2017: Should Linux Benchmarking Still Be Mostly Done With Ubuntu?

[Guest]

Ubuntu, because it's the most popular distribution. However, when comes to 3D benchmarks it will be great to see Kubuntu and Fedora as well. It will show differences between Unity and Mir against KDE/Gnome and Wayland.

[FishPls]

Ubuntu, because it's the most popular distribution. However, when comes to 3D benchmarks it will be great to see Kubuntu and Fedora as well. It will show differences between Unity and Mir against KDE/Gnome and Wayland.

Kubuntu is apparently a pretty bad KDE distro because they ship outdated Qt versions usually.

I personally would like to see Fedora / openSUSE / Arch if Ubuntu is no longer the main distro. Doesn't really matter which one.

[halo9en]

Arch.

[AdamOne]

Personally haven't used Ubuntu in years for my desktop OS. But it is still the most popular distro in one form or another.

I personally use Antergos and I can't fault it. Although I want to build my own Arch at some point.

the thing with distros is that with time one's disposition seems to change with it, some software engineer might might find himself installing an 'buntu somewhere just out of convenience.
there is no distro that suits every situation. when youre done setting arch up and installing a desk env, there are other tools you're accustomed to like automounting devices that arent there and wont work until you take the energy to learn it. its not always fun. if its not a good time in your life then avoid it.

but it is a little sweet.

gentoo seems interesting but man... you prolly gotta have some really strong weed to invest in that

[TumultuousUnicorn]

I'm an Arch Linux user, and I've a lot of respect for Fedora and Gentoo.

Should Linux Benchmarking Still Be Mostly Done With Ubuntu? I want to answer you Michael one thing: you are the captain of the boat, and you can choose which boat you want, so if you prefer Fedora, let's benchmark with Fedora!
Fedora is easy to install, e.g. user-friendly (contrary to Arch Linux or Gentoo, I understand, and I'm not interested by forks), and on bleeding edge (~6 months old packages on Ubuntu *cough*). It uses modern software used by a lot of GNU/Linux distro (like GNOME, Systemd and others), and pulls forward the GNU/Linux world (remember who was the first on Wayland, GLVND and others).

As you said in this new, Ubuntu wants to use Unity, Ubuntu wants to use Upstart, Ubuntu wants use Mir... That is not what a majority of others distro use; Ubuntu is *popular*, but not the ambassador of GNU/Linux world.

[atmartens]

Is the point of the benchmarks to show the relative change in performance over time of a given software program, or to compare the relative performance of different distributions? If it's the former, then pick a single distribution and stick with it. If it's the latter, then you need to compare multiple distributions.

[ernstp]

What about using Ubuntu Gnome? That's the same desktop as Fedora and Debian have by default (right?) and I guess it's common with Arch users also etc.
But still the distro that Valve and most AAA games support for maximum compatibility.

[Ray54]

Although I rarely use Ubuntu, I think it is definitely has to be the standard distribution for benchmarking. Partly, as it is the distribution that most non-commercial and commercial developers will have developed and tested against, and partly as it is relatively easy to translate results to expected performance on other distributions. In some ways the recent lack of changes that may make Ubuntu a slightly boring desktop to some, makes it a better standard for benchmarking. In the longer term, standard Ubuntu may not be the right tool, as MIr, etc, may make it unrepresentative of the wider Linux base. However, for now I would like to keep Ubuntu as the standard benchmark, but with limited regular tests on other popular distributions to allow for easy and predictable performance translation. For many, I think it is Linux vs. Windows 10 or Mac OS that is the most important benchmark.

[DanL]

As many of our cross-distribution performance comparisons have shown, when at similar package versions, Antergos, openSUSE, Debian, and others generally don't perform vastly different from Ubuntu.

Michael, as long as that holds true, you should continue to use Ubuntu (or Xubuntu) if that's what you find to be easiest, quickest and/or most reproducible. Ubuntu seems to be a bit conservative introducing new versions of major components (kernel, X, mesa), but you seem to have no issue working around that and dropping in newer components when relevant to the benchmark. The bottom line is that you can't please all of the people all of the time.

[duby229]

Just my opinion, but yeah of course you should continue benching with Ubuntu for all the obvious reasons. But I think Fedora is a more representative linux distribution for a larger majority for all the obvious reasons. So I would like to see more benches with it too by default.

Oh yeah, I like to see Gentoo benches too! (That's a hint for a person interested to contribute it!)