Announcement

Collapse
No announcement yet.

Intel SNA vs. UXA On Ivy Bridge (July 2012)

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

  • Intel SNA vs. UXA On Ivy Bridge (July 2012)

    Phoronix: Intel SNA vs. UXA On Ivy Bridge (July 2012)

    Here are some recent benchmarks comparing Intel's SNA and UXA 2D acceleration architectures offered by their open-source Intel Linux graphics driver.

    http://www.phoronix.com/vr.php?view=17656

  • #2
    I'm still with UXA because glx-dock (cairo-dock) gets corrupted very quickly with SNA.

    Comment


    • #3
      Intel are doing awesome work, very impressive.

      Comment


      • #4
        And usability?

        Some people on different forums reported weird flickering and artefacts with Intel HD 2000/3000 GPUs rendering this new feature useless. Can someone confirm it?

        Comment


        • #5
          I'm using SNA on my OpenSuse 12.1 under an Ivy Bridge CPU, I don't have any graphical artefact. I noticed that it seems smoother in general.


          I'd like to precise that the line to add in xorg.conf is exactly (in the driver section):

          Code:
          Option "AccelMethod" "SNA"

          Comment


          • #6
            Originally posted by roland View Post
            I'm still with UXA because glx-dock (cairo-dock) gets corrupted very quickly with SNA.
            Do you have a link to this bugreport?

            Comment


            • #7
              Originally posted by ssvb View Post
              Do you have a link to this bugreport?
              No, I was hoping it would go away with 2.20.1 but no luck.

              Comment


              • #8
                I'm happy

                Intel is doing an AMAZING job. Congrats!

                Comment


                • #9
                  Originally posted by Redi44 View Post
                  Some people on different forums reported weird flickering and artefacts with Intel HD 2000/3000 GPUs rendering this new feature useless. Can someone confirm it?
                  Confirmed. I have an i5-2467M (Sandy Bridge). Using SNA I started getting corruptions as soon as X is started. It usually happens with overlay dialogs with drop shadows. Switching back to UXA, corruptions are all gone.

                  Comment


                  • #10
                    How do I figure out how what acceleration method is currently being used? Is there some kind of command I can issue? I just want to be able to make sure I have my xorg options set correctly

                    Comment


                    • #11
                      Originally posted by whaevr View Post
                      How do I figure out how what acceleration method is currently being used? Is there some kind of command I can issue? I just want to be able to make sure I have my xorg options set correctly
                      If you didn't explicitly enable SNA (see one of the previous posts), you are using UXA.

                      Comment


                      • #12
                        Looking at /var/log/Xorg.0.log might give some clues about that, too.

                        Comment


                        • #13
                          Latest intel driver in arch linux has sna support (if you do the xorg.conf change), so I enabled it on my ironlake machine to test. Can't say I've noticed much difference in real world usage yet, but it didn't break anything either so I guess there's that

                          Does anyone know what kind of "real world" usage would show sna's improvements?
                          Last edited by bwat47; 07-31-2012, 05:31 PM.

                          Comment


                          • #14
                            Originally posted by GreatEmerald View Post
                            Looking at /var/log/Xorg.0.log might give some clues about that, too.
                            Correct, I just checked mine and it does seem to tell you if its initialized:

                            [ 12.881] (II) intel(0): SNA initialized with Ironlake backend

                            Comment


                            • #15
                              Originally posted by bwat47 View Post
                              Latest intel driver in arch linux has sna support (if you do the xorg.conf change), so I enabled it on my ironlake machine to test. Can't say I've noticed much difference in real world usage yet, but it didn't break anything either so I guess there's that

                              Does anyone know what kind of "real world" usage would show sna's improvements?
                              Well, I've obviously completedly failed if you haven't noticed any of your common tasks improved by SNA. So what is your common workload?

                              Comment

                              Working...
                              X