How retro game consoles rendered graphics without modern

For over two decades, video game consoles relied on specialized graphics chips designed to perform a single task: convert game data into a television signal. These fixed-function processors could draw pixels but could not run custom code, a fundamental limitation that created the visual quirks players encountered in early games.

The NES exemplified this approach with its Ricoh 2C02 Picture Processing Unit, a rudimentary chip that worked with approximately 10 kilobytes of memory. The chip drew from roughly 8 kilobytes of tile artwork stored on the cartridge and 2 kilobytes of RAM holding background maps. The processor tracked up to 64 sprites in separate internal memory and generated a 256-by-240 pixel display built from 8-by-8 pixel tiles. As a CRT television's electron beam swept across each line of the screen, the chip determined which tile belonged at each position and cross-referenced which sprites occupied that line, then decided which pixel to display. This process happened in lockstep with the beam, preventing the CPU from safely rewriting graphics while they were being drawn.

One consequence of this architecture became infamous among players: sprite flicker. In Super Mario Bros., when too many Goombas lined up on the same horizontal row, some would disappear because the NES graphics chip had a hard limit of eight sprites per scanline. The system would alternate which sprites appeared, creating a flickering effect that became a defining characteristic of the era.

Different consoles approached the problem differently. The Atari 2600, released five years before the NES, required the CPU itself to handle graphics. Because its Television Interface Adaptor lacked a frame buffer, the processor had to feed new values to the chip for every scanline. Even earlier, the 1972 Magnavox Odyssey had no processor or memory at all, drawing its dots and lines using discrete diode-transistor logic. These machines were fundamentally designed around CRT technology, which is why retro games display more authentically on period-correct CRT televisions.

As 3D graphics became standard, console makers developed more sophisticated approaches. The PlayStation 2, unveiled in 1999, featured the Emotion Engine, a 128-bit processor built by Sony and Toshiba, paired with a separate Graphics Synthesizer chip. The two components ran at approximately 295 megahertz and 147 megahertz respectively. The Graphics Synthesizer handled only the final step of rendering—filling triangles with color and texture—while the Emotion Engine performed the geometry work through two vector units. One vector unit could manage up to 85 million 3D transforms per second, converting game models into on-screen triangles. This division of labor across multiple specialized chips became common among early 3D consoles.

The paradigm shifted dramatically when Microsoft released the original Xbox in November 2001. Rather than designing custom graphics hardware, Microsoft built the console around an Intel Pentium III processor and a graphics chip NVIDIA created specifically for the system, based on the GeForce 3. Critically, the GeForce 3, announced just months earlier in February 2001, was the first GPU that allowed developers to write their own pixel and vertex shaders—small programs that controlled how graphics were rendered. This marked the transition from fixed-function chips that could only draw to programmable processors that could compute.

The industry did not immediately standardize on this approach. The Nintendo GameCube, launched the same month as the Xbox, used a PowerPC processor from IBM, a design choice that influenced console architecture for the next decade. The Xbox 360 adopted three PowerPC cores from IBM in 2005, the PlayStation 3's Cell processor was built around a PowerPC core by Sony, Toshiba, and IBM, and both the Wii and Wii U ran updated versions of the GameCube's IBM chip. Graphics processors did transition to PC-style architecture more quickly, with the 360 using an ATI chip and the PS3 using NVIDIA, but central processors lagged behind.

The final shift to fully PC-based architecture came in 2013 with the PlayStation 4, which moved to eight AMD processor cores and graphics derived from AMD's Radeon PC cards. Today's PlayStation 5 and Xbox Series X each run eight AMD Zen 2 cores alongside AMD RDNA 2 graphics on a single chip. Once off-the-shelf GPU architectures became affordable enough to fit console budgets, designing graphics chips from scratch no longer made economic sense. The evolution from dedicated picture processors to programmable graphics units fundamentally transformed how games could look and play.