Efficient diagnostics algorithms for regular computing structures

Abstract

The paper contributes to system level diagnostics by two new diagnostics algorithms for faulty units identification in regular computing systems from testing results. The developed algorithms are based on the symmetric diagnostics model at system level. Effectiveness and complexity of the implemented algorithms were evaluated by experiments over several regular computing architectures (hypercube, torus, 2-dimensional grid). The achieved results were compared to existing diagnostics algorithms at system level.