Announcement

Collapse
No announcement yet.

Etnaviv Gallium3D Is Almost To OpenGL 2.0 Compliance

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

  • Etnaviv Gallium3D Is Almost To OpenGL 2.0 Compliance

    Phoronix: Etnaviv Gallium3D Is Almost To OpenGL 2.0 Compliance

    The Etnaviv Gallium3D driver that provides reverse-engineered, open-source graphics support for Vivante graphics hardware is almost to exposing OpenGL 2.0...

    http://www.phoronix.com/scan.php?pag...-At-OpenGL-2.0

  • #2
    I can recall that I used an vivante arm board back in 2013. Are these things still in use after all?

    Comment


    • #3
      Yes, they were used in the GCW Zero console.

      Comment


      • #4
        Can anyone please add it to Mesa's features.txt so Mesa matrix could display it?

        And 3.0 only? That's too bad, KDE could benefit from 3.1. Is that GPU so old? I thought all modern GPUs already support 4.6 features.

        UPDATE:

        Hm. i.MX 8 lists Vulkan as supported. That doens't make any sense. If it supports Vulkan, it should support 4.x OpenGL feature set.
        Last edited by shmerl; 10-17-2017, 07:47 PM.

        Comment


        • #5
          Originally posted by shmerl View Post
          If it supports Vulkan, it should support 4.x OpenGL feature set.
          The feature set target for Vulkan 1.0 was OpenGL ES 3.1, not OpenGL 4.x, there are some notable differences (fp64 among them).

          Comment


          • #6
            Originally posted by microcode View Post

            The feature set target for Vulkan 1.0 was OpenGL ES 3.1, not OpenGL 4.x, there are some notable differences (fp64 among them).
            What else is the key difference? Ivy Bridge for example got away with implementing fp64 in software.

            Comment


            • #7
              Originally posted by Kemosabe View Post
              I can recall that I used an vivante arm board back in 2013. Are these things still in use after all?
              Solidrun stuff, Wandboard stuff, and others use iMX6 SoCs with a Vivante GPU in them.

              Newer iMX8 SoCs have also Vivante GPUs.
              Last edited by starshipeleven; 10-18-2017, 02:56 AM.

              Comment


              • #8
                Originally posted by shmerl View Post
                Hm. i.MX 8 lists Vulkan as supported. That doens't make any sense. If it supports Vulkan, it should support 4.x OpenGL feature set.
                iMX8 is using Vivante GC7000 (two per SoC, for some reason) GPUs.
                https://en.wikipedia.org/wiki/I.MX#i.MX_8_series

                Their GPU support OpenGL ES 3.1 http://www.vivantecorp.com/index.php...to-mobile.html

                iMX6 chips are 2011 designs, and use older GPUs.

                Comment


                • #9
                  Originally posted by starshipeleven View Post
                  iMX8 is using Vivante GC7000 (two per SoC, for some reason) GPUs.
                  https://en.wikipedia.org/wiki/I.MX#i.MX_8_series
                  You need to read the right spec sheet.
                  https://www.nxp.com/docs/en/fact-sheet/IMX8FAMFS.pdf
                  They do say it has a Vulcan driver in the newer spec sheets talking about the Vivante GC7000.

                  That pdf was reference 13 on the wikipedia link. starshipeleven it does pay to check the references on wikipedia articles as this can do better than google searching something.

                  https://en.wikipedia.org/wiki/Vulkan_(API)
                  Vulkan bare minimal is Opengl ES 3.1

                  And I can answer why it has two gpus.

                  SafeAssure® ASIL-B ready hardware protects critical visual information with fail-over-capable quality of service to any display.

                  It is this bit in the spec sheet. So you can run your output interface twice and if one gpu has a issue and locks up and need a restart you can switch to the backup GPU. Its quite a common reason you see gpu times 2 instead of deploying the twice as powerful version of the same GPU that would take the same space in silicon.
                  This does mean on embedded you do need to be able to use multi gpu for a single output at times to extract the most performance out of particular soc chips. When you are needing durability gpu times 2 is good but can make the graphical management API more complex when the how the split is configured and change over is software managed.

                  Some SOC can see complete isolation between cpu cores and other bits for same reason of running the interface twice with rapid switch over in case of issue.

                  Comment


                  • #10
                    Originally posted by oiaohm View Post
                    You need to read the right spec sheet.
                    I got the same information from wikipedia in 5 seconds flat and it is also correct, so what's the point of this schooling?

                    Comment

                    Working...
                    X