Floating Point Package

This document describes the "Floating Point Package," a self-contained Scientific Programming System designed for the PDP-7 computer. It features interpretive instruction execution and employs double-precision normalized floating-point arithmetic. The system supports two data modes: a two-word mode providing 6 decimal digits of accuracy (range 10±38) and a three-word mode offering 9 decimal digits of accuracy (range 10±99).

The core of the system is an interpreter, entered by the EIM instruction and exited by LIM, which processes a rich set of commands. These commands cover data movement (load, store, complement), arithmetic operations (add, subtract, multiply, divide), control transfers (jump, subroutine jump, compare), and input/output functions. The I/O system supports teletype and paper tape, with flexible input and rigid output formats. Iterative operations are also supported, allowing sequences of commands to be repeated and addresses to be modified.

A comprehensive subroutine library is integral to the package, offering mathematical functions such as Sine, Cosine, Arctangent, Natural Logarithm, Exponential, General Exponentiation (A^B), and Square Root, mostly with 9 decimal digit accuracy. The document also provides detailed instructions for the operational use of the system, including I/O calling sequences, the structure of symbolic and binary tapes, and assembly procedures for user programs.