Announcement

Collapse
No announcement yet.

More Benchmarks From Linux 5.5 Looking Like A Scheduler Snafu Even On Smaller CPUs

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

  • More Benchmarks From Linux 5.5 Looking Like A Scheduler Snafu Even On Smaller CPUs

    Phoronix: More Benchmarks From Linux 5.5 Looking Like A Scheduler Snafu Even On Smaller CPUs

    For the Linux 5.5 kernel that's about half-way through its development phase we have been pointing out some rather significant performance regressions affecting both AMD EPYC and Intel Xeon servers but there are also regressions to be found with desktop class systems too...

    http://www.phoronix.com/scan.php?pag...sed-On-More-HW

  • #2
    If load_balance() in CFS is to blame, I wonder how it compares to BFS (edit: uh, I mean MuQSS) and BMQ.

    Comment


    • #3
      Originally posted by halo9en View Post
      If load_balance() in CFS is to blame, I wonder how it compares to BFS (edit: uh, I mean MuQSS) and BMQ.
      I agree! I would be amazing to compare all them! What about 5.6 and Git branches too?

      Comment


      • #4
        Well--it appears as though they may not have gotten rid of Linus Torvalds quite soon enough. Too many things still creaking along just fine for right now, don't'cha know...


        "Linux - The beginning of the end"
        Updated: September 21, 2018

        https://www.dedoimedo.com/computers/...f-the-end.html


        Comment


        • #5
          Originally posted by danmcgrew View Post
          Well--it appears as though they may not have gotten rid of Linus Torvalds quite soon enough.
          He isn't a moron like Stallman so he does not post dumb opinions on highly volatile matters. This complicates the SJW job quite a bit and it may very well save him.

          #StallmanDidNothingWrong
          #EpsteinDidn'tKillHimself

          Comment


          • #6
            I hope they get to the bottom of this set of regressions - touching this part of the kernel can be all sorts of crazy! if we get to RC5/6 and there are no fixes in sight, then I am sure these scheduler changes will be reverted and worked on.

            not sure why there is all this doom and gloom going on...

            Comment


            • #7
              Originally posted by boxie View Post
              I hope they get to the bottom of this set of regressions - touching this part of the kernel can be all sorts of crazy! if we get to RC5/6 and there are no fixes in sight, then I am sure these scheduler changes will be reverted and worked on.

              not sure why there is all this doom and gloom going on...
              Perhaps being informed--as opposed to being simply a 'user'--would help immensely. Start here (and don't stop)--

              "Linux - The beginning of the end"
              Updated: September 21, 2018
              https://www.dedoimedo.com/computers/...f-the-end.html --



              Comment


              • #8
                Originally posted by danmcgrew View Post

                Perhaps being informed--as opposed to being simply a 'user'--would help immensely. Start here (and don't stop)--

                "Linux - The beginning of the end"
                Updated: September 21, 2018
                https://www.dedoimedo.com/computers/...f-the-end.html --


                the problem is, I am informed. I see the common themes from kernel cycle to cycle. stuff gets tried, stuff works or it doesn't. if it doesn't it gets pulled back and fixed and tried again.

                Comment


                • #9
                  Michael again it seems more interesting, helpful and relevant to point to a commit instead of showing multiple applications affected

                  Comment


                  • #10
                    Originally posted by brauliobo View Post
                    Michael again it seems more interesting, helpful and relevant to point to a commit instead of showing multiple applications affected
                    I will in an article or two, first needing to cover my expenses via ads for the time invested in doing so.
                    Michael Larabel
                    http://www.michaellarabel.com/

                    Comment

                    Working...
                    X