Announcement

Collapse
No announcement yet.

R600 Gallium3D Shader Compiler Milestone Hit

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

  • #71
    Probably more correct to say that we started in 2007 with a gigantic backlog and by doing some work on older chips we haven't whittled down that backlog quite as quickly as if development and documentation efforts had only focused on the newest hardware.

    I don't think starting any earlier on Evergreen would have worked well until we had a solid 6xx/7xx base to build on, and that only happened near the end of 2009.
    Test signature

    Comment


    • #72
      Actually I like to reserve "gigantic" for *really* big things, and we only started 5 years behind, so let's say we started with a "substantial" backlog. If there had been no open source ATI drivers or driver development community at all in 2007 that would have been a "gigantic" backlog
      Test signature

      Comment


      • #73
        So if we pay people to work on the open source drivers, what does that make us ?
        Test signature

        Comment


        • #74
          that makes you totally awesome!

          Comment


          • #75
            The "Free with capital F" crowd is never going to be satisfied. First, they nag for hardware specs because they want to write the drivers themselves. They got them, but where are the drivers? Oh yeah, they didn't write them, because now it's still AMD's job.

            Comment


            • #76
              Originally posted by RealNC View Post
              The "Free with capital F" crowd is never going to be satisfied. First, they nag for hardware specs because they want to write the drivers themselves. They got them, but where are the drivers? Oh yeah, they didn't write them, because now it's still AMD's job.
              Thats not exactly true,

              Duecher and Bridgeman work for AMD, butArlied works for redhat. Luc used to work for SuSe, not sure where he works now. And then there are about 5 or more other regular contributors that I cant remember there names that dont work for AMD either.

              Actually I think most people are more upset about all the in house development they are doing behind closed doors. They are writing the core structure of these drivers by themselves without any input or advice from the community. I personally dont like it. I would much rather see documentation, then development in public.

              Comment


              • #77
                Originally posted by Qaridarium
                witch one is more evil a assassin who kills for money to feed his child's

                or the people who pays people to kill ?


                you and all payed devs are ''Mercenary's'' for earning money fo feed ths own child's---

                in my Point of view evil means the people who pays people to kill not the one who kills to become food.
                Let me translate this into understandable English:

                ``Hey I go by the name of Qaridarium, but please call me Gary, for my name is just as unpronounceable as my own English.

                I have a philosophic rhetorical question for you:
                What is worse:
                A) The boss of a corporate entiety that forces it's personel to code FLOSS drivers, or;
                B) The people hired to code a FLOSS driver,

                It is ofcourse answer A) for all corporations that order to make a FLOSS drivers are wrong. They have evil plans to take over the world and make everybody a corporate ATI slave!

                Floss we will be. For thy my lord. For thy power havith decendant from thy hand, that our shader cores may swiftly carry out thy command. So we shall flow; a river forth through the binary and Flash memmory. Illuminate ATI and AMD and it will get funky,,

                *Bonus points if you recognise the last paragraph*

                Comment


                • #78
                  RealNC was referring to the point that people generally complain to/about AMD if the driver development isn't moving fast enough, not the rest of the development community.

                  Nearly all of the development *does* take place in public, and code generally moves into public repos as soon as we have something "lighting up" (eg the 3D code moved into public as soon as the r600demo code could reliably draw a triangle).

                  The only exception I can think of is the 6xx/7xx 3D driver, where Richard leveraged some in-house diags code and we had to go through a code sanitizing exercise, but in that case we knew the code was going to be replaced with a new driver following the Gallium3D structure anyways so didn't care too much about the details.
                  Test signature

                  Comment

                  Working...
                  X