Game Engine Architecture

408 10. The Rendering Engine

In a strip, the fi rst three vertices defi ne the fi rst triangle. Each subsequent

vertex forms an entirely new triangle, along with its previous two neigh-

bors. To keep the winding order of a triangle strip consistent, the previous

two neighbor vertices swap places aft er each new triangle. A triangle strip is

shown in Figure 10.8.

In a fan, the fi rst three vertices defi ne the fi rst triangle and each subse-

quent vertex defi nes a new triangle with the previous vertex and the fi rst ver-

tex in the fan. This is illustrated in Figure 10.9.

Vertex Cache Optimization

When a GPU processes an indexed triangle list, each triangle can refer to any

vertex within the vertex buff er. The vertices must be processed in the order

they appear within the triangles, because the integrity of each triangle must be

maintained for the rasterization stage. As vertices are processed by the vertex

shader , they are cached for reuse. If a subsequent primitive refers to a vertex

that already resides in the cache, its processed att ributes are used instead of

reprocessing the vertex.

Strips and fans are used in part because they can potentially save memory

(no index buff er required) and in part because they tend to improve the cache

coherency of the memory accesses made by the GPU to video RAM. Even

bett er, we can use an indexed strip or indexed fan to virtually eliminate vertex

duplication (which can oft en save more memory than eliminating the index

buff er), while still reaping the cache coherency benefi ts of the strip or fan ver-

tex ordering.

Indexed triangle lists can also be cache-optimized without restricting

ourselves to strip or fan vertex ordering. A vertex cache optimizer is an offl ine

geometry processing tool that att empts to list the triangles in an order that

Figure 10.9. A triangle fan.

Interpreted

as triangles:^012023034

Vertices V 0 V 1 V 2 V 3 V 4

V 0

V 4

V 3 V 2

V 1