The history of x86 is presented as an accumulation of decisions made for particular customers and constraints, many preserved to keep software compatible. Intel’s 4004 began as a programmable processor for a Japanese calculator; a deal with Busicom gave Intel the right to sell it more broadly. The 8008 grew out of a terminal company’s design, and its inheritance explains lasting quirks: little-endian byte order suited the terminal’s bit-serial processing, while the parity flag helped detect transmission errors. The 8080 added 16-bit register pairs, a stack pointer, and a separate I/O address space, helping power the Altair and CP/M-era computer boom. Its successor, the 8085, simplified power requirements and multiplexed address and data pins.
The 8086 was intended as a temporary answer to Intel’s delayed, ambitious iAPX 432 project, but became the foundation of the PC architecture. Its design enabled 16-bit arithmetic and a larger address space while making 8080 assembly easier to translate, though it did not run 8080 machine code directly. That compatibility goal shaped familiar register roles: AX reflects the accumulator, BX the memory-pointing HL pair, and CX and DX inherited other conventional duties. The later family extended this legacy: the 80286 introduced protected mode and privilege rings; the 80386 brought 32-bit registers, paging, and virtual 8086 mode; and the 80486 added on-chip cache and floating-point hardware plus a five-stage pipeline. The result is an architecture shaped less by a clean blueprint than by successive compromises and backward compatibility.
Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.