A Fixed-Point Arithmetic Compiler.
Abstract
Mini computers used within real time digital systems require mathematical operations to be performed at maximum speed. The most efficient use of their fixed point arithmetic units is made while manipulating fixed point data. Software simulation of floating point mathematics is too slow and requires additional core. A fixed point arithmetic compiler is designed to allow the programmer to express fixed point mathematical processes in equation form. Only the description of the original equation arguments is required. All intermediate results of the equation realization are scaled by the compiler to avoid clipping and truncation errors. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Jul 01, 1977
- Accession Number
- ADA059041
Entities
People
- Paul Craun Iii
Organizations
- Naval Ordnance Laboratory