What is OpenGL?

What is OpenGL

OpenGL is a series of graphics commands, functions that we can call as it is a c programming based API. It allows us to interact with the GPU, load data and render it from the GPU instead of the CPU.

It is a specification that represents how a graphics API should look like.

History of OpenGl

1. In 1991 SGI starts developing OpenGL.
   • runs on multiple platforms (windows, linux, mac etc)
2. Version 1.1 supports texturing and has a Fixed Function Pipeline.
3. Version 1.5 starts pushing OpenGL Extensions (Shaders).
   • Architecture Review Board allowed companies to start sending in extensions based off on their hardware.
4. Version 2.1 starts enforcing a programmable pipeline with GLSL language.
   • We as programmers are able to write and execute programs on the GPU.
   • We are able to offload the work on to the GPU instead of only relying on the CPU.
5. Version 3.3/4.0 starts removing the old functionality
   • Only the programmable pipeline is used.
6. Version 4.6 is what we use today.

Graphics Pipeline

A rendering pipeline is the following:

• The journey of a primitive(vertex, triangle, line, etc) from their 3D data to a 2D plane, our screen for example.

How would point(1.0,0.0,-5.0) look like on our screen?

1. Vertex Specification

Where we setup our vector data on the CPU.

2. Vertex Shader

Shader: programmable part of our pipeline. This is a feature of modern OpenGL that we can write programs on our GPU to control the graphics pipeline.

In the beginning we were not able to change the data being set to the GPU.

It executes on each vertex, positioning that vertex.

3. Tesselation (optional)

The process of breaking up bigger primitives into smaller primitives to create more detail in the rendering.

4. Geometry Shader (optional)

The process of generating more geometry from a given point on the GPU.

• useful for particle systems
• explosion effect

5. Vertex Post-Processing

Here you can do additional post processing for the data. Example: you create new points in using the geometry shader, you can then process that data right after.

6. Primitive Assembly

Assembling the final geometry by taking in all the data from the vertices. Includes other phases too:

• clipping: clips primitives if they are outside of the specified bounds.
• face culling: removes the faces of the primitive that are not necessary for the rendering.

7. Rasterization

The process of drawing in the actual pixels based on the vertices that we have loaded in so far.

• depth test: deciding what to do if shapes overlap

8. Fragment Shader

It executes once on each fragment -> sort of like a pixel. Determine the final color of each of the pixels that we rasterize.

9. Pre-Sample Operation

Different operations can be done here as the last step.