OpenGL es: (2) relationship between OpenGL es, EGL and glsl

OpenGL ES Be responsible for GPU Working , The purpose is adopt GPU Calculation , Get a picture , This image is actually a block in memory buffer, The color information of each dot is stored . And this picture is finally going to be displayed on the screen , So we need a specific window system to operate ,OpenGL ES There are no related functions . therefore ,OpenGL ES Have a good partner  EGL.

EGL, Full name :embedded Graphic Interface, yes OpenGL ES And the ground floor Native platform The interface between Windows systems . So the general process is like this : First , adopt EGL Get the phone screen Of handle, Get the configuration supported by the mobile phone (RGBA8888/RGB565 And so on , Represents the package in each pixel How many bits is the storage space of the color and other information ), Then create a block containing the default buffer Of surface(buffer The size of is calculated by multiplying the screen resolution by the size of each pixel information Calculated ) And for storage OpenGL ES State set context, And put them enable get up . then , adopt OpenGL ES operation GPU Calculate , Save the calculated results in surface Of buffer in . Last , Use EGL, Display the drawn picture on the phone screen .

And in the OpenGL ES operation GPU When calculating , It needs to be introduced OpenGL ES Another good partner of GLSL.

GLSL, Full name :OpenGL Shading Language, yes OpenGL ES Of the shader used in Language , With this language, you can write small programs to run in GPU On .

Here we need to mention CPU and GPU The difference between , Their functions are used to calculate , It is also composed of many nuclei , The difference lies in CPU There are fewer nuclei , But the computing power of a single core is relatively strong , and GPU There are many cores , But the computing power of each core is not particularly strong . at present GPU The main work of is to generate pictures ( Now there are also through GPU Perform high-performance computing _ Parallel operation , But it's not part of the discussion here ), The reason is that the picture is composed of many pixels , Each pixel contains a color 、 Depth and other information , And in order to get this information data , Calculation for each pixel , It can be completed by a unified algorithm .GPU I'm good at dealing with this kind of large-scale data , Use the same algorithm to calculate . And this algorithm , Is the use of GLSL It's written in Shader, for GPU Operation use .

From the perspective of graphics , All pictures are made of triangles . So pass OpenGL ES When drawing pictures , We need to pass OpenGL ES API Create for use in GPU Running on shader, And then through CPU Obtained image vertex information , Pass in GPU Medium Shader in . stay Vertex Shader Through matrix transformation , Convert vertex coordinates from the model coordinate system to the world coordinate system , Then go to the observation coordinate system , To crop coordinate system , Finally, it is projected into the screen coordinate system , Calculate the coordinates of each vertex on the screen . then , By rasterizing , The information of all pixels is obtained by interpolation , And in Fragment shader Calculate the color of all pixels in . Last , adopt OpenGL ES Of API Set the State , The obtained pixel information is depth/stencil test、blend, Get the final picture .

Reference link

1. Android Display System(2)

2. OpenGL ES 2.0 Knowledge collusion

from ：OpenGL ES: (2) OpenGL ES And EGL、GLSL The relationship between - Night traveler - Blog Garden