1 / 29

Questions about GPUs

Discover insights into suspicious behaviors of GPUs and explore improvements in areas like 3D textures, deferred shading, and pre-multiplied alpha textures, as discussed by experts in Bordeaux. Learn about challenges and proposed solutions.

ahintz
Download Presentation

Questions about GPUs

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Questions about GPUs AS-Temps réel – Bordeaux Philippe Decaudin Fabrice Neyret Stéphane GuySylvain Lefebvre

  2. Overview • Suspicious behaviors of GPUs. • What could be improved quickly. • Our wishes.

  3. 1. Suspicious behaviors • 3D Textures, MIP-mapping and anisotropy • Clamp to border and MIP-mapping • Deferred shading for conditions

  4. 3D textures and MIP-mapping • Filtering of 3D textures is isotropic(same MIP-map level for u, v and w) • When texturing a sliceof volume with a 3D texture, the slope of the slice controls the blursloppy => blurry • With 2D textures, this problem is solved by anisotropic filtering, but: - it is not available for 3D textures… - no dsdw

  5. Clamp_to_border and MIP-mapping • Using non-square textures (2a 2b, a b) • CLAMP_TO_BORDER_ARB enabled • MIP-map enabled • at MIP-map levels > log2(a) (with a < b), the texture becomes 1D: clamp doesn’t seem to work correctly…

  6. Clamp_to_border and MIP-mapping • a quad mapped with a white texture 25664 • (u,v) = (0,0) to (6,6) • mapping clamped to [0,1][0,1] • border color is black • MIP-map level forced with GL_TECTURE_MIN_LOD (6,6) Clamp problem bug? Square texture (6464) correct MIP-map level = 1 (25664) MIP-map level = 8 (11) (0,0) Texture 25664

  7. Clamp_to_border and MIP-mapping  all MIP-map levels 128x32 64x16 32x8 16x4 8x2 4x1 2x1 1x1

  8. Deferred shading for conditions Theory: • Custom shader: • costly particular case for some pixels • general case fast. 1 fast, but a few pixels are wrong

  9. Deferred shading for conditions Theory: 1 2 flag wrong pixels in stencil fast, but a few pixels are wrong

  10. Deferred shading for conditions Theory: 1 2 flag wrong pixels in stencil fast, but a few pixels are wrong 3 use slow correction shader only where needed

  11. Deferred shading for conditions 1 2 flag wrong pixels in stencil fast, but a few pixels are wrong 3 • Practice: • Does not work !? • cost the same as full rendering • stencil after fragment program ? use slow correction shader only where needed

  12. 2. What could be improved quickly • Front to back rendering with alpha • Fog and pre-multiplied alpha • Scale / Bias not limited to [0,1] • Tiled textures storage

  13. Front to back rendering with alpha • Theory: front to back (to benefit Z-test culling) • Practice with alpha: must disable depth_test ! landscapes, billboards, volumes… even more costly • Suggested solution:glEnable(ALPHA_DEPTH_TEST)doing regular blending, thenif (Ascreen>Athreshold) Zscreen:=Zfrag

  14. Fog and pre-multiplied alpha • Premultiplied alpha textures (aR,aG,aB,a) to avoid interpolation artefacts. • Blend mode is then set to glBlendFunc(GL_ONE,GL_ONE_MINUS_SRC_ALPHA)

  15. Fog and pre-multiplied alpha • Problem: not compatible with standard Fog equation (glEnable(GL_FOG))Fog equation: Cfrag’= (1-f) Cfrag+ f CfogIn our case, Cfrag = aC (premultiplied alpha) Blend equation: Cdest’=Cfrag’+ (1-a)Cdest Cdest’= (1-f) aC + f Cfog + (1-a)Cdest missing ‘a’ • Suggestion: • if SRC_ALPHA == GL_ONE then • fog equation: (1-f) C + f aCfog

  16. Scale / Bias not limited to [0,1] • Theory:scale, bias to tune transparency, light, color… • Practice:coef must lies in [0,1]!  applications which need coef>1 must multipass (volume rendering, light maps…) • Suggested Solution:Float operation in texture units for scaling.

  17. Tiled textures storage • Principle: • Use a set of tiles to create a larger texture fragment program

  18. Tiled textures storage • Numerous new advanced usage of tiles • Antialiased Parameterized Solid Texturing [Hart et al]ACM Transactions on Graphics2002 • Adaptive Texture Maps [Kraus and Ertl] Graphics Hardware 2002 • Pattern Based Procedural Textures [Lefebvre and Neyret] I3D 2003 • … and more !

  19. Tiled textures storage • Storage requirements • Efficient selection from FP • No waste of memory • Ease filtering / interpolation of tiles • per-tile border color • per-tile MIP-mapping

  20. Tiled textures storage • Current solutions • Store as separate textures • limited bindings • no real selection in FP • Store in an atlas • false neighborhood and no border color • waste space if filtering on tiles • limited to 4096x4096 • Store as a stack in a 3D Texture • waste a lot of space (2n) • ill-defined filtering

  21. t r s level 0 - 32x32x4 level 1 - 16x16x4 level 2 - 4x4x4 Tiled textures storage • Our proposal: Texture stack • 3D Texture with no filtering on R • per-tile border colors • per-tile filtering • easy selection from fragment program

  22. 3. Our wishes • Having features more coherent • programable blending • feedback buffer • read access to all state variables (ex: MIP-mapping) • r/w stencil from FP • Native support of tiled textures • Fragment culling order (stencil, A, depth) • Tools for procedural textures

  23. Having features more coherent • Programable blending: • multi-pass algorithms with intermediate results • needs complex blending • current blending very limited compared to programable pipeline

  24. Having features more coherent • Feedback buffer: • lot of vertex attributes • few can be read back

  25. Having features more coherent • Perspective correction rasterization in object space not always desired  need control on variables interpolated by rasterizer glEnable/Disable(param,GL_PERSPECTIVE_CORRECTION)

  26. Native support of tiled textures • Difficulties • Storage (borders, repeat, …) • MIP-mapping • Linear interpolation

  27. Fragment culling order • Early culling essential for performance • Would let the user specify test order between • alpha test • stencil • depth • discard (from FP)

  28. Tools for procedural textures New texture types (envmap, indir., N, shaders, …) • custom filtering • MIP-mapping levels • have to be infered from ddx ddy • while computed for standard textures • needs to be accessible (read) • resolution specifiable by user

  29. Questions about GPUs AS-Temps réel – Bordeaux Philippe Decaudin Fabrice Neyret Stéphane GuySylvain Lefebvre

More Related