Microcode in Intel's 8087 floating-point chip: the scale instruction

Intel's 8087 floating-point coprocessor, introduced in 1980, became the industry standard for floating-point arithmetic by offering superior accuracy and compatibility with the IBM PC, making floating-point operations up to 100 times faster. Researchers reverse-engineering the chip's microcode have discovered that even the seemingly simple FSCALE instruction—which scales numbers by powers of two—requires over 140 micro-instructions and three levels of subroutine calls to handle various special cases, revealing the sophisticated complexity underlying the 8087's design.