Testing The Different Ubuntu 10.04 Kernels

[phoronix]

Phoronix: Testing The Different Ubuntu 10.04 Kernels

The release of Ubuntu 10.04 LTS "Lucid Lynx" is quickly approaching next month and it will arrive with a whole set of new features and improvements including a faster boot process, a long-awaited new theme, the Nouveau driver to replace the crippled xf86-video-nv driver, the unveiling of the Ubuntu One Music Store, integration of Plymouth, Ubuntu ARM advancements, and many other advancements for this Linux distribution. While it may not be as exciting as looking at these new end-user features, in this article we are testing out the available kernels for Ubuntu 10.04. Besides the standard Linux 2.6.32 kernel used in the Lucid release, there is also a specialized server kernel as well as a new -preempt kernel is now available. We are looking at how these different kernels perform and how they compare to the mainline Linux kernels with the 2.6.32, 2.6.33, and 2.6.34-rc1 releases.

http://www.phoronix.com/vr.php?view=14663

[Shining Arcanine]

Why are there so many Ubuntu Linux reviews? Why not review other Linux distributions, like Gentoo Linux or Arch Linux? Ubuntu seems to be getting preferential treatment in that not only is it featured in every single review, but multiple reviews are done of its alpha releases. Seeing reviews of Ubuntu when other perfectly viable distributions are being completely ignored is getting old fast, especially since it tends to show performance ceilings local to Ubuntu and not global to the Linux community as a whole.

[Michael]

Why are there so many Ubuntu Linux reviews? Why not review other Linux distributions, like Gentoo Linux or Arch Linux? Ubuntu seems to be getting preferential treatment in that not only is it featured in every single review, but multiple reviews are done of its alpha releases. Seeing reviews of Ubuntu when other perfectly viable distributions are being completely ignored is getting old fast, especially since it tends to show performance ceilings local to Ubuntu and not global to the Linux community as a whole.

With this article for instance, it came out of a direct request from Canonical. Other vendors are welcome to come forward and communicate requests that then turn into new tests, but most often it's from Canonical. And the fact that Ubuntu is effectively the most popular desktop Linux distribution it caters to the largest audience.

[Shining Arcanine]

With this article for instance, it came out of a direct request from Canonical. Other vendors are welcome to come forward and communicate requests that then turn into new tests, but most often it's from Canonical. And the fact that Ubuntu is effectively the most popular desktop Linux distribution it caters to the largest audience.

I was tempted to say that Canonical was probably involved with this in my previous post, but I decided against it because I felt that such a claim was unsubstantiated. I guess I was right in thinking that the sheer quantity of reviews was too great to be a coincidence.

Anyway, the flood of Ubuntu reviews seems to be making Phoronix less of an independent review site and more of an arm of Canonical's marketing department. As much as I appreciate what Canonical is doing for open source, I think Ubuntu should not dominate Phoronix's reviews to the degree it has and that it would be more beneficial to open source software in general if performance across different distributions were benchmarked to identify global minima and maxima that affect Linux in general, rather than just the local minima and maxima that affect Ubuntu. It would probably be beneficial to Ubuntu users too, because then Phoronix would be providing numbers that show where Ubuntu really ranks in the Linux world, which could result in performance improvements in Ubuntu after Canonical's developers see the results.

[Michael]

BTW, in the queue for Tuesday (I think that's the date) is a comparison of the latest development builds of openSUSE, Mandriva, PCLinuxOS, and Ubuntu.

[Shining Arcanine]

BTW, in the queue for Tuesday (I think that's the date) is a comparison of the latest development builds of openSUSE, Mandriva, PCLinuxOS, and Ubuntu.

That is good to hear. Is there any chance that Arch Linux and Gentoo Linux could be added to the mix in the future? The approaches both of those distributions take are different enough from that of most mainstream distributions that adding them to the mix might identify performance regressions that would not otherwise appear and could make viewing the benchmark results more interesting.

[unix_epoch]

Request for future tests that include preempt and low-latency kernels: please add some kind of latency/responsiveness test, like cyclictest or interbench. Even posting more detailed stats on the game framerates would be very helpful (i.e. min, max, median, standard deviation, etc.).

[Shining Arcanine]

Request for future tests that include preempt and low-latency kernels: please add some kind of latency/responsiveness test, like cyclictest or interbench. Even posting more detailed stats on the game framerates would be very helpful (i.e. min, max, median, standard deviation, etc.).

More detailed statistics would be nice. Perhaps Phoronix could add box plots to the graphs.

[Yfrwlf]

I don't mean to take sides for or against, as I'm not trying to, but it has been pointed out many times here that there's really no end to the details and numbers of tests you could dive into. Their goal is to give general real-world Linux user tests, not so much hacker-oriented tests. The goal is to keep Phoronix more relevant to "average" Linux users no doubt, though it still certainly is geek-slanted as you could easily say any hardware review and software benchmarking site would be.

Phoronix does review other Linux distros, and I fully agree that it is very important to take a step back and look at other Linux configurations and question performance differences with them to get a good perspective on things. Phoronix does, as was mentioned, have articles that span several distros as well as ones focused on certain ones. Out of those that are focused, a higher percentage are Ubuntu-focused, and as stated that is because Ubuntu is the most popular and thus more highly used distro and more relevant to a greater number of Linux users.

So all-in-all, I think they do a pretty good job of focusing on what is of interest to their widest audience. Could they have a "hardcore hacker" section of the site and have appropriate articles for that section? Perhaps, but unless they separated it out in an elegant way, it may turn off some of their more general audience readers. For those wanting to dive into the code and the exact reasons why performance changes exist, it would be cool though, but for everyone else?...not so much probably.

[FireBurn]

That is good to hear. Is there any chance that Arch Linux and Gentoo Linux could be added to the mix in the future? The approaches both of those distributions take are different enough from that of most mainstream distributions that adding them to the mix might identify performance regressions that would not otherwise appear and could make viewing the benchmark results more interesting.

It's unlikely Gentoo will ever be properly benchmarked by Michael, Gentoo takes a lot longer to set up that a binary distro and compilation on an Atom CPU would take a while, reinstalling from scratch to test something else would also take time.

Compared to the time to install Ubuntu or Fedora against their user bases it probably doesn't seem viable even if the results have the potential to be more interesting

Oh and for the books I am a Gentoo user, I won't comment on Arch as I've never used it