Recent advances and new avenues in hardware-level reliability support

Discussion on some of the recent advances in logic- and architectural-level techniques to deal with transient errors serves as a "springboard" to motivate the need for hardware-level application-aware runtime checks, which the application can invoke from within its instruction stream per its dependability requirements. In contrast with traditional approaches of software and hardware duplication, alternative techniques such as fine-grained application-aware runtime checking offer more efficient, low-overhead detection, correction, and recovery.