As technology scales deep into the sub-micron regime, transistors become less reliable. Future systems are widely predicted to suffer from considerable aging and wear-out effects. This ominous threat has urged system designers to develop effective run-time testing methodologies that can monitor and assess the system's health. In this work, we investigate the potential of online software-based functional testing at the granularity of individual microprocessor core components in multi-/many-core systems. While existing techniques monolithically test the entire core, our approach aims to reduce testing time by avoiding the over-testing of under-utilized units. To facilitate fine-grained testing, we introduce DaemonGuard, a framework that enables the real-time observation of individual sub-core modules and performs on-demand selective testing of only the modules that have recently been stressed. Moreover, we investigate the impact of the cache hierarchy on the testing process and we develop a cache-aware selective testing methodology that significantly expedites the execution of memory-intensive test programs. The monitoring and test-initiation process is orchestrated by a transparent, minimally-intrusive, and lightweight operating system process that observes the utilization of individual datapath components at run-time. We perform a series of experiments using a full-system, execution-driven simulation framework running a commodity operating system, real multi-threaded workloads, and test programs. Our results indicate that operating-system-assisted selective testing at the sub-core level leads to substantial savings in testing time and very low impact on system performance. Additionally, the cache-aware testing technique is shown to be very effective in exploiting the memory hierarchy to further minimize the testing time.

In-field on-line testing techniques have recently been proposed for permanent fault detection caused by wear-out/aging-related defects manifesting during the lifetime of a system. Selective Software-Based Self-Testing (SBST) is one such paradigm focusing primarily on the recently stressed functional units of a multicore system at a sub-core granularity, in an attempt to reduce the application performance penalty caused by periodically testing the entire system. In this work, we complement our O/S-enabled framework DeamonGuard for on-demand (selective) SBST to support fault recovery capabilities. Towards this goal, we propose an efficient check pointing and rollback recovery mechanism which, upon fault detection, can restore the system to the most recently valid correct state and resume the normal operation assuming disabling of the faulty core, thereby leading to a healthy (but degraded) system. The work in this paper concentrates on reducing the number of stored checkpoints required when testing at a sub-core granularity, and minimizing the recovery penalty of such framework. We evaluate and demonstrate the overhead of the proposed recovery mechanism, and our results indicate a practical reduction in the number of stored checkpoints as well as a significant improvement in recovery latency for the cases where the faults are correlated with the stressed units.