What Linux Distribution Should Be Benchmarked The Most?

[dh04000]

That's just childish. Which rules are you talking about?

The made up ones in his head, lol.

[dh04000]

Those numbers are worthless and anyone who uses them as a basis for any sort of argument is either ignorant, or intentionally trying to mislead (ie. a liar).

Internet traffic sites, like netapplications and the like, only measure page hits from the pages of their few partner companies who participate in the program. That already gives a very biased sampling, concentrating too much on businesses and too little to home users (and ignoring countries outside Europe/USA pretty much completely). But to top that off, their methodology is flawed, with things like multiple users behind a NAT counted as a single user, and inaccuracies in detecting which OS is used.

Counting methods that identify by browser user agent are extra worthless, as many versions of Mint have the user agent set to report Ubuntu by default - thus resulting in many Mint installations being counted as Ubuntu.

Fact is that there is no reliable way to measure distro popularity. We can get ballpark figures, but nothing accurate.

So basically, any real number generated that doesn't agree with the "feeling/conclusion" you have, is misleading and I'm a liar for suggesting to use a real benchmark instead of going with your gut?

Lol, ok. We got ourselves a real badass over here guys.

[chrisb]

For the trolls: Be carefull what u wish for, if ubuntu would to fall it wont fall alone...

Newsflash: it's ubuntu that builds on GNU/Linux, not the other way around.

You're right, there is absolutely nothing that builds on Ubuntu... apart from Edubuntu, Ubuntu GNOME, Kubuntu, UbuntuKylin, Lubuntu, Mythbuntu, Ubuntu Studio, Xubuntu, BackBox, Baltix, Fluxbuntu, FrogLinux, Guadalinex, Itis Linux, ChristianUbuntu, Ubuntu Mini, Sabily, Ubuntulite, UbuntuRescueRemix, Dubuntu, liUbuntu, Tilix, Runtu, gNewSense, Ichthux, Linux Secure, Mint, Pioneer, MMourcebuntu, TeXbuntu, Vinux, BlankOn, Elbuntu, Zorin OS, Gnoppix, gOS, ImpiLinux, Kiwi_Linux, nUbuntu, PUD Gnu/Linux, Pyramid Linux, TheOpenCD, Ulteo, Bardinux, FreezyLinux, OzOS, moonOS, Ubuntu Eee, SuperX, Satux, BigLinux, LliureX, OpenTLE, Polippix, BoliviaOS, Bubuntu, iMagic OS, Securpc, Zevenos, and Estobuntu.

[verde]

So basically, any real number generated that doesn't agree with the "feeling/conclusion" you have, is misleading and I'm a liar for suggesting to use a real benchmark instead of going with your gut?

Lol, ok. We got ourselves a real badass over here guys.

Steam usage as far as Debian regards is MASSIVELY in favor of Ubuntu. Doubting and disgusing the Ubuntu general "market-share" supremacy by some people here, its clear indicator that they are haters in denial!

[chrisb]

I'm sorry to inform you that LINUS TORVALDS is PROUD of Android and he is using CHROMEBOOK!

Yes, he likes Android, but he's running Fedora on that Chromebook.

[uid313]

Easiest in he short-term would just go with a non-Mir Ubuntu-derivative such as Kubuntu or Xubuntu.

[mrugiero]

LFS (Linux from scratch)

That's ridiculous.

You're right, there is absolutely nothing that builds on Ubuntu... apart from Edubuntu, Ubuntu GNOME, Kubuntu, UbuntuKylin, Lubuntu, Mythbuntu, Ubuntu Studio, Xubuntu, BackBox, Baltix, Fluxbuntu, FrogLinux, Guadalinex, Itis Linux, ChristianUbuntu, Ubuntu Mini, Sabily, Ubuntulite, UbuntuRescueRemix, Dubuntu, liUbuntu, Tilix, Runtu, gNewSense, Ichthux, Linux Secure, Mint, Pioneer, MMourcebuntu, TeXbuntu, Vinux, BlankOn, Elbuntu, Zorin OS, Gnoppix, gOS, ImpiLinux, Kiwi_Linux, nUbuntu, PUD Gnu/Linux, Pyramid Linux, TheOpenCD, Ulteo, Bardinux, FreezyLinux, OzOS, moonOS, Ubuntu Eee, SuperX, Satux, BigLinux, LliureX, OpenTLE, Polippix, BoliviaOS, Bubuntu, iMagic OS, Securpc, Zevenos, and Estobuntu.

You can name even a bigger list of distros, and those distros will still not be a game changer. They can rebase on anything else. And that's pretty much trivial.

Arch linux or Gentoo ~unstable would be more appropriate since they have the latest packages. Ubuntu has taken a path with mir and its plain rotten.

That's too much work for just a benchmark. Remember he will need to change the config too often.

Steam usage as far as Debian regards is MASSIVELY in favor of Ubuntu. Doubting and disgusing the Ubuntu general "market-share" supremacy by some people here, its clear indicator that they are haters in denial!

Steam only indicates who wins between gamers. Also, Ubuntu has got the biggest group, but it's not more than the others summed up (which means as far as display system goes, Mir will probably be used by less people than Wayland, and this is likely to be a crucial part on performance).

I think the ideal is to test all the significantly different stacks, i.e., Ubuntu (just Mir justifies it to be benchmarked on its own, but there's also upstart that will be relevant when benchmarking startup time), any Wayland based distro, any X.org based one, and, when testing startup times, one with systemd.
I guess Lubuntu will be the only relatively massive one relying on X, and then Xubuntu or Kubuntu (I'm not saying Fedora because of the debug build, which might lead to misinformation about the performance of this kind of stack) are likely to use Wayland eventually.
As for testing specific features from development packages, Ubuntu for Mir depending ones, Xubuntu or Kubuntu for everything else.

[souenzzo]

That's too much work for just a benchmark. Remember he will need to change the config too often.

No, arch is simpler than most other. Just "pacman -S catalyst-dkms" for updated catalyst driver, and just "makepkg && pacman -U catalyst-legacy" for an older driver.

[Vim_User]

I am not a fan of Ubuntu/Canonical, but this is just ridiculous. How do you people think that using Ubuntu instead of Fedora/openSuse/Arch/Slackware/... will affect the results of those benchmarks? The only thing that comes to my mind (and where Ubuntu had problems in the past) is the compositor/WM that can influence the gaming benchmarks, a thing that can easily be fixed by using a version with a non-compositing WM, for example Lubuntu. But if Michael makes a comparison of kernel versions, graphic drivers or anything else the OS that is used shouldn't matter at all, otherwise the benchmarks setup is flawed in itself.
Also, since when is systemd the standard that must be used for benchmarks and why do you people expect that it would give you other benchmark results than an Upstart/OpenRC/SystemV distro?

Again, I am not a fan of Ubuntu/Canonical, but this is just pure hatred for no good reason.

[felipe]

The most important rule.

You are raising an interesting point. Where does a "genuine" Linux distribution start and where does it end?

Is android-x86 a genuine Linux Distribution? After all it ships Linux and has a "desktop" environment.
Is openelec a genuine Linux Distribution? Altough it does not have a real desktop.
Is Rebecca Black a genuine Linux Distribution? It's the only one shipping Wayland.
Is OpenWRT a genuine Linux Distribution? It does not have a desktop at all.
Are all these rescue systems out there "genuine" Linux distributions?

So what makes a "genuine" Linux Distribution? The desktop environment? The display server?

Obviously the kernel is one hard criteria. But the rest?




That's just childish. Which rules are you talking about?

The principle rule of the free software. The rule more important,the rule who gave a birth this kind of operative systems. Of Course i'm talking about the ethics. Canonical doesn't have ethics. They Include adware in Ubuntu. And they stole the work from another projects and sold as they own