Announcement

Collapse
No announcement yet.

Linux 5.5-rc3 Benchmarks Are Still Pointing To Slips In Performance

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

  • Linux 5.5-rc3 Benchmarks Are Still Pointing To Slips In Performance

    Phoronix: Linux 5.5-rc3 Benchmarks Are Still Pointing To Slips In Performance

    Early on in the Linux 5.5 cycle during the merge window we saw some wild swings in performance including some positive gains but also performance regressions. Given last weekend's Linux 5.5-rc3 release having merged some scheduler fixes and other fallout from early on in Linux 5.5, I was curious to see if those regressions have been addressed... Sadly, they are not...

    http://www.phoronix.com/scan.php?pag...rformance-Slip

  • #2
    Kernel goes worse every day. Regressions left and right. Both perf and functionality wise. 5.4 breaks nvidia prime , nvidia hdmi sound etc. Also since the 5.3 there are various sound pop-in,crackling issues.

    I guess it is at a point where everyone says "yolo" and pushes stuff.

    Comment


    • #3
      Originally posted by Leopard View Post
      Also since the 5.3 there are various sound pop-in,crackling issues.
      I'm glad I read your comment, I was getting some weird sound noise recently and was wondering what's wrong (Intel CPU, AMD GPU and MB are brand new).

      Comment


      • #4
        Originally posted by Leopard View Post
        Kernel goes worse every day. Regressions left and right. Both perf and functionality wise. 5.4 breaks nvidia prime , nvidia hdmi sound etc. Also since the 5.3 there are various sound pop-in,crackling issues.

        I guess it is at a point where everyone says "yolo" and pushes stuff.
        more like "LOYP" Linus Only Yelled in the Past.

        Comment


        • #5
          Nobody cares about these regressions? Linus abandones the project?

          Comment


          • #6
            Every non-trivial change in kernel should be benchmarked.

            Comment


            • #7
              Originally posted by pyler View Post
              Nobody cares about these regressions? Linus abandones the project?
              Troll?

              Comment


              • #8
                Okay, so the new scheduling logic seems to have more regressions that expected. It should be interesting to see how this develops.

                Comment


                • #9
                  I wonder how new scheduling code was added to mainline without proper benchmarking. It should be reverted before 5.5 if not fixed properly.

                  Comment


                  • #10
                    Originally posted by pyler View Post
                    I wonder how new scheduling code was added to mainline without proper benchmarking. It should be reverted before 5.5 if not fixed properly.
                    The way I read it on the LKML suggests that it was benchmarked and found to improve performance in all of the test cases, hence the authors pushed it for mainlining. This is probably the step in the right direction in the long run once the regressions are resolved.

                    Comment

                    Working...
                    X