REESE: a method of soft error detection in microprocessors

Future reliability of general-purpose processors (GPPs) is threatened by a combination of shrinking transistor size, higher clock rates, reduced supply voltages, and other factors. It is predicted that the occurrence of arbitrary transient faults, or soft errors, will dramatically increase as these trends continue. The authors develop and evaluate a fault-tolerant microprocessor architecture that detects soft errors in its own data pipeline. This architecture accomplishes soft error detection through time redundancy, while requiring little execution time overhead. Our approach, called REESE (REdundant Execution using Spare Elements), first minimizes this overhead and then decreases is even further by strategically adding a small number of functional units to the pipeline. This differs from similar approaches in the past that have not addressed ways of reducing the overhead necessary to implement time redundancy in GPPs.