Announcement

Collapse
No announcement yet.

Testing Primitive Binning With Vega 10 On RadeonSI

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

  • Testing Primitive Binning With Vega 10 On RadeonSI

    Phoronix: Testing Primitive Binning With Vega 10 On RadeonSI

    One month back Marek Olšák landed support in the RadeonSI Gallium3D driver for primitive binning with Vega 10 GPUs but now that feature is likely to be disabled by default...

    http://www.phoronix.com/scan.php?pag...m-Binning-Test

  • #2
    Interesting that it made no difference for you. I've been looking at implementing binning in radv, and when I enable the dfsm bits I typically get a significant slowdown. Are you sure the change got picked up?

    Comment


    • #3
      Originally posted by BNieuwenhuizen View Post
      Interesting that it made no difference for you. I've been looking at implementing binning in radv, and when I enable the dfsm bits I typically get a significant slowdown. Are you sure the change got picked up?
      Sure it got picked up unless anything super peculiar... it's set for both the PTS terminal and separate Steam process. Works out fine for testing other R600_DEBUG settings, so unless any odd behavior or requirements I should be aware of?
      Michael Larabel
      http://www.michaellarabel.com/

      Comment


      • #4
        Originally posted by Michael View Post

        Sure it got picked up unless anything super peculiar... it's set for both the PTS terminal and separate Steam process. Works out fine for testing other R600_DEBUG settings, so unless any odd behavior or requirements I should be aware of?
        Thanks for checking. I'm not aware of anything special needing to happen, and haven't tested GL, so I've got nothing to preclude your results.

        Comment


        • #5
          Originally posted by BNieuwenhuizen View Post

          Thanks for checking. I'm not aware of anything special needing to happen, and haven't tested GL, so I've got nothing to preclude your results.
          Happy to rerun the tests or if anything else comes to light. Had communicated with Marek via email when he expressed interest in seeing some tests done and followed his guidance. If you ever have any benchmark requests, feel free to email me directly at michael at phoronix.com and happy to carry out more tests, etc.
          Michael Larabel
          http://www.michaellarabel.com/

          Comment


          • #6
            I think many people are interested in such VEGA benchmarks. Thank you Michael.
            It's still puzzling why VEGA (despite these new shiny features) mostly only behaves like a higher clocked Fiji...

            Comment


            • #7
              Originally posted by mibo View Post
              I think many people are interested in such VEGA benchmarks. Thank you Michael.
              It's still puzzling why VEGA (despite these new shiny features) mostly only behaves like a higher clocked Fiji...
              got me wondering too.
              I can only say the stock cooler in balanced mode (Windows) is too damn hot, near 100C while powersave works for 99% of games at max settings @ 4K but an EK block should make that issue disappear.
              it gets roughly 1350 mhz in my computer @ balanced, my case, my ambient - despite living close to the artic circle I have it rather warm inside so compared to a fury it's not a whole lot faster than fury so apart from added memory capacity I don't know what's new but instead of being a niche architecture it'll be a mainstream arch. with new APU's so I hope we see some good performance improvements through drivers.

              I've looked at some benchmarks of scaling, I've tested 56 bios vs 64 and performance is so near that I start to think the arch is really designed for the lower end hardware...
              Speculation and curiosity is strong on vega.

              Comment

              Working...
              X