Announcement

Collapse
No announcement yet.

ETC2 Texture Compression Looks Good For OpenGL

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

  • ETC2 Texture Compression Looks Good For OpenGL

    Phoronix: ETC2 Texture Compression Looks Good For OpenGL

    With OpenGL ES 3.0 and OpenGL 4.3 there is now mandatory texture compression support in the form of ETC2, the Ericsson Texture Compression method...

    http://www.phoronix.com/vr.php?view=MTE1ODU

  • #2
    This is good but, when will wayland will full suport OpenGL so we can use it in he future?

    Comment


    • #3
      Originally posted by Thaodan View Post
      This is good but, when will wayland will full suport OpenGL so we can use it in he future?
      This is also required by the new openGL ES spec, which Wayland uses. Also, the only reason Wayland uses openGL ES is that the only available full openGL implementation at the moment is GLX which has a dependency on X.

      Comment


      • #4
        Any statement yet from the big three on when we can expect this on desktop gpus, in hardware? Bridgman?

        Comment


        • #5
          So how well does ETC2 stack up with ASTC? Seems we now have several texture compression alternatives to S3TC, what would be the best for games performance and memorywise?

          Originally posted by curaga View Post
          Any statement yet from the big three on when we can expect this on desktop gpus, in hardware? Bridgman?
          I'm sure that the major GPU makers (Intel, AMD and nVidia) will have their implementations and would be good to see the feedback from their respective devs. Hopefully we'll see benchmarks on all these compression methods soon
          Last edited by DeepDayze; 08-12-2012, 05:31 PM.

          Comment


          • #6
            interesting reading, i would like to know if Sandy/Ivy Bridge will support it soon?

            Comment


            • #7
              Originally posted by DeepDayze View Post
              So how well does ETC2 stack up with ASTC? Seems we now have several texture compression alternatives to S3TC, what would be the best for games performance and memorywise?
              To be fair, we have a few years at least for these to mature and get various improvements and optimizations in hardware and encoders. You can't ship a game using either of these today. Or next year. Or even three years from now. Far too few users would have hardware that supports the features. (drivers might support them, but if they're not hardware supported, we won't care.)

              It's a lot like SSE, where SSE3/4 are only just recently becoming feasible to use in a game. It wasn't that long ago that even SSE2 was just a theoretical optimization despite having been in shipping hardware for years. It's a bit easier with code since the binaries are small and you can just ship several versions of the compiled game (Crysis did this, for instance), but for content you'd be duplicating potentially gigabytes of data for each compression algorithm or hardware feature set.

              The Steam hardware survey is really useful to figure out things like this. You can see that some 99% of Steam users have SSE2 now and over 98% have SSE3 and there's no reason not to use them anymore, but only 50% have SSE4 so it's still a no-go (and it added the dot-product operator, dammit... soon). You can see that WinXP is under 15% of users (and see that most of those have older hardware), and hence D3D10/11 are finally viable to use as your sole graphics API for high-sales-volume games. You can see that less than 7% have single-core CPUs, and only 50% of people are still using dual-core CPUs, so you can plan to require threaded engines today and prepare for requiring 4+ cores in the near future. You can see that there are still almost 30% of users with 512MB or less of video ram, so keeping low-quality (and highly compressed) textures is still a must for a little while longer. You can see that 50% still have less than 4GB of RAM and wonder why nobody is telling everyone that RAM is practically free right now. You can see that OS X usage for Steam users is still hovering at 5%, which is higher than some people expect, lower than many people hope, and can be extrapolated to what Linux might look like.

              You can also look at GPUs, see rate of uptake, and figure out what GPU features you can rely on. After ASTC support actually arrives in real hardware, you can track how long it takes for a significant-enough portion of the market to have hardware supporting, and when you can start using it in shipping titles.

              Comment


              • #8
                Originally posted by elanthis View Post
                You can see that WinXP is under 15% of users (and see that most of those have older hardware), and hence D3D10/11 are finally viable to use as your sole graphics API for high-sales-volume games.
                If only games were developed for PCs and not ported from aging consoles...

                Comment


                • #9
                  Originally posted by ShadowBane View Post
                  This is also required by the new openGL ES spec, which Wayland uses. Also, the only reason Wayland uses openGL ES is that the only available full openGL implementation at the moment is GLX which has a dependency on X.
                  Is it so hard to clear GLX from X11 depencies or rewrite a lib that gives full OpenGL support?

                  Comment


                  • #10
                    Originally posted by Thaodan View Post
                    Is it so hard to clear GLX from X11 depencies or rewrite a lib that gives full OpenGL support?
                    The real problem is manpower, Making libGL work without GLX would be nice, and will happen at some point in the future, but right now there isn't enough intrest for this to be a major push.

                    Comment

                    Working...
                    X