Rendering acceleration by mesh optimization

The rendering speed is limited by the slowest per-frame operation. Such per-frame operations are the

• scene graph traversal which is done by the CPU. During this stage the represented data will be generated and prepared for the following steps in the graphics pipeline.
• transformations. This stage is limited by the rate at which the graphics hardware can process vertices. Therefore, those limits can be influenced by the data only.
• rasterization which is limited by the rate at which the graphics hardware can update the frame buffer. Here, one deciding factor is the size of the viewport.

In the subtopics the results of two alternatives are presented. Both reduce the number of vertices that have to be sent to the graphics subsystem in order to speed-up the rendering:

Stripping of quadrilaterals	Mesh simplification