Announcement

Collapse
No announcement yet.

ATI R600g Gains Mip-Map, Face Culling Support

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

  • Agreed. The thing that works in our favour, though, is that as long as anyone working on shader decode starts at the end of the pipe and works backwards, they'll probably run out of development time about the same time the smallest GPUs run out of shader power

    I haven't had time to tinker with any code yet but my feeling is that everything from bitstream parsing to IDCT and inter-prediction should stay on the CPU, while motion comp (intra-prediction) and deblock filtering should go on the GPU. That seems like a good split in the sense that computationally expensive stuff would be on the GPU while "moving fiddly little bits around" would stay on the CPU.

    It's not clear that moving more of the work onto the GPU (ie going further back up the decode pipe) would be a win anyways.

    Comment


    • Originally posted by droidhacker View Post
      tball: have you considered asking either of them for assistance or funding?
      No I haven't. This is a sparetime project only, and I don't want to be bounded by any promises. I am going to USA for some time in a couple of month, and I don't know if I have time developing on GPU decoding over there.

      Comment


      • Originally posted by Hans View Post
        No I haven't. This is a sparetime project only, and I don't want to be bounded by any promises. I am going to USA for some time in a couple of month, and I don't know if I have time developing on GPU decoding over there.
        I guess I am confusing people here. Well Hans = Tball :-)
        i created the user Hans, because I forgot my password for tball. Luckily firefox has stored the password, so I am now back with tball.

        Once in a while I use chromium browser, which is login in automatically as Hans :-)

        Comment


        • Don't need to make any promises. If you tell them of your interests, they may HIRE you or hire someone to do some grunt work and/or even take over the project and do it FOR you.

          Wishful thinking? Never hurts to ask. Especially google -- this could be exactly what they need to get VP8 really off the ground.... and they have TONS of money to throw around.

          Comment


          • Originally posted by bridgman View Post
            One essential part of holding something back is, well, holding it back. I'm not going to give you a list of the stuff we *didn't* release
            I believe he was asking about stuff you *initially* didn't release, but where you changed your mind and actually released it later (which according to you happened in several instances).

            I listed what you added in R5xx 3D acceleration 1.5 that wasn't in 1.4: probably some of that meets the above criteria.

            Comment


            • That wasn't how I understood Q's question but you may be right.

              Comment


              • Originally posted by bridgman View Post
                That wasn't how I understood Q's question but you may be right.
                Q has a very mischievous streak. I suspect that he knows what he's saying a lot better than he makes it look. He probably worded the question precisely to ask both things.

                Comment


                • I wouldn't be surprised. Anyways, I'm not answering both

                  Comment


                  • Bridgman, can you say why you thought fast Z-clears or MSAA were on the hold-back list?

                    Comment


                    • Yeah, we just didn't have time to go through all the potential patent & IP issues and they weren't on the critical path for driver development anyways.

                      We could have delayed the initial info release until these were worked through but splitting them out and coming back to them later seemed like a better approach. We don't talk about things we held back until we actually release them though, since we don't know at the time if we *are* going to be able to release them.

                      Comment

                      Working...
                      X