "intelligent".

    So why couldn't ribosomes be what assigns the "meaning"? Why is an additional entity required? Well, he says, ribosomes are needed to decode DNA, but ribosomes are themselves encoded in DNA, so there's a chicken-and-egg problem. Sounds a bit to me like irreducible complexity, and we know how well that's worked out. It's exactly the same argument -- "Half an eye is useless, so the eye couldn't evolve by steps!" It turns out that half an eye is useful, just not for quite the same things a whole eye is -- and in this case, there are entirely plausible ideas for how this mechanism could itself have evolved piecewise. (Yes, evolution can happen without DNA, it's just that DNA is the mechanism that won.)

    Finally, and this is where I'm really disappointed in the supposed "software engineer" who linked to this -- the lecturer starts talking about the information that's encoded in DNA, and how much work we've had to put into just finding out what the sequence is -- that "our best computers can scarcely..." what, assemble DNA sequences? Takes some pretty heavy hardware (you want at least 24 gigs of RAM), but it's far from our best. Yes, the Human Genome Project was massive, but these days, it's fast and cheap to sequence stuff -- you could almost do it as a hobby, it's that cheap. On top of all of this, when he says that the amount of information in DNA is "almost infinite" -- there are a number of things I could nitpick, but really? It's not only finite, it's trivial by modern standards -- the entire genome is, according to Wikipedia, just over 3 billion base pairs. If each can be A, T, G, or C, that's 4 options, so you can store it in 2 bits, or 4 of them to a byte, so about 750 million bytes -- just barely doesn't fit on a single CD-R.

    None of the above paragraph really affects his argument, but it shows just how little he knows about information in the modern age.

    Is there anything about this argument that is worth the time I just put into it?

    --
    Don't thank God, thank a doctor!
  17. Re:Why should we care? by hedwards · · Score: 2

    Because, most people don't load up with 4gigs of ram (or more) just to run the desktop environment. Sure Unity doesn't use anywhere near 4 gigs, but every megabyte that's used by the desktop environment is a megabyte that's not available for whatever it is that you're using the computer for.

    But more than that, the sort of sloppy coding practices which lead to bloat also lead to other problems, ones which are of larger importance.

    At the end of the day, I know that it's not realistic to eliminate all the bloat, but there shouldn't be huge gobs of RAM being wasted because the developer couldn't be arsed to optimize things.

  18. Re:Desktop Not Redrawing by r6144 · · Score: 2

    I think support for Intel 8xx graphics has been rather poor since the change nearly two years ago to use the GEM infrastructure. My i845 locked up about once an hour in Fedora 12 unless I revert to the non-accelerating vesa driver. This turns out to be a GPU bug which happens to be triggered very rarely with the old drivers. A few patches have been found to work around this problem, but I haven't tried them. AFAIK comprehensive GPU documentation from Intel is only available for i810 and i965, so for everything in between, it is rather difficult to have bugs fixed, and Intel doesn't seem to have much interest in having their old hardware continuing to work, either.

    If you just want a usable desktop (not accelerated graphics), use the vesa driver. If you want it to be fixed, report it to freedesktop.org and expect to spend some time compiling kernels.

  19. Re:Not only that... by erice · · Score: 3, Informative

    but since when has memory footprint been a "benchmark?" Really, we're talking roughly half a gig here, and who's running these on a system without at least 2?

    Most netbooks are 1GB. half a gig is a lot just to run the desktop on a machine like that.

  20. Re:Not only that... by TheLink · · Score: 2

    I think something is wrong somewhere. Windows XP SP3 can run very comfortably in 512MB virtual machine. And the 512MB virtual machine doesn't even have to use all 512MB of the host machine's RAM.

    So what does the GUI give you for all that memory use? Or is it mostly bloat?

    I doubt it provides "faster performance". Seems to me most GUI designers either don't really care about performance or are clueless about it. They keep creating extra steps to do common stuff. Or insert artificial delays so that they can show flashy animations (e.g. click, pause for flashy animation, then only click to launch/activate). OSX's expose is actually slower than just clicking on the taskbar button representing the window you want to "raise". And moving your hand to click on the task button is often slower than "alt+<number>".

    By the way, to the GUI designers out there, if "GNU screen" manages windows/tasks faster/better than your GUI, you are doing something wrong. A GUI that is friendly to new users does not have to be slower than "screen" in the hands of expert users.

    --
  21. Re:Not only that... by TheRaven64 · · Score: 2

    I was under the impression that the point of Unity was to be useful on small form-factor devices. I have a little ARM-based machine, with a 10" screen, which shipped with Ubuntu. GNOME sucks on it - most of the dialog boxes don't fit on the screen - so I was interested to see if Unity was an improvement. Turns out? It won't even start Unity, because it claims my system is not up to it. Designing a UI for small devices that requires a high-end machine to run seems a bit silly. Even when we've all got quad-core Cortex A9 chips in our netbooks, no one is going to want to use the environment that reduces your battery life...

    --
    I am TheRaven on Soylent News