16

I installed xubuntu desktop on top of ubuntu 12.04. The usage of RAM and CPU is in general lower than in ubuntu. But when it comes to some serious computations I did not see much of a difference between the two. For example, I ran a Matlab program that took 99 seconds to run on ubuntu and 98 seconds on xubuntu. Then I ran Geekbench benchmarking and the overall scores ubuntu and xubuntu are 2996 and 2997, respectively. My laptop is a 4 year old dell 1535. Would things run faster on a clean install of xubuntu?

  • 1
    You should try Lubuntu. It is even faster than Xubuntu. It runs on my two-generation-older (Centrino) laptop as fast as Unity does on my Core2Duo desktop. – MarkovCh1 Aug 7 '12 at 3:31
  • it is faster if initially ubuntu with unity made it slower. if you are used to ubuntu-unity (or kubuntu-KDE) on a powerful new computer running perfectly, you will not see the same decisive difference that is really obvious on an older 1GB-RAM machine when trying xubuntu (and even more so, lubuntu). With 2GB of RAM or more i guess the advantage of using lighter desktop environments is gone. more here:askubuntu.com/a/181304/47206 – user47206 Dec 16 '12 at 20:32
  • I didn't notice some drastic changes in speed when changing to Xubuntu. One thing I noticed, is that the computer does not overheat like in Ubuntu. I also use Matlab, and in Ubuntu, on hard computations, CPU temperatures were around 90 degrees celsius. On Xubuntu, it barely passes the 65 degrees limit. I was surprised to see this, and it is a pleasure to work on a silent laptop. :) – Beni Bogosel Jan 5 '15 at 10:06
18

This is an issue that I looked into pretty seriously because I had an older laptop that was not running very smoothly at all. It took forever for programs to open and execute various tasks, and got slower and slower during the time I would use it, forcing me to reboot periodically just to refresh the system. I experimented between Xubuntu and Ubuntu, and kept track of times to measure the actual difference.

The technical answer is, yes, Xubuntu is faster than regular Ubuntu.

But, the more accurate answer is, no, probably not in any way that will make a difference to you.

If you just opened up Xubuntu and Ubuntu on two identical computers and had them sit there doing nothing, you would see that Xubuntu's Xfce interface was taking up less RAM than Ubuntu's Gnome or Unity interface.

But, so what? You don't just run computers to have them sit there. And here's where the rubber hits the road.

Xubuntu isn't just Xfce, it's also a whole suite of bundled applications to handle most tasks you want to do. For example, if you wanted to do word processing, Xubuntu comes with Abiword preinstalled. On regular Ubuntu, you're offered LibreOffice Writer by default.

Abiword may take less resources, but you pay for that with less features. And while I probably don't use 90% of the features available in Writer, the difference was noticeable enough that I found I would rather use Writer.

Which brings us to the real issue when determining whether Xfce is faster or not. Even though underneath Writer, Xubuntu's Xfce interface is using less resources than Ubuntu's Gnome or Unity, the question you're coming down to is this: Is Xfce using less resource to a degree that would make a difference in how fast LibreOffice Writer runs?

The answer is, no, not really. LibreOffice uses so much resource on its own that it quickly eats up any savings in using Xcfe. For all intents and purposes, when running something like LibreOffice on Xubuntu, the savings that Xcfe offers you is so negligible as to not matter.

So, if you're content to stay within the suite of applications offered by Xubuntu, then you might see some difference.

If, however, your hope is to run more or less the same programs as you would with Gnome or Unity but are hoping to get an appreciable performance boost, then there is no meaningful difference. By "meaningful difference", I mean a difference in how fast programs open or execute tasks to a degree that a human can tell the difference. I don't know about you, but saving .03 seconds if it takes a minute and a half to open a program doesn't mean anything to me.

In the end, I found that buying more RAM for my laptop - going from 512MB to 1GB - was the real solution I was looking for. The dream is that by getting an optimized desktop manager you can breathe new life into an old machine, and I'm sure some people have a situation where they pulled that off. But my experience was that trying to tweak the interface and applications ultimately had too many feature drawbacks for too little return.

So the question about which is faster, Xubuntu or Ubuntu, comes down to this: what applications do you want to use? If you are willing to go with the whole Xubuntu suite and all the bundled applications that are meant to be more minimal in resource use, then you might notice a performance improvement. But as soon as you start using bigger applications, the interface savings will become moot.

19

The Kernel is the same, the underlying system is the same, the tools are the same, the desktop is different. it's the same Operating System it's just using different Desktop Environment.

  • 1
    Exactamente!!!! – mikewhatever Aug 7 '12 at 4:57
  • 3
    This is perfectly accurate and technically correct, and provides no information that would help in any decision making. Reminds me of that old Microsoft joke. – Questioner Nov 2 '14 at 3:01
4

In my experience: Xubuntu is a lot faster. In my netbook: Samsung NC110 A02 couldn't stand unity (everything was slow, much clipping etc.). On the other hand: my Asus F3ka have been better with Ubuntu (12.04) but also, slow as hell. It took hours to do simple things, also I don't like how they have hidden most of settings (since 11.10 it isn't a customizable O/S). I tried Xubuntu. And since then I have it in all of my computers. Fast, stable, customizable and light OS. It eats only about 190 MB of RAM. I repaired a lot of old computers with it. I could not even image what Ubuntu 12.04 will looke like if on them.

  • It is faster if you alternate between application a lot. But if all you are doing is running Matlab scripts or other batch tasks then you won't notice a difference. That's why @ahamkah reported equal times for running matlab. – rpmcruz Nov 8 '15 at 12:40
  • Going forward: it wouldn't make any difference if pc with ubuntu and xubuntu are both powered off. – bartekordek Nov 9 '15 at 15:02
  • When you run a batch process and no other process is running, only the kernel and a few other things matter. CPU time and memory is dedicated to that process and so the desktop is largely irrelevant. – rpmcruz Nov 10 '15 at 16:17
  • If you don't run it then yes. – bartekordek Nov 12 '15 at 7:07
3

If you are doing the same 1 billion multiply operations, without accessing windows or any other operating system function, and nothing else is going on in the background, it shouldn't really matter what operating system or desktop environment you are using Ubuntu, Red Hat, Windows. The hardware is the constraint. If whatever operating system is bringing the program into memory and then letting the processor or processors do their best your hardware will mostly determine the speed (assuming all of the operating systems handle the multiple processors, hyperthreading, and so on).

When compiled with a better compiler, or compiled with more optimization, you may not be performing quite the same number of operations, however. Ubuntu, Xubuntu, Red Had, probably are using about the same compiler. It may make a difference how matlab is compiled between Ubuntu/Debian and Windows, however.

When you get into doing reads and writes to disk you shouldn't necessarily expect a benchmark to work the same on Windows, BSD, or Linux, but should probably expect the Linux numbers to be about the same, including Ubuntu and Xubuntu, (but the kernel's will be configured differently and that may make a difference).

If you are putting someone at the desktop and have taught them how to perform some task, Ubuntu and Xubuntu will use different resources when they perform them later, Xubuntu probably fewer resources. The individual may take more or less time to perform the task with either Ubuntu or Xubuntu. In this case it makes all the difference what you decide to measure.

2

Yes, the desktop environment is faster than the unity shell on top of gnome.

  • 2
    @StephenMyall This does provide details. It says the DE (and not associated apps, for instance, which also differ) is faster. This isn't the best possible answer and it certainly could be improved. But it does provide enough information that someone can search the web to find specific information about how the different desktop environments perform. Furthermore, as an answer, it can be improved; as a comment, it couldn't be. In my opinion, only when something does answer the question is fundamentally an opinion should it be converted to a comment. – Eliah Kagan Aug 14 '12 at 10:52
  • 1
    The Unity shell is interwoven with Gnome applications. But it is on top of GTK, not Gnome. – rpmcruz Nov 8 '15 at 12:41
1

I came here looking for a technical answer. I didn't find that. I'm looking into this for an Ubuntu server install on a pretty low resources box.

So here is a better technical answer with resources comparison: http://mylinuxexplore.blogspot.com/2012/10/ubuntu-1210-vs-kubuntu-1210-vs-xubuntu.html

Hope this helps someone.

0

Probably yes. Regular Ubuntu brings Unity and a lot of more programs that are kept executing in memory, while Xubuntu brings less of those.

If you want a really fast and less memory consuming, you can try download Ubuntu server and install only what you need, or even use a WM likde Openbox instead of a DE.

-1

My friend, Matlab would be slower on any system that's what it is. Matlab is written in C++ and you are not compiling it from source for your system to build binaries. Its all compiled code and it is one of the slowest scientific computation environment. It doesn't really matter whether you run it on an RHEL, ubuntu or xfce version of Ubuntu. Don't look for these kind of bench marks. Look for things like sensor outputs throughout the day for normal and extreme usage in your system. Smart data is available for sensors. Just see how the resources are being used through out the day. Don't do benchmarks with Matlab. Atleast write a real scientific code in c++ and fortran and run it in both environments and see if there is any difference (would be pointless still if you are gonna do something stupid do it right). If you are the looks person use Ubuntu. If you are the stability person go for xubuntu. It hardly crashes unless you do something stupid.

  • Matlab is Java actually. – rpmcruz Nov 8 '15 at 12:41

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.