Announcement

Collapse
No announcement yet.

The v2 Rotary Interactivity Favor Scheduler

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

  • The v2 Rotary Interactivity Favor Scheduler

    Phoronix: The v2 Rotary Interactivity Favor Scheduler

    Version 2 of the Rotary Interactivity Favor Scheduler (RIFS) was published on Friday. The RIFS scheduler for the Linux kernel is designed for driving a low-latency Linux desktop...

    Phoronix, Linux Hardware Reviews, Linux hardware benchmarks, Linux server benchmarks, Linux benchmarking, Desktop Linux, Linux performance, Open Source graphics, Linux How To, Ubuntu benchmarks, Ubuntu hardware, Phoronix Test Suite

  • #2
    Interesting. They also mention a latency benchmark app called latt. I don't know if it's already in pts but it would be a great addition.

    Comment


    • #3
      Almost no buzz on the net about this scheduler! Real low-profile stuff... anyone using it out there? Comparisons vs BFS?

      Comment


      • #4
        Originally posted by elmariachi View Post
        Almost no buzz on the net about this scheduler! Real low-profile stuff... anyone using it out there? Comparisons vs BFS?
        it's true that it handles interactivity under heavy load better than bfs, but almost every cpu intensive task seems to be slower with the patch. i'm going to this patch again tomorrow.

        Comment


        • #5
          Originally posted by raj7095 View Post
          it's true that it handles interactivity under heavy load better than bfs, but almost every cpu intensive task seems to be slower with the patch. i'm going to this patch again tomorrow.
          Well, that is the price you pay for interactivity or low latency.
          Seems like the main problem is that io kthreads aren't preemptible with the kernel most desktops use. Please Ubuntu, Debian, Mint, Fedora, and Suse, make the preempt kernel default.
          If someone is running a server then either offer a server spin, or the less preemptive kernel for them in the repos.
          Without these more strongly preemptive kernels schedulers just don't matter very much. Sure, you'll see some differences, but you'll see more differences between them if you change kernels. This is why I think Android didn't go with bfs(or some other scheduler than cds).

          Has anyone heard of testing of this sort being openly performed?
          Last edited by liam; 14 May 2012, 02:39 PM.

          Comment


          • #6
            I tried it over the weekend, it has issues...

            Sure for the foreground task it seems to be giving more CPU time to it, but it is a bit too agressive.

            Basically I was playing fullyscreen HeroesOfNewerth and would tabback to browse the net or IRC.

            I usually run a kernel with the Brain Fuck Scheduler with no issue.
            I tried a new kernel with this RIFS and switching between tasks has a clear lag (ok maybe a price to pay if the foreground app is given more priority). However, after the desktop had been up for about 4hours and HoN had been quit and restarted a number of times I ran into an issue where I entered a game and I was getting 5fps (usually 45) and everything was sluggish. restart of HoN didn't fix, restart of X didn't fix (ie something more fundemental).
            I rebooted to my BFS to find they didn't pause and got a leave (and -10mmr )

            so yer not that great of a schedules, im sticking with stock or BFS

            Comment


            • #7
              Hmm.

              Originally posted by Naib View Post
              I tried it over the weekend, it has issues...

              Sure for the foreground task it seems to be giving more CPU time to it, but it is a bit too agressive.

              Basically I was playing fullyscreen HeroesOfNewerth and would tabback to browse the net or IRC.

              I usually run a kernel with the Brain Fuck Scheduler with no issue.
              I tried a new kernel with this RIFS and switching between tasks has a clear lag (ok maybe a price to pay if the foreground app is given more priority). However, after the desktop had been up for about 4hours and HoN had been quit and restarted a number of times I ran into an issue where I entered a game and I was getting 5fps (usually 45) and everything was sluggish. restart of HoN didn't fix, restart of X didn't fix (ie something more fundemental).
              I rebooted to my BFS to find they didn't pause and got a leave (and -10mmr )

              so yer not that great of a schedules, im sticking with stock or BFS
              Hi have you tried rebooting your box?If you have tried rebooting that means it is not the bug of RIFS. If the lagging problem is gone after you rebooted your machine, it is my mistake and please contact with me by email:[email protected] to help me to work on the bug by giving me the config or something else. ;-)
              Last edited by 3766691; 15 May 2012, 09:12 AM.

              Comment


              • #8
                Originally posted by raj7095 View Post
                it's true that it handles interactivity under heavy load better than bfs, but almost every cpu intensive task seems to be slower with the patch. i'm going to this patch again tomorrow.
                Thanks for giving a like on my RIFS

                Comment


                • #9
                  Originally posted by 3766691 View Post
                  Hi have you tried rebooting your box?If you have tried rebooting that means it is not the bug of RIFS. If the lagging problem is gone after you rebooted your machine, it is my mistake and please contact with me by email:[email protected] to help me to work on the bug by giving me the config or something else. ;-)
                  The lagging problem was gone after reboot and switching back to my backup kernel

                  Comment


                  • #10
                    Originally posted by Naib View Post
                    The lagging problem was gone after reboot and switching back to my backup kernel
                    Thanks for waiting. Which version of3.x kernel are you using?

                    Comment

                    Working...
                    X