16-bit Microprocessors
J. C. Cluley · 1987
Current 8-bit microprocessors have been developed steadily for over a decade and now provide a rich instruction set and a high rate of computation. Despite this the 8-bit data bus and the 16-bit address bus impose a limitation to the volume of program and data that can be handled conveniently. The 8-bit data limitation means that single-byte arithmetic is not sufficiently accurate for many purposes, since each item can be specified with a resolution no better than 1 part in 256, and rounding-off errors steadily worsen this resolution as more arithmetic is performed. Thus many operations must be carried out with double-length numbers. This more than doubles the execution time of the program and increases the program storage space. Floating point numbers may consist of 24 or 32 bits for the significant digits and 8 bits to position the binary point; handling these will be much slower than 8-bit arithmetic.