Floating Point Computation (Computer Architecture Book 4)

This book discusses the floating point data format in computation. It is somewhat architecture-neutral, but does restrict the discussion to binary computation in digital computers based on software and microelectronics technology. To understand why we need the complexity of floating point for scientific, engineering, and financial calculations, we need to review number systems, integer calculations in binary and decimal, and other representations systems, as well as the concept of negative numbers and zero. This work contains a broad list of floating point units and software packages.

Both software and hardware approaches are discussed for 8-bit to 64-bit integer machines. The IEEE standard for floating point is discussed, as well as the previous mainframe era standards.

A glossary and list of references are included.