Page 1 of 4 123 ... LastLast
Results 1 to 10 of 37

Thread: Forking a "Linux Desktop" kernel

  1. #1
    Join Date
    Apr 2007
    Location
    Portland, Oregon
    Beans
    348
    Distro
    Ubuntu 9.10 Karmic Koala

    Forking a "Linux Desktop" kernel

    I just read this really interesting article that asks many of the questions I have been asking myself lately after attending the UbuntuLive conference.

    The basic premise is that maybe there SHOULD be a Desktop Linux kernel. Here's a few snippets to look over to see if you think the article is interesting enough to read:

    "Linux kernel is too bloated

    It was made for servers because that's where its initial success was possible and where most of the money came. ... Developers are also developing on machines which are too powerful compared to the machines that ordinary desktop users are using,"

    "Lack of communication with desktop users

    Linux kernel developers are normatively disconnected from the rest of the users, most of which are desktop users, the kind making up 90% of all computer users. ... a lot of the time [the developers will] be oblivious to the needs of the ordinary desktop users."

    "The point that CK makes is related to performance issues on hardware which is so powerful that it shouldn't have performance issues at all. Why do we still need a specially optimized Linux kernel just to properly enable music production? Shouldn't real time audio processing just work absolutely flawlessly on computers containing dual core CPUs? Even Windows XP doesn't have this requirement."

    "One of the solutions being pondered is forking a Linux kernel into a kernel which would specifically be aimed at desktop users and their needs."

    So what do you all think? I think there is DEFINITELY a need for a Linux Desktop kernel. Ubuntu could champion this easily and include it on the LiveCD's as a "checkbox" to select between the standard and Desktop kernel during installation.

  2. #2
    Join Date
    Jun 2006
    Location
    Syracuse, NY
    Beans
    3,264

    Re: Forking a "Linux Desktop" kernel

    Are there enough programmers? I would LOVE to help, but I lack the skills (as of now). I think the real solution is to get a panel to oversee benchmarking on the kernel in ways meaningful to desktop users. Have representatives from all the major desktop distros there and have them run some tests and put pressure on kernel developers in our favor. We don't need to divide our strength, we need our voice to be heard. We need a Union if you will, and not one of those POS greedy organizations that are America's unions now, but an actual group that will stand up and give us a voice.
    Desktop: AMD Athlon64 X2 3600+, Nvidia 8600GT, 3GB RAM, 80GB hd, Windows 7 Beta
    Lappy: Sony Vaio FW-140E, Intel P8400 2.26Ghz, 3GB Ram, 250GB HD, Intel x4500MHD, Windows 7 Beta & Kubuntu 8.10 w/ KDE 4.2

  3. #3
    Join Date
    May 2007
    Beans
    946

    Re: Forking a "Linux Desktop" kernel

    Quote Originally Posted by psyopper View Post
    I just read this really interesting article that asks many of the questions I have been asking myself lately after attending the UbuntuLive conference.

    The basic premise is that maybe there SHOULD be a Desktop Linux kernel. Here's a few snippets to look over to see if you think the article is interesting enough to read:

    "Linux kernel is too bloated

    It was made for servers because that's where its initial success was possible and where most of the money came. ... Developers are also developing on machines which are too powerful compared to the machines that ordinary desktop users are using,"

    "Lack of communication with desktop users

    Linux kernel developers are normatively disconnected from the rest of the users, most of which are desktop users, the kind making up 90% of all computer users. ... a lot of the time [the developers will] be oblivious to the needs of the ordinary desktop users."

    "The point that CK makes is related to performance issues on hardware which is so powerful that it shouldn't have performance issues at all. Why do we still need a specially optimized Linux kernel just to properly enable music production? Shouldn't real time audio processing just work absolutely flawlessly on computers containing dual core CPUs? Even Windows XP doesn't have this requirement."

    "One of the solutions being pondered is forking a Linux kernel into a kernel which would specifically be aimed at desktop users and their needs."

    So what do you all think? I think there is DEFINITELY a need for a Linux Desktop kernel. Ubuntu could champion this easily and include it on the LiveCD's as a "checkbox" to select between the standard and Desktop kernel during installation.
    Desktop Kernel? Sounds like a kernel designed for GUI use. Wouldn't that eliminate native CLI use? That's a bad idea.

  4. #4
    Join Date
    Jan 2007
    Location
    California
    Beans
    350
    Distro
    Ubuntu 12.10 Quantal Quetzal

    Re: Forking a "Linux Desktop" kernel

    This might just be me, but I fail to see how bloat on a server is any better than bloat on a desktop.

    I can see how developing on a machine where you don't feel the bloat can lead to an OS too heavy for the mainstream, but certainly there are any number of packages/processes in the kernel that could be improved by making them lighter/more efficient improving performance on both desktop and server alike.

    IMHO, more power should not be an excuse for heavier software and efficiency should always be a top goal in software design at all levels, and for all environment.

    It might never happen again in the desktop market, but still.
    Euler_fan
    Ubuntu User # 15369 at the Ubuntu Counter Project
    Woot for GPG

  5. #5
    Join Date
    Jun 2006
    Location
    Syracuse, NY
    Beans
    3,264

    Re: Forking a "Linux Desktop" kernel

    Quote Originally Posted by init1 View Post
    Desktop Kernel? Sounds like a kernel designed for GUI use. Wouldn't that eliminate native CLI use? That's a bad idea.
    Nonono, the issue is that the linux kernel comes optimized for servers. Many of the changes deal with gaining an extra 0.1% on database benchmarks (as most kernel developers are paid by companies that make their money on servers... they want the best server OS possible). Believe it or not, lots of these changes to help servers actually hurt desktop users. The types of tasks the kernels do is different enough that something that speeds up database queries hurts responsiveness on the desktop. Right now there is no voice to the kernel devs that this is happening. How can an average desktop user say that since the recent kernel update tabs on firefox open slower (I took this example from the con article).

    This has nothing to do with GUIs or programs installed, but rather how the kernel handles typical desktop tasks compared to typical server tasks.
    Desktop: AMD Athlon64 X2 3600+, Nvidia 8600GT, 3GB RAM, 80GB hd, Windows 7 Beta
    Lappy: Sony Vaio FW-140E, Intel P8400 2.26Ghz, 3GB Ram, 250GB HD, Intel x4500MHD, Windows 7 Beta & Kubuntu 8.10 w/ KDE 4.2

  6. #6
    Join Date
    Mar 2007
    Location
    Beginners Team IRCChannel
    Beans
    2,811
    Distro
    Kubuntu 10.10 Maverick Meerkat

    Re: Forking a "Linux Desktop" kernel

    Quote Originally Posted by igknighted View Post
    Nonono, the issue is that the linux kernel comes optimized for servers. Many of the changes deal with gaining an extra 0.1% on database benchmarks (as most kernel developers are paid by companies that make their money on servers... they want the best server OS possible). Believe it or not, lots of these changes to help servers actually hurt desktop users. The types of tasks the kernels do is different enough that something that speeds up database queries hurts responsiveness on the desktop. Right now there is no voice to the kernel devs that this is happening. How can an average desktop user say that since the recent kernel update tabs on firefox open slower (I took this example from the con article).

    This has nothing to do with GUIs or programs installed, but rather how the kernel handles typical desktop tasks compared to typical server tasks.
    Agreed (based on my limited knowledge of kernels over the years).

    I dunno if forking is the end all solution, might cause too much trouble. We do need to bring more attention to the desktop in order to get overall performance and experience improved. We can't just rely on folks like ck's patches to do it for us. Maybe Shuttleworth should start an initiative based on this and talk with Linus about setting up some sort of streamlined way of getting Kernel problems to devs (average desktop users certainly don't want mailing lists... I think that's been established).

  7. #7
    Join Date
    May 2007
    Beans
    946

    Re: Forking a "Linux Desktop" kernel

    Quote Originally Posted by igknighted View Post
    Nonono, the issue is that the linux kernel comes optimized for servers. Many of the changes deal with gaining an extra 0.1% on database benchmarks (as most kernel developers are paid by companies that make their money on servers... they want the best server OS possible). Believe it or not, lots of these changes to help servers actually hurt desktop users. The types of tasks the kernels do is different enough that something that speeds up database queries hurts responsiveness on the desktop. Right now there is no voice to the kernel devs that this is happening. How can an average desktop user say that since the recent kernel update tabs on firefox open slower (I took this example from the con article).

    This has nothing to do with GUIs or programs installed, but rather how the kernel handles typical desktop tasks compared to typical server tasks.
    OK, that's better. I didn't understand, but I do now.

  8. #8
    Join Date
    Jul 2005
    Beans
    193

    Re: Forking a "Linux Desktop" kernel

    YES! The inherent problem of Linux is how slow it is on the desktop! I've been saying this all a long...

  9. #9
    Join Date
    Jun 2006
    Location
    Syracuse, NY
    Beans
    3,264

    Re: Forking a "Linux Desktop" kernel

    Quote Originally Posted by PrimoTurbo View Post
    YES! The inherent problem of Linux is how slow it is on the desktop! I've been saying this all a long...
    Honestly, I would be willing to bet that if there is ANY noticeable slowdown from XP it is a driver-related issue. Most of the kernel related stuff is more "why isn't this blowing the bloated winXP out of the water" level things. It's hard to say for sure though. So much hardware is supported amazingly out of the box (WAY more than windows), but if it isn't the chances of finding good 3rd party drivers aren't good. This could change due to an API change forthcoming in the 2.6.23 kernel, so hang in there.
    Desktop: AMD Athlon64 X2 3600+, Nvidia 8600GT, 3GB RAM, 80GB hd, Windows 7 Beta
    Lappy: Sony Vaio FW-140E, Intel P8400 2.26Ghz, 3GB Ram, 250GB HD, Intel x4500MHD, Windows 7 Beta & Kubuntu 8.10 w/ KDE 4.2

  10. #10
    Join Date
    Jan 2007
    Location
    the third world
    Beans
    Hidden!

    Re: Forking a "Linux Desktop" kernel

    Ah. Like always there is a lot of managerial 'we should do this'/'vision for the Linux desktop' nonsense.

    I doubt anybody is going to hack on a "Desktop kernel" (whatever that is), though.

    A good, secure server OS makes for a good desktop OS too. I don't see why anyone would want to create a fork for desktop use.
    IESVS FELLAT IN INFERNVM

Page 1 of 4 123 ... LastLast

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •