Use of Systematic Code Based on Data Bits Weighing for Concurrent Error Detection Considering Error Structure Analysis

Abstract

This paper describes the algorithm of formationof new code with weight-based summation. Authors prove that this code detect all single and twofold distortions in data bits. Also other properties of this code considering hardware and information redundancy are stated. The main attention is paid to the ability of this code to detect 100% of single errors in the tested circuit with relatively low redundancy and analysis of structure of its errors that allow to reach this goal. Results are verified by the experiments using the set of LGSynth '89 benchmarks.