Announcement

Collapse
No announcement yet.

Tear-Free Acceleration For ATI EXA, Xv

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

  • Tear-Free Acceleration For ATI EXA, Xv

    Phoronix: Tear-Free Acceleration For ATI EXA, Xv

    For those of you that have been using the open-source xf86-video-ati driver, need we remind you of its rapidly-improving state and feature set? One of the latest additions to this open-source ATI driver that supports the old ATI R100 graphics cards up through the new Radeon HD 4800 series (RV770) is tear-free acceleration. The current implementation of this tear-free acceleration is for EXA and Textured Video (X-Video) and should eliminate any "tearing" issues that some users experience...

    http://www.phoronix.com/vr.php?view=NjU5NQ

  • #2
    Does it mean we'll have a 6.9.1 "tear-free" ati driver pretty soon?

    My old Nvidia with proprietary driver displays the most tearish 2d I've ever seen... My next card will be an Ati, no more doubt.

    Comment


    • #3
      One of the latest additions to this open-source ATI driver that supports the old ATI R100 graphics cards up through the new Radeon HD 4800 series (RV770) is tear-free acceleration. The current implementation of this tear-free acceleration is for EXA and Textured Video (X-Video) and should eliminate any "tearing" issues that some users experience.
      That sounds like there is already EXA and Xv support for the R7xx series. But that's not true is it?
      I had to read this part 3 times before I understood it correctly

      Comment


      • #4
        Sounds like the radeonhd devs are losing groung in what seems to be an internal developer's competition of who will bring out the cooler driver.

        Anyone cares to remind me why we're having two open working drivers again?

        Comment


        • #5
          As somebody who tried to get Xv and other things working in the fglrx bad-old-days, I read the title as, "Tears-of-frustration free".

          J1M.

          Comment


          • #6
            Originally posted by sundown View Post
            Sounds like the radeonhd devs are losing groung in what seems to be an internal developer's competition of who will bring out the cooler driver. Anyone cares to remind me why we're having two open working drivers again?
            This patch (and radeon itself) handles four generations of cards which radeonhd does not support. We are trying to keep largely common acceleration code between radeon and radeonhd anyways, so I don't think there is any competition here.

            Check out the "quick & dirty 2d" branch of radeonhd and let us know what you think.

            Comment


            • #7
              Awesome! As soon as this code hits master I'll switch back to radeon.

              Comment


              • #8
                Originally posted by bridgman View Post
                We are trying to keep largely common acceleration code between radeon and radeonhd anyways, so I don't think there is any competition here.
                I see, but then what is the reason for having two drivers? they should merge, or separate completely (up to 600 managed by radeon, the upper by radeonhd).

                Comment


                • #9
                  The cleanest cut would probably have been between r500 and r600, so all <=r500 would have been supported by radeon (as the r500 and the previos Radeon Chips share much acceleration-code) and everything >=r600 would have been up to radeonhd. But at the time radeonhd was started it was about supporting the Radeon-Chips which were not already supported by radeon then, and thus they started working on r500.

                  Comment


                  • #10
                    Yeah -- if you look at the display controller then 4xx/5xx is the logical split, but if you look at acceleration then 5xx/6xx makes a lot more sense.

                    The radeonhd initiative was driven by the need for modesetting support, while the post-4xx radeon work was driven by the need for acceleration support. We hope the two efforts can pull together over time. I suspect that once we reach a common understanding on requirements, priorities, design / coding practices and the importance of things like kernel modesetting the rest will be easy
                    Last edited by bridgman; 07-16-2008, 04:23 PM.

                    Comment


                    • #11
                      Originally posted by bridgman View Post
                      Yeah -- if you look at the display controller then 4xx/5xx is the logical split, but if you look at acceleration then 5xx/6xx makes a lot more sense.

                      The radeonhd initiative was driven by the need for modesetting support, while the post-4xx radeon work was driven by the need for acceleration support. We hope the two efforts can pull together over time. I suspect that once we reach a common understanding on requirements, priorities, design / coding practices and the importance of things like kernel modesetting the rest will be easy
                      I read that the reason Red Hat is devoting man hours in this is because they have customers demand. Or if I remember right, they said they had one customer that could see a business model for OSS gfx drivers.

                      So I am wondering, is Novell's and Red Hat's demands/needs the same?

                      E.g. does Red Hat put emphasis on something else than Novell?

                      Comment


                      • #12
                        AFAIK the original reason for pushing ahead with radeon was that RH had a specific customer program which needed 2d acceleration and didn't feel that radeonhd would be able to make the schedule based on progress to date. By adding atombios support to radeon they were able to use all of the existing 2d accel code since the 2d block was pretty much identical to previous generations.

                        A contributing factor was that both the radeon and avivo developers were OK in principle with writing a new, atombios based driver for 5xx and above rather than building on their existing code, but when radeonhd turned out to be mostly hard-coded the idea of adding atombios-based support to radeon started to look more attractive.
                        Last edited by bridgman; 07-16-2008, 05:49 PM.

                        Comment


                        • #13
                          Wow.

                          I really think that Red Hat and IBM are interesting companies. I fear to think how much Red Hat spend on Ice Tea!

                          Have Novell said anything on why they are dedicating so much effort and money?

                          Comment


                          • #14
                            Hmmm, I might have to ditch my geforce 6200 in my tower in favor of my old Radeon 9000 Pro. The extent of my 3d needs are Chromium BSU (when I'm too lazy to go into the other room and fire up R-Type III). As long as I get damn-near-spotless 2d support (which Nvidia has been sliding downhill with) I'm happy.

                            Comment


                            • #15
                              Originally posted by Louise View Post
                              Have Novell said anything on why they are dedicating so much effort and money?
                              I think everyone is trying to achieve the same ends -- to advance what can be done with Linux and expand the potential market. Both Novell and RH are trying to "give back" where they can. There are some differences in the RH and Novell customer bases today and so they do have some different short term priorities. Please don't ask me for details
                              Last edited by bridgman; 07-16-2008, 05:55 PM.

                              Comment

                              Working...
                              X