FPP1, A Floating Point Package for PDP-8 Computers.
Abstract
A 23-bit floating point (FP) package, FPP1, for PDP-8 computers was developed from the DEC YQ4B package with the primary aim of reducing its size, so that an adequate calculation and control software package could be fit into the 4K memory of a PDP-8 that was used to control an X-ray diffractometer in a wide variety of experiments in the AMMRC program of materials research. It should be particularly useful for systems with memory size limitations, as in the older PDP-8s or as part of an arithmetic package committed to ROM in modern units. It follows essentially the same approach and basic procedures as YQ4B, the major changes being the exclusion of E-format input and output (I/O) and the use of a different square-root algorithm, and it retains the same accuracy, generally six decimal digits in normal computations. It has the same mnemonic and numerical codes for the basic FP instructions and function subroutines and many of the same page 0 locations and address tags as the DEC package, so it should be compatible with minimal changes with many programs using a 4K memory. The linkages for multiple-field operations are left to those users to supply. FPP1 includes all the operations of YQ4b, excluding E-format I/O, plus two-word fix and float routines and an improved output roundoff. It occupies certain locations for a total of 1101 words, 26% less than the 1492 words of YQ4B, the saving being approximately 10% of a 4K memory.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 01, 1983
- Accession Number
- ADA134236
Entities
People
- Christopher B. Walker
Organizations
- United States Army Research Laboratory