Originally Posted by Twinnuke
This generation of consoles has come down to magic tricks magic tricks magic tricks. How well can we render only what you see to save on video memory. How many processes can we cut out and in which order so we can have more power to do this and that. How much memory can we save if we only access this by its address instead of pointers. It's all magic tricks
(GPU Access to Memory without an API is why Consoles were able to keep up with 3 generations of Graphics cards)
Almost all game programming is like this, only render what's in the field of view, occlusion culling, mesh and texture LOD, dynamic draw distances, animation frame skipping at larger distances or smaller amounts of screen space, particle system LOD, shader LOD used cheaper shaders at further distances, dynamic culling distances to hide smaller objects past certain distances. The list goes on, to get consoles to run as well as newer PCs they just have to use more techniques than the PC.