Announcement

Collapse
No announcement yet.

Are Intel Skylake Graphics Any Faster With Linux 4.4?

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

  • Are Intel Skylake Graphics Any Faster With Linux 4.4?

    Phoronix: Are Intel Skylake Graphics Any Faster With Linux 4.4?

    Following the recent Mesa 11.2-devel Git tests on Skylake I also ran a comparison to see if the OpenGL performance differed at all when comparing Linux 4.3.0 to Linux 4.4 in its second release candidate form...

    http://www.phoronix.com/scan.php?pag...4-SKL-Graphics

  • #2
    - No.

    Comment


    • #3
      Michael,

      I'm interested in the ide power consumption of mobile SKL, BDW and HSW chips compared to Windows 10. Any chance you can run such a comparison? Please!!!!!

      After over two years of experimetation in enabling all powersave options my HSW notebook still burns nearly twice as much idle power compared to Windows 10. I wonder if BDW/SKL is doing any better - or worse?

      Cheers
      Saski

      Comment


      • #4
        Originally posted by saski View Post
        Michael,

        I'm interested in the ide power consumption of mobile SKL, BDW and HSW chips compared to Windows 10. Any chance you can run such a comparison? Please!!!!!

        After over two years of experimetation in enabling all powersave options my HSW notebook still burns nearly twice as much idle power compared to Windows 10. I wonder if BDW/SKL is doing any better - or worse?

        Cheers
        Saski
        Nope, sorry, not without several donors/premium members coming up for that testing.... Would need to port some code over to be able to read my AC meter in the same manner under Windows as I do in Linux, carrying out three Windows installations is time intensive, etc.
        Michael Larabel
        http://www.michaellarabel.com/

        Comment


        • #5
          Originally posted by saski View Post
          Michael,

          I'm interested in the ide power consumption of mobile SKL, BDW and HSW chips compared to Windows 10. Any chance you can run such a comparison? Please!!!!!

          After over two years of experimetation in enabling all powersave options my HSW notebook still burns nearly twice as much idle power compared to Windows 10. I wonder if BDW/SKL is doing any better - or worse?
          Evidently you either haven't enabled all the powersave options, or they're not all exposed to the tuning knobs you're using otherwise there'd be no difference. Presumably you've tried running powertop to see what's going on? Windows is very well tuned for power consumption out of the box, and vendors tend to improve that with tweaks for their hardware. Linux distributions just aren't typically tuned for power consumption at all, stability always comes first as there's no way of knowing how well a random system will work with aggressive power saving.

          Comment


          • #6
            Originally posted by s_j_newbury View Post
            Evidently you either haven't enabled all the powersave options, or they're not all exposed to the tuning knobs you're using otherwise there'd be no difference. Presumably you've tried running powertop to see what's going on? Windows is very well tuned for power consumption out of the box, and vendors tend to improve that with tweaks for their hardware. Linux distributions just aren't typically tuned for power consumption at all, stability always comes first as there's no way of knowing how well a random system will work with aggressive power saving.
            I've configured the kernel to expose any powersave option it can. I've also set ASPM to powersave. powertop --auto-tune should do the job and enables powersave everything it can and it seems that it does when I look at the tunables. I even forced i915.ebable_psr=1 so that the system enters Package PC7 to some degree. So I guess there is something not exposed to be enabled for agressive power save.

            Comment


            • #7
              RAPL?

              Comment

              Working...
              X