Announcement

Collapse
No announcement yet.

Future of ATI Linux drivers

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

  • #21
    Originally posted by bridgman View Post
    Do you think that would be useful ? I don't think it was something we ever considered. Is there a benefit to having software scaling via Xv, maybe some app that only works with Xv and not with another API ? I *think* we'd have to write the scaling and CSC code anyways, just in x86 rather than shaders, wouldn't we ?
    Well, currently the problem is that without Xv support, it literally breaks the use of most people's computers for their intended uses. So until the Xv (and other 2D accel) are fixed, the whole deal is sour.

    There are a lot of people running MythTV under Ubuntu, for example. If you're running a media server with the graphics card connected to a large format display (a projector, for example), and you can't get usable video output, then it makes the whole system worthless.

    The same goes for people who use a PC or laptop as a multi-role device, for work and for personal use. I fall into that category. I'm trapped into using an X1400 because that's what was available in the laptop that I purchased, and I have a fixed long-term equipment investment in that. The fact that I can watch a video without a problem in Windows but it looks like garbage under Linux is very frustrating, especially when it's obvious that it's not the fault of the OS.

    It makes the device useless for doing video editing, as well. What about those "workstation" customers? Or does "workstation" only mean 3D modeling in ATI's view?

    Comment


    • #22
      Originally posted by curaga View Post
      Hmm. At least with Nvidia moving 2d operations to 3d units dropped the performance considerably, every user is complaining how their new 8800 gt is way slower with Firefox and desktop in general than their old 7950 ultra..

      So, how is the situation with AMD currently? Are the 2d operations slower, equal, or faster with latest (3d) hardware compared to the ones with dedicated 2d units?
      Generally the 3d hardware is faster, but only on the more recent chips. The older chips had specialized 2d hardware because it could do "simple things really fast", faster than the more general 3d engine.

      The main change, though, is not so much "3d hardware faster than 2d hardware", but rather "the stuff people call 2d operations today need the 3d engine anyways"
      Test signature

      Comment


      • #23
        ** PLEASE IGNORE THIS POST -- see #28 below **

        Originally posted by Porter View Post
        Well, currently the problem is that without Xv support, it literally breaks the use of most people's computers for their intended uses. So until the Xv (and other 2D accel) are fixed, the whole deal is sour.
        But... don't most of the player apps support other output mechanisms, including X11 ? I understand the importance of video -- just trying to understand how an unaccelerated Xv implementation would help.

        Originally posted by Porter View Post
        There are a lot of people running MythTV under Ubuntu, for example. If you're running a media server with the graphics card connected to a large format display (a projector, for example), and you can't get usable video output, then it makes the whole system worthless.
        Sure, but realistically unaccelerated Xv is not going to be satisfying here either. I think we need acceleration for that, and that needs the 3d engine.

        Originally posted by Porter View Post
        The same goes for people who use a PC or laptop as a multi-role device, for work and for personal use. I fall into that category. I'm trapped into using an X1400 because that's what was available in the laptop that I purchased, and I have a fixed long-term equipment investment in that. The fact that I can watch a video without a problem in Windows but it looks like garbage under Linux is very frustrating, especially when it's obvious that it's not the fault of the OS.
        Again, I don't understand how an unaccelerated Xv interface would help if playing to (say) X11 did not work for you.

        Originally posted by Porter View Post
        It makes the device useless for doing video editing, as well. What about those "workstation" customers? Or does "workstation" only mean 3D modeling in ATI's view?
        The specialized code paths are primarily for 3d modelling. Again, how is an unaccelerated Xv interface going to help ? I understand the need for video, just not the "Xv without acceleration" part.
        Last edited by bridgman; 14 February 2008, 09:06 PM.
        Test signature

        Comment


        • #24
          Bridgman,

          Thanks for your participation!

          I wasn't suggesting unaccelerated Xv, I was just saying that any 2D accel support is better than none. It was my understanding that X11 video playback was broken on fglrx as well, that scaling to anything above 1:1 causes terrible pixelation, just as it does in Xv.

          Or am I smoking crack? Does basic video playback work properly (no pixelation or horizontal tearing) using X11? Maybe I'm missing something, and should go back and revisit the issue. Apologies to everyone if that is the case.

          Comment


          • #25
            Well i will just wait and see what will happen. It doesn't seem that discussing stuff here would change anything.

            The bridge watching over this forum is pretty picky who may travel along. Of course thats fine and has to be done but i cant free myself from the feeling that john?/bridgeman does not really get what linux is or why it exists. Even if people wish for it. It was never made to replace windows.

            Sometimes i wish Linus would do something like microsoft and just disallow prop. drivers. Then we would have a final decision: no support or good one. But linus says that the distibutions have to take that step before the official linux kernel will follow.

            And dont missunderstand me. I am NOT wishing for that because it would give us better driver but maybe it could change the manufacturers view of linux.

            martin alias cleric

            Comment


            • #26
              but i cant free myself from the feeling that john?/bridgeman does not really get what linux is or why it exists. Even if people wish for it. It was never made to replace windows.
              I agree completely. The issue here is not Windows -- it's that Linux has made a lot of progress replacing Unix. Those markets (server, workstation, HPC) were perfectly comfortable with closed-source products, and some of the newer Linux segments like embedded (your DRM-infested DVD player probably runs Linux internallY) often have no choice but to use closed drivers because their market requires DRM.

              That said, there is also a large group of users for whom free and open drivers are essential. For those users, we are making sure you have access to good open source drivers as well.

              One of the challenges has always been that different people have very different views of what Linux is about. What do you think it is ?

              The bridge watching over this forum is pretty picky who may travel along.
              I didn't fully understand this. Am I "the bridge" here, or are you talking about Phoronix moderators and other forum members ?
              Last edited by bridgman; 14 February 2008, 08:59 PM.
              Test signature

              Comment


              • #27
                Originally posted by Porter View Post
                Bridgman,

                Thanks for your participation!

                I wasn't suggesting unaccelerated Xv, I was just saying that any 2D accel support is better than none. It was my understanding that X11 video playback was broken on fglrx as well, that scaling to anything above 1:1 causes terrible pixelation, just as it does in Xv.

                Or am I smoking crack? Does basic video playback work properly (no pixelation or horizontal tearing) using X11? Maybe I'm missing something, and should go back and revisit the issue. Apologies to everyone if that is the case.
                (deleted evidence of my mixing up cleric's & Porter's posts

                The fglrx driver already has accelerated Xv using TexturedVideo on 5xx and up and VideoOverlay for the older ASICs which had video processing built into the overlay block. Right now there is tearing on the TexturedVideo implementation (a diagonal line), but you should see improvements there soon. X11 is a good option on the radeonhd driver today but Xv is a better choice on fglrx.
                Last edited by bridgman; 14 February 2008, 09:10 PM.
                Test signature

                Comment


                • #28
                  Porter, I just noticed that I mixed your original post up with cleric's. Forget everything I said about unaccelerated Xv interfaces -- that was from Cleric's question.

                  Cleric, my mistake. It was you asking about the unaccelerated Xv interface not Porter. Can you fill me in on how that would help vs. just using X11 ?

                  Personally ... i wouldn't care if they implement it in software and emulate the xv interface. That would help a lot of people. Just as a filler of course..
                  Thanks,
                  John
                  Last edited by bridgman; 14 February 2008, 09:11 PM.
                  Test signature

                  Comment


                  • #29
                    Originally posted by bridgman View Post
                    That said, there is also a large group of users for whom free and open drivers are essential. For those users, we are making sure you have access to good open source drivers as well.
                    Care to speculate on where i might find this excellent offering for an ati card near me?

                    Comment


                    • #30
                      It depends on which card you have, but either radeon or radeonhd.

                      The radeon driver seems to be pretty well regarded except for 3d which was largely reverse engineered. The first 3d docco package should fill in a lot of the gaps for the earlier ASICs as well as 5xx family.

                      Radeonhd seems to be solid on the display/modesetting side, and is picking up 2d and DRM as we speak... um... post. The 3d code will be pretty much common from 3xx through 5xx and RS690, so to a large extent 3d on those all those chips should improve together, and right now it's looking like 6xx will share a lot of the code as well.
                      Test signature

                      Comment

                      Working...
                      X