After reviewing Ubuntu 15.10 a few months ago, I came up with an Ubuntu (15.10) flavor comparison as well. So after reviewing Ubuntu 16.04 LTS, and especially since this is a LTS (Long Term Support) release, I decided to come up with yet another Ubuntu 16.04 LTS flavors comparison that involves Ubuntu, Kubuntu and Ubuntu GNOME because they come with the 3 main desktop environments of GNU/Linux: Unity, KDE Plasma and GNOME.
But just like the previous one of its kind, this too will be based on the performance aspect and the stability of the each operating system, and I won’t talk about the new features of the desktop or the applications. But as a general introduction, all three flavors use the Kernel 4.4 & Xorg 1.18.3. Ubuntu’s Unity desktop features the version 7.4.0, Kubuntu features the KDE Plasma 5.5.5 (and KDE Applications 15.12), and Ubuntu GNOME features GNOME 3.18 release.
So consequently, you’ll find performance comparison between Boot-up Times, Memory Usage, System Responsiveness and Shutdown Delay, mainly. I’ve also added data concerning hardware recognition & ACPI, and few other performance related data which you’ll get to read.
This is How I Did it…
Before posting my judgement, I did my best to gather performance related data as accurately as possible, including the level of the stability of the OS, that although not to the extreme I did with performance. I followed the following ‘protocols’ for achieving those objectives:
- I installed each operating system on the same hardware (obviously).
- I have a single partition created for the sole purpose of testing new GNU/Linux distributions. And all three flavors were installed into it, one after the other.
- Before installing each operating system, I securely wiped out the partition so that I end up with ‘clean’ sectors.
- After installing, I let the newly installed OS to boot 5-6 times for letting things to settle down (because there are some applications and especially utilities such as ‘ureadahead‘ -- a tool that significantly speeds up boot-up times of Ubuntu, that require a couple of reboots to work properly).
- I tried not to ‘touch’ (manual tweaks etc) anything on the OS. Although I did add the system monitor (I used it to measure the memory usage) to the status-bar or to a launcher (if one is present). I did that because if I were to open it each time using the start-menu or otherwise, it would’ve increased the memory usage, thus negatively affecting the accuracy of the memory usage reading.
- For further enhancing the accuracy, of each test, I took five samples (except for the system responsiveness test. But on two occasions I had to run it twice).
- Then I evaluated the hardware recognition (I don’t think anyone needs an introduction to what this means) & stability. For evaluating the stability, I made sure to use the newly installed operating system as my main operating system quite extensively, for at least two days. In this instance, I ran Ubuntu 16.04 LTS for about 5 days, Kubuntu 16.04 LTS for about 2 days and Ubuntu GNOME 16.04 LTS for about 4 days. I looked for things like major bugs that could make the OS unstable (for instance, in Ubuntu 15.10 I came across a bug that forcefully logged me out of the desktop every now and then. And when that happened, I lost any unsaved data!) and application crashing etc.
Below is my Dell laptop that I used to test these Ubuntu flavors:
Intel Core i3-2330M CPU, Intel HD 3000 GPU, 4GB RAM (DDR3), Toshiba 7200 RPM (320GB) SATA HDD, Intel N-1030 Wireless adapter, Realtek network adapter ('RTL8168'), LED display with 1366x768 resolution (60Hz/60FPS). It's a Dell Vostro V-131 notebook.
The Kbuntu 16.04 LTS Installer
Both Ubuntu 16.04 LTS and Ubuntu GNOME 16.04 LTS use the same installer. It’s called ‘Ubiquity’. And as I’ve said many times, this is one of the most reliable and intuitive installers out there. In all these years of testing, I don’t recall even a single major issue that I came across when using it.
Kubuntu 16.04 LTS on the other hand uses a different installer called ‘Calamares’. It’s being used by many KDE based distributions (Kubuntu, KaOS, Manjaro, Netrunner etc). Just like ‘Ubiquity’ it too is a fantastically intuitive program, but unfortunately, couple of times, I came across a major issue with this installer that made it almost impossible to install the operating system that used it, if it wasn’t for my experience (didn’t put it in an egoistic sense 😛 ).
What happens is that the installer totally crash at ‘Keyboard Setup’ and aborts the installation. I first came across it as far as in 2014 while I was reviewing Netrunner 14 Frontier. I’ve ran across this issue in a previous Kubuntu release too (15.04, I think). But interestingly, I’ve used few other KDE based GNU/Linux distributions that used Calamares, in between, and Calamares ran just fine on those occasions. But with Kubuntu 16.04 LTS this happened once again, as you can see from the below screenshot.
If you come across this issue, then this is the way to fix it. Make sure the computer is not connect to the Internet (turn off Wi-Fi, 3G etc). Then reboot the computer and run the installer again. When you come to setting up the Time Zone, leave the default Time Zone untouched.
Then when you get to setting up the keyboard, also make sure to not change anything. This should fix the issue (it has worked for me in the past occasions and it worked this time too).
Sure this time the installer won’t be able to detect your Time Zone, language and keyboard type automatically, but you can fix it by simply changing them from the ‘System Settings’ after installing the OS anyway.
Do I consider this to be a major issue?
Obviously! After all, this is a LTS release in which stability and security outsmart other ‘exciting’ elements (up-to-date software, major new features etc). I was lucky to find a way to bypass it, but a new ‘Linux’ user wouldn’t have even stood a chance.
I also noticed, although this isn’t a major issue for me at all, that some of the widgets of the Plasma desktop (CPU Load Monitor, Network Monitor etc) didn’t quite work as expected. They failed to correctly render the widget layout, as you can see from the below image.
Update: Actually, the problem was that I wasn’t using the correct method for re-sizing the widget because apparently, after bringing the mouse pointer over a widget, now you have to press and hold both the Ctrl key and the left mouse button to get the widget’s control-bar. Not the most intuitive way I must say. It worked nonetheless, and a big thanks goes to Michal (comments) for pointing that out.
Both Ubuntu 16.04 LTS and Ubuntu GNOME 16.04 LTS use the same software center. And as I mentioned in my Ubuntu 16.04 LTS review, this software center can’t install ‘.deb’ files, and it’s the same under Ubuntu GNOME 16.04 LTS as well. The below screenshot is from Ubuntu 16.04 LTS (of course), I just forgot to take a screenshot in Ubuntu GNOME, and after writing this review, I immediately wiped out the data on that partition (I added this info to this review after few hours of publishing the article originally). Sorry about that guys.
Kubuntu 16.04 LTS uses a software center of its own and I was easily able to install a third-party ‘.deb’ package (Google Chrome to be precise) without any issues whatsoever. Excellent.
As you can see, Ubuntu 16.04 LTS is the fastest to boot, beating both Kubuntu and Ubuntu GNOME by 35% & 29.5%, respectively.
Personally, except for Ubuntu GNOME, I’m not surprised by these readings. Kubuntu (or most KDE based distributions in general) has never been that fast while booting. To its credit however, Kubuntu 16.04 LTS boots 21% faster compared to its predecessor, Kubuntu 15.10.
I came across an issue that tremendously slowed down Ubuntu GNOME 16.04 LTS and I have good reasons to believe it’s plymouth’s (the tool that displays the boot-logo) fault. What happened was that, precisely 6 out of 10 times (yes I counted them) when booting, plymouth failed to show the graphical boot logo of Ubuntu GNOME and only showed a black screen. When this happens the boot-up process gets stuck for about 15-17 seconds! After this long & agonizing delay the system starts to boots up again from where it left off, though. And justifying my suspicion, when this happens, within a very short period after I log into the desktop, I get an error saying ‘plymouth’ has crashed.
But 4 out of those 10 times where the plymouth doesn’t fail to show the boot-logo, Ubuntu GNOME 16.04 LTS boots quite fast. According to my numbers, the boot-up times were anywhere around 32 seconds to 34 seconds which is very close to Ubuntu 16.04 LTS’s boot-up times. And it makes sense too because unlike Kubuntu, both Ubuntu and Ubuntu GNOME have so much in common. Not only they’re all based on the Ubuntu Core, but Ubuntu and Ubuntu GNOME share the same desktop applications, and Unity even uses lot of GNOME technologies for rendering its desktop shell as well.
Anyhow, coming back to the issue, yes I tried updating the system, but the issue is still there. Is this a big issue? No, not necessarily. But keep in mind that this is a LTS release, therefore, these types of issues can be frustrating from that point of view.
Update: After 2 days of using Ubuntu GNOME 16.04 LTS, I just received an update and it seems to have fixed the whole situation actually. I’ve boot into the OS many times after updating and have been using Ubuntu GNOME for 2 more days, so far haven’t seen it, not even once. And the system now boots within 32-34 seconds as well.
Kubuntu 16.04 LTS used the least amount of memory compared to both Ubuntu 16.04 LTS (3.1%) and Ubuntu GNOME 16.04 LTS (10.5%). Kubuntu was also impressive when compared to its predecessor where the memory usage had increased only by a small amount (3%), where both Ubuntu (48.1%) and Ubuntu GNOME’s (66%) memory usage has increased by a lot, compared to their predecessors.
Note: I usually add the power usage (idle) next. But as I mentioned during my Ubuntu 16.04 LTS review, the battery of my laptop is dead, therefore I don’t have any data concerning the power usage. Sorry about that.
CPU Usage at Idle
A good operating system at idle should keep the CPU usage around zero, although, there’s nothing the OS can do if a buggy program eats up the CPU. Low CPU usage at idle also means less power consumption as well.
I’m happy to say that all three Ubuntu flavors did extremely well here. As you can see, at idle the CPU usage was very close to zero and remained so for lengthy time-frames, although here & there some app would pop-out which is pretty normal.
Ubuntu 16.04 LTS:
Kubuntu 16.04 LTS:
Ubuntu GNOME 16.04 LTS:
Hardware and ACPI
I do admit that the hardware on my Dell laptop is more ‘Linux friendly’. Ubuntu specifically has been able to run on them quite well in my experience. And it’s no exception this time around either, because all Ubuntu flavors were able to correctly recognize and configure my hardware (except for the fingerprint reader. It has never fully worked on GNU/Linux. The manufacture doesn’t even care to release a driver!). But keep in mind what I said. My hardware is more ‘Linux friendly. Therefore, if you have different hardware (such as proprietary Nvidia & AMD/ATI GPUs, sound cards etc for instance) you may not experience the same level of stability I experienced.
Other features such as Suspending and waking up also worked like a charm on all three flavors. Also, as far as the application crashing is concerned, they were all very impressive, even in Kubuntu because KDE has never been extremely stable in my experience. For the 2 days I used it, not even a single app crashed. Kudos to KDE developers.
The only app crash I saw was the ‘plymouth’ incident that I mentioned earlier which happened in Ubuntu GNOME, but that too was fixed with a recent update.
The hard disk drive is the primary storage device of a computer. But unfortunately, it’s also one of the most slowest (whether it’s a rotational disk or a solid state drive) component of a computer. If all the reads and writes that occur in it is not properly handled, your operating system can get very slow and unresponsive. And I’ve been using a very simple test to find out how responsive an operating system is when the hard disk is put under stress.
What I do is simple. First I copy a file (which is usually about 1.5 GB) within two locations in the Home folder of the currently logged in user. And as soon as the OS starts to copy it, I immediately try to open a multimedia file. Then as soon as I’m done double clicking on the multimedia file, then I use the start-menu (or Dash if it’s Ubuntu) to search and open a few programs. And while the OS is busy handling all that, I try to navigate to a folder that contains somewhat large number of files through the file manager.
If the OS was able to open most of the programs that I tried to open, if the file manager responds well and all the while if the multimedia playback didn’t get interrupted too much (because after all, all this puts a lot of pressure on the OS), and also, if the mouse pointer didn’t lose its sensitivity too much, then I consider the operating system to a responsive one.
Ubuntu 16.04 LTS:
As mentioned in the review, it didn’t go well at first. In fact it sucked! Other than an app or two (all of which were lightweight ones), until the file copy was fully finished copying, the rest did not get opened nor did the multimedia was played either. But since this has happened to me in the past under Ubuntu, I made a simple tweak to the system, and it completely fixed the situation. Almost all the programs were opened, VLC (I installed it manually using ‘apt-get’ in all three flavors) was able to open the multimedia file quickly and its playback was never interrupted. Mouse pointer also didn’t lose its sensitivity that whole time, although the file manager kind of got stuck for 2-3 seconds, but rest of the system functioned normally. At the end of the day, I loved Ubuntu 16.04 LTS’s responsiveness after the fix.
Note: All these three screenshots are mere illustrations and were not taken at the time of actual testing.
Kubuntu 16.04 LTS:
The KDE Plasma desktop of Kubuntu did extremely well too, if not better! Just like its predecessor (which also did very well), Kubuntu 16.04 LTS didn’t require any manual tweaks either. It rocked out of the box 🙂 .
Ubuntu GNOME 16.04 LTS:
The GNOME flavor had virtually the same issue as of Ubuntu 16.04 LTS so I won’t go into the details. It’s a carbon copy of what happened in Ubuntu 16.04 LTS. So then I applied the tweak I applied to Ubuntu 16.04 LTS, and unsurprisingly, it completely fixed the situation as well.
So all in all, I was very happy with the responsiveness of the three flavors, although to its credit, as mentioned, Kubuntu performed extremely well by default, and it didn’t require any tweaks.
Ubuntu 16.04 LTS and Ubuntu GNOME 16.04 LTS have (almost) the same shutdown delay. They both outperformed Kubuntu 16.04 LTS by being about 244% faster.
The main focus of an Ubuntu LTS is stability and performance. And in that regard, Ubuntu 16.04 LTS did really well. And performance-wise, the memory increase is the major negative element. Other than that, it’s very stable and as mentioned in my original review, many of the subtle issues have also been addressed as well.
Kubuntu 16.04 LTS has a buggy installer. And I consider that to be a huge concern for some end-users. It crashed in my case when it tried to switch the language & keyboard from the default selection. But that doesn’t mean it’ll fail and crash on every language other than US English. Still, it’s a nasty and a very embarrassing bug from the developers’ perspective. They should’ve had tested it more thoroughly. Other than that, as far as the stability of the OS was concerned, it was excellent. Performance-wise, compared to the other two, it was slow to boot and shutdown. But compared to its predecessor, things have been improved nonetheless.
Ubuntu GNOME 16.04 LTS too had a issue. Although it’s wasn’t a major one compared to the Kubuntu installer issue, it delayed the boot-up process by around 15-17 seconds which was somewhat irritating as well. Luckily, it was fixed through a system update. Other than that, it too was very stable.
So that’s it folks. I gave you the numbers concerning performance and shared my experience as far as the stability was concerned. The choice is yours to make. But again, please remember that my hardware is quite compatible with ‘Linux’ to begin with. So that whole ‘quite stable’ statement, well, take it with a grain of salt. However, that’s not to say that it’s completely useless. Because for instance, the fact that I didn’t see a single user application crash (except on Ubuntu GNOME) has nothing to do with my hardware (on some situations they’re responsible for crashes, though). Thus, to a certain degree, it’s still a useful observation.
Before heading towards the download page also make sure to thoroughly read the release notes because they contain information about bugs and known issues that give you valuable insights of what to expect.
Thank you for reading and good luck.