Hacker News new | ask | show | jobs
by fredoralive 5 days ago
It's basically a case of hardware acceleration. Most games consoles and fancier micros have some of dedicated graphics chip that handles the heavy lifting of generating graphics. So for basic game graphics the CPU largely acts as a manager adjusting things like tilemaps and their viewport offsets, and sprite locations (possibly updating things several times mid frame, for fancier effects). The exact setup differs between systems, tilemaps and spites like the SNES is common, though you also have setups with some sort of framebuffer and a Blitter to speed up drawing to it instead.

A traditional IBM PC has a "dumb" framebuffer, where everything is done by the PC. Simply scrolling the background by 1 pixel basically means redrawing a lot of the screen, and you have to keep track of what graphics behind sprite would need to be redrawn after they move etc. As a bonus, on early consumer level 386 and 486 machines you have a mighty processor, but the graphics card is often still on a 16 bit 8MHz(ish) ISA bus. The PC does have an advantage that it's more flexible, so stuff like 3D was easier to do than on a tile-and-sprite setup (especially once we had stuff like VESA and PCI).