Alternate frame rendering
Alternate Frame Rendering (AFR) is a method of graphics rendering in personal computers which combines the work output of two or more video processors (or separate video graphics cards) for a single monitor, in order to improve image quality or to accelerate the rendering performance. The technique, in essence, is that one graphics processing unit (GPU) computes all the odd video frames, the other renders the even frames. This technique is useful for generating 3D video sequences in real time, improving or filtering textured polygons and performing other computationally intensive tasks, typically associated with computer gaming, CAD and 3D modeling.[1]
One disadvantage of AFR is a defect known as micro stuttering.
Parallel rendering methods
AFR belongs to a class of parallel rendering methods, which subdivide a four dimensional image frame sequence (x,y,z and time) into smaller regions, each of which is then assigned to a different physical processor within a multi-processor array. Note that the regional boundaries may be defined in space or in time. Also, the multiple processors can be implemented within a single video card or separate video graphics cards can be combined, subject to the motherboard and I/O slot limitations. When separate video cards are used, they must be specifically designed to allow a "cross-link" between them.
If a computer has two video cards that combine their outputs into a single video monitor, then one of three methods could be used to create the images.
- Alternate Frame Rendering (AFR): One graphics processing unit (GPU) computes all the odd video frames, the other renders the even frames. (i.e. time division)
- Scissors: One GPU renders the top half of each video frame, the other does the bottom. (i.e. plane division)
- Checker board: As the name implies, the image is split into smaller squares, which are assigned to different cards