Four-Word Floating-Point Package

This document describes the Four-Word Floating-Point Package (Digital-8-20-F-Bin), a software utility designed for use with the Digital-8-5-S 3-word floating-point system. This version offers increased precision by carrying accuracy to 35 bits using four 12-bit words for storage.

Key details include:

• Storage Requirements: The program occupies octal registers 7; 40–61; and 5600–7577.
• Usage: It is loaded via the Binary Loader (Digital-8-2-U) or a DECtape system, with a calling sequence identical to the 3-word package.
• Functionality: It supports various floating-point operations but excludes square root and square functions.
• Execution: The document provides specific timing estimates for operations such as FADD, FSUB, FDIV, FMPY, FGET, FPUT, FNOR, and FEXT.
• Implementation: The latter half of the document provides the full assembly language source code listing for the interpreter and associated floating-point subroutines.