chrisforbes/mesa

Issues

• EXT_transform_feedback for ironlake

  #80 opened 7 years ago by mirh
  10

• Promote small dense UBO access to push constants

  #50 opened 7 years ago by chrisforbes
  3

• tess: Support barrier function

  #64 opened 7 years ago by chrisforbes
  0

• tess: Support vertex indexed URB read in HS

  #66 opened 7 years ago by chrisforbes
  0

• tess: Specialize HS generation on ES domain type

  #68 opened 7 years ago by chrisforbes
  0

• arb_flv: Support conformant out-of-range handling via SF read offset mangling

  #70 opened 7 years ago by chrisforbes
  0

• tess: Implement gl_PatchVerticesIn in HS

  #71 opened 7 years ago by chrisforbes
  0

• tess: Implement gl_PatchVerticesIn in DS

  #72 opened 7 years ago by chrisforbes
  0

• conservative rasterization: Figure out exactly what the windows driver is doing

  #77 opened 7 years ago by chrisforbes
  0

• tess: If >1 hw instance, use and reset semaphore in HS epilog

  #60 opened 7 years ago by chrisforbes
  1

• tess: Support masked URB writes for HS output

  #57 opened 7 years ago by chrisforbes
  0

• tess: Support URB readback from HS outputs

  #58 opened 7 years ago by chrisforbes
  0

• tess: Support URB read from patch URB entry in DS

  #63 opened 7 years ago by chrisforbes
  0

• tess: Consider SIMD8 dispatch in DS.

  #62 opened 7 years ago by chrisforbes
  0

• size of stride-0 attributes is wrong

  #79 opened 10 years ago by chrisforbes
  0

• atomics: Consider using append counter on surface state

  #73 opened 10 years ago by chrisforbes
  0

• Gen7+: Consider append counter atomics rather than addressed atomics, when not arrayed.

  #78 opened 10 years ago by chrisforbes
  1

• tess: If >1 hw instance, allocate and zero 32dw of URB space for semaphores

  #59 opened 10 years ago by chrisforbes
  2

• tess: In HS/DS prolog, fix execution mask for upper half if last instance and #verts odd.

  #61 opened 10 years ago by chrisforbes
  1

• tess: Specialize HS generation on input vertex count

  #69 opened 10 years ago by chrisforbes
  1

• tess: In HS epilog, if we are the final thread, release all the vertex URB handles.

  #67 opened 10 years ago by chrisforbes
  1

• tess: Flag some state when glPatchParameter[if]v is used.

  #76 opened 10 years ago by chrisforbes
  1

• tess: Configure topology filter to match patch size

  #65 opened 10 years ago by chrisforbes
  1

• Investigate Gen7 tesselation stuff

  #46 opened 10 years ago by chrisforbes
  1

• msaa: Figure out exactly what NV MFAA does, and why fiddling with sample positions is better than adjusting projection matrix by <1px

  #75 opened 10 years ago by chrisforbes
  1

• msaa: Allow user-defined sample positions

  #74 opened 10 years ago by chrisforbes
  0

• Gen4/5: Experiment with FS-based clipdistance

  #56 opened 11 years ago by chrisforbes
  0

• View format overrides -- store the mesa_format in the intel_texobj.

  #55 opened 11 years ago by chrisforbes
  1

• Implement 2x MSAA hack for Gen6/7

  #54 opened 11 years ago by chrisforbes
  2

• ARB_texture_multisample depth sampling

  #53 opened 11 years ago by chrisforbes
  1

• Investigate multiple viewport rendering

  #42 opened 11 years ago by chrisforbes
  0

• Guess better about whether a FS is likely to be used with an FBO.

  #51 opened 11 years ago by chrisforbes
  2

• Add diagnostic support for FS dispatch efficiency

  #52 opened 11 years ago by chrisforbes
  0

• Gen7+: Consider URB atomics when decrement operation not used

  #49 opened 11 years ago by chrisforbes
  0

• ARB_draw_indirect support (Gen7+)

  #32 opened 11 years ago by chrisforbes
  5

• Fix glitchy behavior in CS:S death notifications

  #41 opened 11 years ago by chrisforbes
  0

• Collect a list of cases which force ARB_draw_indirect to take a slow path

  #39 opened 11 years ago by chrisforbes
  0

• Investigate bogus cycle count estimates

  #37 opened 11 years ago by chrisforbes
  0

• Gen7 BLORP for R8G8B8A8 <-> R10G10B10A2?

  #48 opened 11 years ago by chrisforbes
  1

• Investigate rendering issue with nouveau etc -- some rendering gets clipped to topleft 1Kx1K.

  #43 opened 11 years ago by chrisforbes
  1

• Missing GLSL1.4 feature on Ironlake: UBO

  #33 opened 12 years ago by chrisforbes
  1

• GLSL FS backend: Extremely poor code generation for loops that reference constant arrays!

  #35 opened 12 years ago by chrisforbes
  3

• Investigate HiZ support for ILK

  #44 opened 12 years ago by chrisforbes
  1

• Fix oilrush on SNB+

  #47 opened 12 years ago by chrisforbes
  1

• ARB_fragment_program support: allow benign respecification of precision hints

  #34 opened 12 years ago by chrisforbes
  2

• Investigate misclipping in water in HL2 on ILK

  #38 opened 12 years ago by chrisforbes
  4

• Write dynamic-resolution demo

  #45 opened 12 years ago by chrisforbes
  0

• Investigate setting 3DSTATE_DRAWING_RECTANGLE to something other than the whole surface

  #40 opened 12 years ago by chrisforbes
  0

• Investigate whether fragment color clamping setup in precompile can be chosen better

  #36 opened 12 years ago by chrisforbes
  0

• Write piglits to demonstrate stencil bug, and ensure no regression

  #31 opened 12 years ago by chrisforbes
  1