A Five-Way Linux Distribution Comparison In 2010

[sundown]

If you really mean it, I would say it is pretty fanboyish, but I could also assume you were just trying to be funny.

Of course I'm joking.

[bnolsen]

Sorry...32bit isn't very interesting for comparison. Please rerun with modern hardware in mind.

[d4ddi0]

What? no Gentoo benchmarks?

[Cape]

?!
Right... so... same kernel, same userspace programs, same blobby drivers... There's no ***** difference dud'!!! I don't have time to check but for me there's been either a flaw in the benchmarking process or canonical is maintaining phoronix.

Also, don't use arch - it's slow and unstable. ( http://wiki.archlinux.org/index.php/...vertisement.29)

[Shining Arcanine]

Why are you making my life so difficult?!? Now I have to so many of my friends who flaunt Arch Linux to shove it!!! LOL

Seriously though, I have been so sick of this aura of superiority that Arch Linux users seem to develop.

I could produce a chart showing that Microsoft Windows is more than 1 million times faster than any version of Linux, but that does not make the chart true. The benchmark results are clearly flawed; Arch Linux users themselves are pointing out that their distribution has no stock configuration. That is similar to Gentoo Linux, although Gentoo Linux is source based, which makes controlling how things are compiled and linked much easier than Arch Linux. It is ironic that some Arch Linux users are saying that their system needs to be optimized when it is far less customizable than Gentoo Linux. I still find the idea that the Arch Linux system was not probably optimized a valid point, because there is no reason for having a mainstream distribution should do so poorly in benchmarks. Perhaps a generic i386 kernel is being used.

By the way, Phoronix has made a mockery of proper benchmarking by paying no attention to whether or not the results are valid. When something is as wildly off as the Arch Linux benchmarks were, benchmarkers need to investigate to see whether the benchmarks as they were tainted by the introduction of some assumption/mistake that was wrong by rerunning them, analyzing them to ascertain the cause and running them again to verify that hypothesis. This is something that Phoronix clearly failed to do and this sort of behavior is unethical.

[Michael]

?!
Right... so... same kernel, same userspace programs, same blobby drivers... There's no ***** difference dud'!!!

The kernels aren't the same nor is all of the userland the same. There are also no blob drivers involved.

[Michael]

benchmarkers need to investigate

Anyone is welcome to reproduce the results.

[Michael]

Sorry...32bit isn't very interesting for comparison. Please rerun with modern hardware in mind.

That's what's usually done and then there are the people complaining that I ignore those running older hardware, so once in a while it's switched up.

[monraaf]

Ouch!

[Cape]

The kernels aren't the same nor is all of the userland the same. There are also no blob drivers involved.

Unfortunately you are not specifying what hardware/drivers you are using for arch, so i'm assuming you are using Radeon Drivers (blob) as in the others benchmarks.

Userspace Programs should be the same (or as similar as possible) since you are running a distro benchmark, not a multi distro debugging session.

Arch uses kernel 2.6.33 (no special config involved) while ubuntu is using 2.6.32, since arch is not using special configs - actually it should be a little slicker than ubuntu kernel - i don't understand the enormous lag you have showed.