Aktív témák

• #592 FreeM aktív tag

  2004-04-23 19:30:41 #592

  Összes hozzászólása itt Válaszok az összes hozzászólására itt Válaszok erre a hozzászólásra

  Privát üzenet küldése

  

  FreeM

  aktív tag

  Andrey Khonich ( Crytek ) nyilatkozata a SM3.0-ról:
  
  ''In VS3.0 Shader model actually is possible to support general displacement mapping (with smart Shader design when vertex Shader has to do something during waiting for texture access).
  
  In PS3.0 shaders it’s possible to decrease number of shaders using dynamic branching (one Shader for general lighting) and in such way decrease number of Shader switches and as result increase speed, and also we can utilize dynamic conditional early reject for some cases in both PS and VS and this also will increase speed. As to NV40 generally possible to use co-issues better to take advantage of super-scalar architecture (we can execute 4 instructions per cycle in a single pipeline).
  
  - Handle several light sources in single pixel shaders by using dynamic loops in PS3.0.
  - Decrease number of passes for 2-sided lighting using additional face register in PS3.0.
  - Use geometry instancing to decrease number of draw-calls (remove CPU limitations as much as possible).
  - Unrestricted dependent texture read capabilities to produce more advanced post-processing effects and other in-game complex particles/surfaces effects (like water).
  - Full swizzle support in PS3.0 to make better instructions co-issue and as result speedup performance.
  - Increase quality of lighting calculations using 32 bit precision in pixel shaders on NV40.
  - Take advantage of using 10 texture interpolators in PS3.0 Shader model to reduce number of passes in some cases.
  - Easily do multiple pass high-dynamic-range rendering on FP16 targets.
  - Speed up post-processing a lot by using MRT capabilities and mip-mapping support on non-power-of-two textures.
  
  For PS3.0 shader model 32 bit precision becomes high precision. 24-bit and lower are partial precision. Currently I see advantages of using 32 bit over 24 bit only in calculating of depth values in pixel shader (custom Z values). Also it could be possible to use 32 bit precision with big values, but it’s not the case in current generation engine.''
  
  
  
  Tim Sweeny ( Epic Games )
  ''PS 3.0 supports dynamic conditionals and much larger Shader programs, so it becomes possible to render in a single pass what would previously take many passes. For example, a renderer can more practically combine multiple lightsources into a single pass, using a dynamic conditional to handle the properties of each individual light in a single pass.
  
  Regarding the general question of ''Why are conditionals needed in pixel Shader programs'', I recommend comparing the situation to that of C programs. How many useful C programs don't contain conditionals? Approximately none!
  
  Only a marketing guy would consider 24-bit floating point to be ''full precision''!
  
  32-bit floating point has been an industry standard for 18 years, since the original IEEE 754 specification was ratified. Besides the obvious need for 32-bit precision in any sort of high-range calculation (such as position vector math, world transformations, etc - all things one is likely to do frequently per-pixel), we feel there's a huge advantage to using industry standard arithmetic in its predictability and reproducibility across different GPUs. The 24-bit precision mode allowed by the DirectX9 spec was backed by a reasonable cost/performance/precision judgment back in the early days of DirectX in 2002-2003. But nowadays there isn't a good reason for hardware to support less than full 32-bit floating point precision.''

Aktív témák