1. 23 Aug, 2019 1 commit
  2. 22 Aug, 2019 1 commit
  3. 19 Aug, 2019 1 commit
  4. 18 Aug, 2019 1 commit
    • Connor Abbott's avatar
      ac/nir: Fix store_scratch with a non-full writemask · c550d367
      Connor Abbott authored
      By adding one more helper to ac_llvm_build, we can also easily keep
      vector stores together.
      
      Fixes the
      tests/spec/glsl-1.30/execution/fs-large-local-array-vec4.shader_test
      piglit test.
      
      Fixes: 74470bae ("ac/nir: Lower large indirect variables to scratch")
      Reviewed-by: Marek Olšák <marek.olsak@amd.com
      c550d367
  5. 06 Aug, 2019 2 commits
  6. 31 Jul, 2019 2 commits
  7. 29 Jul, 2019 1 commit
  8. 20 Jul, 2019 6 commits
  9. 07 Jul, 2019 1 commit
  10. 04 Jul, 2019 5 commits
  11. 03 Jul, 2019 5 commits
  12. 25 Jun, 2019 1 commit
  13. 24 Jun, 2019 1 commit
  14. 19 Jun, 2019 1 commit
    • Connor Abbott's avatar
      ac,radeonsi: Always mark buffer stores as inaccessiblememonly · 3bf8981c
      Connor Abbott authored
      
      
      inaccessiblememonly means that it doesn't modify memory accesible via
      normal LLVM pointers. This lets LLVM's dead store elimination, memcpy
      forwarding, etc. ignore functions with this attribute. We don't
      represent descriptors as pointers, so this property is always true of
      buffer and image stores. There are plans to represent descriptors via
      pointers, but this just means that now nothing is inaccessiblememonly,
      as LLVM will then understand loads/stores via its usual alias analysis.
      
      Radeonsi was mistakenly only setting it if the driver could prove that
      there were no reads, and then it was cargo-culted into ac_llvm_build
      and ac_llvm_to_nir. Rip it out of everything.
      
      statistics with nir enabled:
      
      Totals from affected shaders:
      SGPRS: 152 -> 152 (0.00 %)
      VGPRS: 128 -> 132 (3.12 %)
      Spilled SGPRs: 0 -> 0 (0.00 %)
      Spilled VGPRs: 0 -> 0 (0.00 %)
      Private memory VGPRs: 0 -> 0 (0.00 %)
      Scratch size: 0 -> 0 (0.00 %) dwords per thread
      Code Size: 9324 -> 9244 (-0.86 %) bytes
      LDS: 2 -> 2 (0.00 %) blocks
      Max Waves: 17 -> 17 (0.00 %)
      Wait states: 0 -> 0 (0.00 %)
      
      The only difference was a manhattan31 shader.
      Acked-by: Timothy Arceri's avatarTimothy Arceri <tarceri@itsqueeze.com>
      Acked-by: default avatarNicolai Hähnle <nicolai.haehnle@amd.com>
      Reviewed-by: default avatarMarek Olšák <marek.olsak@amd.com>
      3bf8981c
  15. 17 Jun, 2019 1 commit
  16. 12 Jun, 2019 1 commit
  17. 04 Jun, 2019 3 commits
  18. 03 Jun, 2019 1 commit
    • Samuel Pitoiset's avatar
      ac,radv: remove the vec3 restriction with LLVM 9+ · 6970a9a6
      Samuel Pitoiset authored
      
      
      This changes requires LLVM r356755.
      
      32706 shaders in 16744 tests
      Totals:
      SGPRS: 1448848 -> 1455984 (0.49 %)
      VGPRS: 1016684 -> 1016220 (-0.05 %)
      Spilled SGPRs: 25871 -> 25815 (-0.22 %)
      Spilled VGPRs: 122 -> 122 (0.00 %)
      Scratch size: 11964 -> 11956 (-0.07 %) dwords per thread
      Code Size: 55324500 -> 55301152 (-0.04 %) bytes
      Max Waves: 235660 -> 235586 (-0.03 %)
      
      Totals from affected shaders:
      SGPRS: 293704 -> 300840 (2.43 %)
      VGPRS: 246716 -> 246252 (-0.19 %)
      Spilled SGPRs: 159 -> 103 (-35.22 %)
      Scratch size: 188 -> 180 (-4.26 %) dwords per thread
      Code Size: 8653664 -> 8630316 (-0.27 %) bytes
      Max Waves: 60811 -> 60737 (-0.12 %)
      Signed-off-by: Samuel Pitoiset's avatarSamuel Pitoiset <samuel.pitoiset@gmail.com>
      Reviewed-by: default avatarMarek Olšák <marek.olsak@amd.com>
      6970a9a6
  19. 16 May, 2019 1 commit
    • Marek Olšák's avatar
      ac: rename SI-CIK-VI to GFX6-GFX7-GFX8 · ccfcb9d8
      Marek Olšák authored
      
      Acked-by: default avatarDave Airlie <airlied@redhat.com>
      
      We already use GFX9 and I don't want us to have confusing naming
      in the driver. GFXn naming is better from the driver perspective,
      because it's the real version of the gfx portion of the hw. Also,
      CIK means Bonaire-Kaveri-Kabini, it doesn't mean CI.
      
      It shouldn't confuse our SDMA, UVD, VCE etc. code much. Those have
      nothing to do with GFXn and they have their own version numbers.
      ccfcb9d8
  20. 14 May, 2019 1 commit
  21. 13 May, 2019 1 commit
  22. 02 May, 2019 1 commit
  23. 23 Apr, 2019 1 commit