On Error Correction In Macro-based Circuits

Abstract

We consider the problem of correcting errors in a macro-based circuit. Our formulation of the problem allows the correction of errors that arise both in the context of design error correction, before the circuit is realized, and in the context where a physical circuit needs to be corrected. Two error classes are defined, namely, component errors and line errors. Both single and multiple errors are considered. Accurate correction procedures are given for single errors. Heuristics are given for correcting multiple errors. Experimental results are given to demonstrate the correction procedures presented.