Announcement

Collapse
No announcement yet.

Significant Performance Improvements For RadeonSI LLVM With New Patches

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

  • Significant Performance Improvements For RadeonSI LLVM With New Patches

    Phoronix: Significant Performance Improvements For RadeonSI LLVM WIth New Patches

    Tom Stellard is looking to kick things off well in 2015 for the open-source AMD GPU LLVM back-end with some new patches to boost the performance...

    http://www.phoronix.com/vr.php?view=MTg3NzM

  • #2
    si-scheduler-v3 was fine for me last month , but this one perf-Dec31-2014 (and i already tried it yestrerday still has that performance improvemant, but it also has broken rendering ... in Unigine Valley, Tesseract, etc...
    Last edited by dungeon; 01-01-2015, 05:27 PM.

    Comment


    • #3
      Lets take a screenshot:

      Comment


      • #4
        Nothing for us, AMD trinity APU folks?

        Comment


        • #5
          Originally posted by dungeon View Post
          si-scheduler-v3 was fine for me last month , but this one perf-Dec31-2014 (and i already tried it yestrerday still has that performance improvemant, but it also has broken rendering ... in Unigine Valley, Tesseract, etc...
          Did you see any performance improvement in non-OpenCL tasks?
          ## VGA ##
          AMD: X1950XTX, HD3870, HD5870
          Intel: GMA45, HD3000 (Core i5 2500K)

          Comment


          • #6
            Originally posted by darkbasic View Post
            Did you see any performance improvement in non-OpenCL tasks?
            Yeah, this is similar to:

            http://www.phoronix.com/scan.php?pag...tem&px=MTgzMjE

            Only now has that render bug in many games so i did not benchmarking yet, but on eye that is it well yeah perf is similar like nearly 2 months ago with si-scheduler-v3

            Comment


            • #7
              Am I the only one who would rather be able to use the OpenCL (missing image support etc. if you need anything more complex than simple computing) than knowing that the implementation is fast?

              Comment


              • #8
                Yes, I agree currently OpenCL sucks. Lots of workloads simply don't work...
                ## VGA ##
                AMD: X1950XTX, HD3870, HD5870
                Intel: GMA45, HD3000 (Core i5 2500K)

                Comment


                • #9
                  That would be a pretty wallpaper if the green border was faded out instead of a sharp triangle

                  Comment


                  • #10
                    It's great to see the open-source AMD Linux stack already off to a good start in 2015 given the incredible AMD OSS improvements made in 2014 and moving closer to performance parity with AMD Catalyst over the past year of improvements to the Linux kernel Radeon DRM, Mesa/Gallium3D, and LLVM.
                    I don't quite undestand the difference between various AMD drivers. AMD Catalyst drivers on Windows and Linux are also using LLVM. Do they have two different LLVM based compilers - one open-source and other closed source. Or is the compiler part shared between both drivers stacks?

                    Comment

                    Working...
                    X