Floating-point architectures (continued)
ANSI/IEEE 754-1985 (almost all CPUs since 1980) 32-bit, 64-bit, 80-bit (Intel and Motorola 68K), 128-bit (software, except IBM S/390 G5 and HAL SPARC64-GP)
Wider format means wider exponent range (unlike IBM S/360 and DEC PDP-11, VAX)