Announcement

Collapse
No announcement yet.

Linux 5.0 Kernel Performance Is Sliding In The Wrong Direction

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #21
    Originally posted by DoMiNeLa10 View Post
    While I'd love to say that the ship is sinking, the golden era of Linux is over, that Torvalds killed the project by adding the CoC, and that we should all move to DragonFlyBSD's kernel, I think it's too early to say. Right now it doesn't really look like post-CoC kernels have a trend to degrade performance, and there's a possibility that this build of 5.0 was misconfigured, or that some bad commit went through and wasn't reverted yet. This isn't a proper 5.0 release, so the best thing to do is to wait and benchmark once more later.
    Just please, don't blame Linus for that,
    That decision was forced on him..

    A lot of things that before were more rigid, now needs to be accepted, like if the Linux Kernel was a "Social Project",
    And so any one can tweak its part of it so be a happy person in life...

    But Linux Kernel is not that, its a big Piece of Software, were if you have soft rules, it will derail..

    Blame the Social Justice Warriors( SJW ), ( a group of people, that for who have learned history of war...they should be called "5ft Collumn"... )

    Comment


    • #22
      Originally posted by moilami View Post
      This is because Linux CoC
      Very true. I have a feeling that tells me open-source software quality has been going down after the additions of "Code of Conduct"s in projects...

      Comment


      • #23
        Originally posted by andyprough View Post

        Yup. The problem could be on Ubuntu's side.
        Nah, the problem is on the kernel side 'cause Linus has become Mr Nice Guy

        Comment


        • #24
          Originally posted by tildearrow View Post

          Very true. I have a feeling that tells me open-source software quality has been going down after the additions of "Code of Conduct"s in projects...
          Developers can't be afraid to call out crap code due to fear of reprisal and I think CoCs are sometimes used to undermine that. It's important to have honest peer review in order to maintain clean code to be completely honest. You can do this without being an ass though, but people also need to not mistake criticism for being an ass because they're not the same thing.
          Last edited by jrdoane; 23 February 2019, 02:21 PM.

          Comment


          • #25
            Come on, people. Blaming the CoC for this is silly. Join any kernel development-related email list and I assure you that developers are not afraid to call out crap code. You can call out crap code based on objective measures without turning it into a personal attack. The CoC addresses the latter. Linux is sufficiently big and complex at this point and software build systems have become sufficiently sophisticated that it's probably time to compile a lot more custom kernels based on what they'll be used for. For example, for my structural biology people who run mostly python/C/CUDA code, I absolutely plan to remove all the sceptre/meltdown crap from their workstation kernels; a totally unnecessary performance hit, given the context. The optimal scheduler for desktop systems is different from the optimal scheduler for a server, and so on. The gentoo people are on the right track. The next step is to make it a lot easier to customize kernels for particular uses; the exact opposite of the Ubuntu/RHEL model.

            Comment


            • #26
              Originally posted by tildearrow View Post

              Very true. I have a feeling that tells me open-source software quality has been going down after the additions of "Code of Conduct"s in projects...
              Seriously? We're going to argue that saying you can't mistreat other people or use slurs leads to poor code? This is why Linux users are portrayed as Neanderthals with beards in the popular media; we're not doing anything to dispel the image.

              Wojcian posted a benchmark showing it's a Ubuntu issue rather than a kernel issue, everyone ignores it and starts going on about the evils of being asked not to sexually harass women? Sheesh.

              Comment


              • #27
                Originally posted by alcalde View Post
                Wojcian posted a benchmark showing it's a Ubuntu issue rather than a kernel issue, everyone ignores it
                You are right. It looks from the benchmark Wojcian posted that the issue is probably on the Ubuntu side.

                Comment


                • #28
                  Now that it's been proven to be wrong, I'm waiting for all the people in these thread who falsely accused the kernel developers' adoption of the CoC to be the cause of this "regression" to apologize. Of course you won't, will you? That would require being a real grown up.

                  Comment


                  • #29
                    They are script kiddies at best, or second-rate trolls. It's a shame they pollute the comments. At least such comments may have potential for ML detection of morons. . Phoronix articles are first rate but for intelligent comments it's lwn.

                    Comment


                    • #30
                      Originally posted by hreindl View Post

                      it was a good job only bechmark Ubuntu and making a topic "buuuuh Linux 5.0 will be slow"?
                      this whole topic is one of the reasons i don't give a shit about generic benchmarks at all
                      To be fair I'll also test 4.18 kernel with my custom config to see how it performs against 5.0. What I don't like Ubuntu has debugging enabled even for final releases. I'm not sure how it is in other distributions. FreeBSD has this disabled (and this is good!) as far as I remember. Probably the same in Windows, OS X, so it puts Ubuntu and maybe few other distributions in disadvantage for no reason.

                      Comment

                      Working...
                      X