Summary of architecture of graphic drawing system

What is the purpose of the initial production of the graphics system?
Initially, the purpose of making a dedicated graphics system was to liberate the general computer from the task of constantly refreshing the display. The Display processor module, which is added between the main computer and the monitor, has a traditional architecture, but adds instructions to display the entities on the display. The instructions that generate the image can be stored centrally on the host computer and then sent to the display processor. These instructions can be stored as a display list in the display processor's own memory. The display processor then repeats the instructions in the display list at a sufficient rate to avoid flickering so that the host computer can perform other tasks.

What is the throughput of the system?
The rate of data passing through the system, called the system throughput (throughput).

What is the system delay?
The time required for data to pass through the system, called the system's latency (latency). When evaluating the performance of pipelining, the tradeoff between latency and throughput must be weighed.

In graphics, what are the advantages of pipeline processing?
The advantage of pipelining is that it can significantly shorten computation time when doing the same for many data. This is why it is used in computer graphics, because graphics need to handle a huge number of vertices and pixels in the same way.

What constitutes the geometry data (geometry) of the scene?
The type of the entity and the set of vertices. Because objects in a scene are made up of a set of group elements, each element contains a set of vertices.

What are the 4 main steps to drawing a pipeline?
Vertex processing, cropping, and element assembly, rasterization, and element processing.

What are the main functions of the vertex processing module?
1. Perform coordinate transformation
2. Calculate the color value of each vertex

What is the function of coordinate transformation during the imaging process?
In the imaging process, there are many steps to be made in the transformation of the coordinate system. For example, in the virtual camera imaging mode, one of the main functions of observation is to transform the representation of an object from its defined coordinate system to the representation of the camera's coordinate system. Again, in any coordinate system, the representation must eventually be converted into a display coordinate system.

Why transform operations apply to pipelining architectures.
Because each transformation under the coordinate system can be represented by a matrix. The multiple transformations of a coordinate system can be expressed as a multiplication of matrices, so that multiple matrices are combined into a matrix by multiplying them. The third matrix is obtained by multiplying one matrix with another, and a transformation sequence is obviously well suited to the implementation of pipelining architecture. In addition, because the matrix used in computer graphics is always small (4*4), we can also do parallel computation in the transformation module of pipelining.

Why do we have to cut?
Because imaging systems cannot image the entire scene at once, based on this restriction, we have to crop.

How do I crop the visual field of a virtual camera?
This is achieved by clipping the body. Objects located inside the clip can be imaged, and objects outside the clipping body will not be imaged, and we say they are clipped out. Objects that span the boundaries of the clipping body are partially visible in the image. We will deal with some efficient cutting algorithms.

Why do we need to assemble the elements?
Because the cropping must be for the element-by-entity, not for vertex-by-pixel. Therefore, at this stage of the drawing pipeline, before clipping is performed, we must assemble the vertices to image the segments and the elements such as polygons.

What is the difference between the cropped entity and the stored information in the frame cache?
The resulting entities of the cropping module are represented by vertices, while the frame cache stores pixels. Further processing must be done between them.

What is the operation of the Rasterization module?
The module outputs a set of slices for each entity. An element is a potential pixel that carries information about it, including its color and location, and depth.

What is chip processing?
This is the last module of the drawing pipeline, using the elements generated by the Rasterization module to update the pixel in the frame cache, which is the slice processing. If the application is generating three-bit data, then some of the slices are not visible because the surfaces they define are behind other surfaces.

Copyright NOTICE: This article for Bo Master original article, without Bo Master permission not reproduced.

Summary of architecture of graphic drawing system