Announcement

Collapse
No announcement yet.

Clear Linux Continues To Have Graphics Performance Advantage Over Ubuntu

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

  • Clear Linux Continues To Have Graphics Performance Advantage Over Ubuntu

    Phoronix: Clear Linux Continues To Have Graphics Performance Advantage Over Ubuntu

    Earlier this week I published some Windows 10 vs. Ubuntu Linux OpenGL benchmarks showing how the native gaming performance is different between the competing platforms. Ubuntu Linux lost nearly all of those results with the Intel Mesa driver to Windows 10. In this article are those previous benchmarks plus now having Intel Clear Linux benchmarks added in the mix. Months ago in previous tests we've found Clear Linux to have faster Linux graphics performance than other distributions.

    http://www.phoronix.com/vr.php?view=23422

  • #2
    Impressive ClearLinux results !

    Comment


    • #3
      For an OS with such different goals, it continues to impress... me, at least!

      Comment


      • #4
        So where does it come from? Is it the compiler flags? Should we pay Gentoo more attention again as a logical result for non-intel users?

        Comment


        • #5
          intel have their own distro and the gpu performance is worst than windows? what a bad joke

          Comment


          • #6
            Originally posted by andre30correia View Post
            intel have their own distro and the gpu performance is worst than windows? what a bad joke
            Windows is highly tuned for gaming performance, and Microsoft have the Xbox which runs Windows.

            Linux isn't really popular for gaming, it have smaller teams of developers for device drivers for graphics and is not finely tuned for gaming.

            It is no surprise that Windows is ahead of Linux in gaming performance.

            Comment


            • #7
              Uh so most benchmark give windows 10 driver a >2 or 3 times performance advantage of linux driver? Is it wise to still follow that goal? When virtualization with gpu-passthrough will give ~10%-30% performance hit? (I'm not sure you can do that with intel graphics tough).

              Windows is highly tuned for gaming performance, and Microsoft have the Xbox which runs Windows.
              Yeah, but it doesn't run any intel hardware (cpu or graphic) neither it use opengl.

              It is no surprise that Windows is ahead of Linux in gaming performance.
              One can hope...

              Comment


              • #8
                Originally posted by uid313 View Post

                Windows is highly tuned for gaming performance, and Microsoft have the Xbox which runs Windows.

                Linux isn't really popular for gaming, it have smaller teams of developers for device drivers for graphics and is not finely tuned for gaming.

                It is no surprise that Windows is ahead of Linux in gaming performance.
                Games favouring DX and the rest of the stack doesn't help either. I'm thankful that gallium-nine exists, hopefully Vulkan can pick up and replace DX12.

                Comment


                • #9
                  Microsoft didn't tune anything OpenGL wise. Something is funny here. There shouldn't be a reason the Linux driver is multiples slower. I wonder what the actual technical cause is, and not just speculative popularity-of-OS answers...
                  ​​​​​

                  Comment


                  • #10
                    Originally posted by g7RbdHRt View Post
                    Uh so most benchmark give windows 10 driver a >2 or 3 times performance advantage of linux driver? Is it wise to still follow that goal? When virtualization with gpu-passthrough will give ~10%-30% performance hit? (I'm not sure you can do that with intel graphics tough).
                    Virtualization with gpu-passthrough did not give me stable frame rates and my performance hit was more than 30% in games compared to Windows 10. I am sure this is not IO related yet I did not do extensive profiling to find out exactly what the cause was. I do competitive gaming, so unfortunately due to spikes I cannot depend on "virtualization" at this point in time. For the casual gamer it could be acceptable performance hit. I would love to use zfs for my main filesystem or btrfs when it becomes stable. For now I am stuck with ntfs on my desktop.

                    Software
                    1st: Ubuntu 15.04 failed to get iommu working, had to do a few OS/grub recoveries along the way.
                    2nd: Arch with heavily patched 3.18 kernel, worked could run steam and play games.
                    3rd: Arch community 3.19 image, worked with less hassle.

                    Hardware
                    * CPU: i7 3770
                    * RAM: 32GB (4x8) G.Skill 1866 C9
                    * M/B: Gigabyte GA-Z77X-UD3H
                    * GPU: 3GB R9 280X
                    * SSD: 256GB Crucial M550
                    * LCD1 24" Samsung SyncMaster BX2485
                    * LCD2 24" Sentech TV (South African brand connect to Intel HD 4000)
                    * Two sets of Mouse & Keyboard (I was able to separate the USB controllers yay!)

                    PS: I have no idea if performance is more stable with newer hardware / kernels. I would like to try this again when my nitro RX 480 arrives and I find time.

                    PPS: I had to update my M/B BIOS to get the audio to work in Linux. I am now running the F20e beta BIOS which sometimes causes system crashes depending on how many USB devices I use, this happens under any OS even in the BIOS itself. Damn Gigabyte...

                    Comment

                    Working...
                    X