Architecture
R200's 3D hardware consists of 4 pixel pipelines, each with 2 texture sampling units. It has 2 vertex shaders and a legacy Direct3D 7 TCL unit, marketed as Charisma Engine II. It is ATI's first GPU with programmable pixel and vertex processors, called Pixel Tapestry II and compliant with Direct3D 8.1. R200 has advanced memory bandwidth saving and overdraw reduction hardware called HyperZ II that consists of occlusion culling (hierarchical Z), fast z-buffer clear, and z-buffer compression. The GPU is capable of dual display output (HydraVision) and is equipped with a video decoding engine (Video Immersion II) with adaptive hardware deinterlacing, temporal filtering, motion compensation, and iDCT.
R200 introduced pixel shader version 1.4 (PS1.4), a significant enhancement to prior PS1.x specifications. Notable instructions include "phase", "texcrd", and "texld". The phase instruction allows a shader program to operate on two separate "phases" (2 passes through the hardware), effectively doubling the maximum number of texture addressing and arithmetic instructions, and potentially allowing the number of passes required for an effect to be reduced. This allows not only more complicated effects, but can also provide a speed boost by utilizing the hardware more efficiently. The "texcrd" instruction moves the texture coordinate values of a texture into the destination register, while the "texld" instruction will load the texture at the coordinates specified in the source register to the destination register.
Compared to R100's 2x3 pixel pipeline architecture, R200's 4x2 design is more robust despite losing one texture unit per pipeline. Each pipeline can now address a total of 6 texture layers per pass. The chip achieves this by using a method known as 'loop-back'. Increasing the number of textures accessed per pass reduces the number of times the card is forced into multi-pass rendering.
The texture filtering capabilities of R200 are also improved over its predecessor. For anisotropic filtering, Radeon 8500 uses a technique similar to that used in R100, but improved with trilinear filtering and some other refinements. However, it is still highly angle-dependent and the driver sometimes forces bilinear filtering for speed. NVIDIA's GeForce 4 Ti series offered a more accurate anisotropic implementation, but with a greater performance impact.
R200 has ATI's first implementation of a hardware-accelerated tessellation engine (a.k.a. higher order surfaces), called Truform, which can automatically increase the geometric complexity of 3D models. The technology requires developer support and is not practical for all scenarios. It can undesirably round-out models. As a result of very limited adoption, ATI dropped TruForm support from its future hardware.
| DirectX 8.0 Pixel Shader 1.1 |
DirectX 8.1 Pixel Shader 1.4 |
||
|---|---|---|---|
| Max. Texture Inputs | 4 | 6 | |
| Max. Program Length | 12 instructions (up to 4 texture sampling, 8 color blending) |
22 instructions (up to 6 texture sampling, 8 texture addressing, 8 color blending) |
|
| Instruction Set | 13 address operations, 8 color operations | 12 address / color operations | |
| Texture Addressing Modes | 40 | virtually unlimited |
Read more about this topic: Radeon R200
Famous quotes containing the word architecture:
“The principle of the Gothic architecture is infinity made imaginable.”
—Samuel Taylor Coleridge (1772–1834)
“I don’t think of form as a kind of architecture. The architecture is the result of the forming. It is the kinesthetic and visual sense of position and wholeness that puts the thing into the realm of art.”
—Roy Lichtenstein (b. 1923)
“They can do without architecture who have no olives nor wines in the cellar.”
—Henry David Thoreau (1817–1862)