Some results and open problems concerning memory reconfiguration under clustered fault models

Abstract

Reconfiguration of memory arrays using spare rows and spare columns has been shown to be a useful technique for yield enhancement. This problem is NP-hard in general and hence, previous work has focused on branch-and-bound algorithms for smaller problems and approximation algorithms for larger problems. Recently, the performances of several algorithms have been evaluated under a probabilistic model for memory defects that assumes faults in memory cells occur independently. In this paper, the authors describe some results for the array reconfiguration problem under compound probabilistic models that allow for clustering of faults. >