Announcement

Collapse
No announcement yet.

AMD Publishes Open-Source Graphics Code For Mullins & Beema

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

  • AMD Publishes Open-Source Graphics Code For Mullins & Beema

    Phoronix: AMD Publishes Open-Source Graphics Code For Mullins & Beema

    AMD has started rolling out their open-source Linux graphics driver support code for their recently announced Mullins and Beema APUs...

    Phoronix, Linux Hardware Reviews, Linux hardware benchmarks, Linux server benchmarks, Linux benchmarking, Desktop Linux, Linux performance, Open Source graphics, Linux How To, Ubuntu benchmarks, Ubuntu hardware, Phoronix Test Suite

  • #2
    Congratulations AMD!

    It might not be pre-launch like Intel, but we're getting closer to that.
    Even if it looks it was mostly a matter of adding "case CHIP_MULLINS" and "|| xxx == CHIP_MULLINS", I still find this pretty awesome.

    Comment


    • #3
      I don't have that hardware, but great to see Mullins/Beema opensource support that early .

      Running now Athlon 5350 2046@2457 MHz / memory 1600@1872 MHz runs great for 2 days now , of course with SATAs turned off for now, maybe later bioses will fix that maybe not .

      Comment


      • #4
        15-16 weeks is absurd.
        AMD needs to fix this crap.
        There has to be a better way.

        Comment


        • #5
          Originally posted by grndzro View Post
          15-16 weeks is absurd.
          AMD needs to fix this crap.
          There has to be a better way.
          ????

          Sorry but: huh?



          Yeah maybe it would have been great if they could have made it into 3.15 but we all know that Linus is not fond of late additions of that magnitude to release candidates. And we are in the RC phase.
          Stop TCPA, stupid software patents and corrupt politicians!

          Comment


          • #6
            Originally posted by grndzro View Post
            15-16 weeks is absurd.
            AMD needs to fix this crap.
            There has to be a better way.
            If you can't wait, run blob earlier maybe or just use non released kernel code .

            Comment


            • #7
              Originally posted by Adarion View Post
              Yeah maybe it would have been great if they could have made it into 3.15 but we all know that Linus is not fond of late additions of that magnitude to release candidates. And we are in the RC phase.
              ...and even that doesn't prevent distributor who won't be using 3.16 to provide patch to their kernel and handle the newer hardware on older kernels.
              Specially since, as mentionned above:
              Originally posted by Spittie
              Even if it looks it was mostly a matter of adding "case CHIP_MULLINS" and "|| xxx == CHIP_MULLINS", I still find this pretty awesome.
              it looks that patching support isn't that much complicated.

              Which make sense.
              The role of a kernel module is to handle initializing/power management/modesetting. This doesn't have any reason to change that much between recent hardware revision.
              Then it only passes through the commands that it receives by state tracker or back-ends (like from Gallium3D's radeonsi backend). Slightly newer hardware probably means slightly different commands, but it's not the kernel's job to handle them. That would be the job of Gallium3D's back-end.

              Comment


              • #8
                Originally posted by Adarion View Post
                ????

                Sorry but: huh?



                Yeah maybe it would have been great if they could have made it into 3.15 but we all know that Linus is not fond of late additions of that magnitude to release candidates. And we are in the RC phase.
                He's made exceptions in the past for hardware enablement as long as the code doesnt introduce bugs in existing code. Reasoning being that the situation literally can't get worse for someone with unsupported hardware (ignoring insta-kernel panics, but those should be caught)
                All opinions are my own not those of my employer if you know who they are.

                Comment


                • #9
                  Seems AMD heard the cries of anguish.
                  Phoronix, Linux Hardware Reviews, Linux hardware benchmarks, Linux server benchmarks, Linux benchmarking, Desktop Linux, Linux performance, Open Source graphics, Linux How To, Ubuntu benchmarks, Ubuntu hardware, Phoronix Test Suite
                  Stop TCPA, stupid software patents and corrupt politicians!

                  Comment


                  • #10
                    Slightly wrong product numbers

                    Hello Michael, quoting this news article:
                    The AMD Mullins APUs include the A10 Micro-6700T, A4 Micro-6400T, and E1 Micro-6200Tl. The Beema APUs include the E2-6010, E2-6110, A4-6210, and A6-6310.
                    and finding the same mistakes in your original Beema/Mullins news, I just want to pont out that two number/character combinations are wrong:

                    For Mullins, it's "E1 Micro-6200T" rather than "E1 Micro-6200Tl" and the weakest Beema is named an E1, not an E2, making it "E1-6010", which you called "E2-6010".

                    Sorry for being pedantic

                    See AMD's slides for reference.

                    Comment

                    Working...
                    X