Résumé
This work analyses and classifies strategies for sampling results of concurrent error detection (CED) schemes in transient fault scenarios. It shows that dealing with results - indicating the occurrence of transient faults in circuits - can require additional mechanisms to make the error indication useful for system's recovery procedures. The paper highlights that not all error indications are noticed by certain strategies of varied costs, and therefore their efficiencies in sampling results as well as the performance, power, and area overheads added to the CED schemes must be considered. The work then finishes presenting a qualitative comparison between existing strategies in function of design goals.