>>41112615Basically, your average consumer gpu(be it a card or integrated with the cpu/soc) is lots of tiny processors that are really good at doing a specific task. By dividing meshes into triangles all the data can be broken up into chunks of the same information just with different values:
vertex1.x, vertex1.y, vertex1.z
vertex2.x, vertex2.y, vertex2.z
Ect...
triangle1.vert1=vertex1, triangle1.vert2=vertex2, triangle1.vert3=vertex3
triangle2.vert1=vertex1, triangle2.vert2=vertext2, triangle2.vert3=vertex4,
Ect...
so each little processor in the gpu can do its task without the extra overhead of figuring out what format the data is in, it just runs the numbers and sends it on to the next batch of tiny processors to determine what color each rendered pixel should be based on vertex color, the color of the texture at the corresponding coordinates determined by the UV mapping, lighting, ect. Sometimes there's even shaders used to calculate new geometry based on the input geometry, often used for grass, hair, smoothing out semi-compressed models for close-ups, snow/sand particles, things like that.
